4.2.1 Telephone 1.0 and policy 1.0

In the development of the telephone network in the Netherlands, private entrepreneurship played a leading role. For instance, the City of Amsterdam selected the Nederlandsche Bell Telephoon Maatschappij (NBTM) to build a telephone network under license from US-based International Bell Telephone Corporation (IBTC).⁴ The network was placed into service in 1881. Private initiatives resulted in 43 local networks being created in the major cities. These private networks were operational for a significant period of time, between 10 and 34 years, for an average of 24.5 years. In 1896 upon the expiry of the NBTM license, the City of Amsterdam decided based on public interest considerations (quality and not-for profit objectives) to assume the exploitation of the local telephone network. The municipalities started to assume ownership and operational control with the transfer of the NBTM network in Amsterdam in 1896. This example was followed by seven transfers from private to municipal ownership and twenty new municipal network start-ups.⁵ The municipal ownership lasted from 3 to 44 years, on average 21 years. From 1906 the state began to build and exploit new networks and to take over the private and municipal networks, essentially starting in 1916 and completed in 1927, with the exception of the municipal networks in the three largest cities: Amsterdam, Rotterdam and the Hague which transitioned in 1940 under the Nazi occupation.

For the construction of the inter-local network only one licence was granted, to NBTM which subsequently operated the network as a monopoly. The inter-local network transitioned to the State in 1897 after just 10 years of operation. (Schuilenga, Tours, Visser and Bruggeman, Reference Schuilenga, Tours, Visser and Bruggeman1981; De Wit, Reference De Wit1998)

4.2.2 Radio 1.0 and policy 1.0

The first experimental radio broadcasting service in the Netherlands was provided by a radio-amateur à Steringa Idzerda in The Hague in 1919. (De Boer, Reference De Boer1969; Blanken, Reference Blanken1992) In 1923 the Nederlandsche Seintoestellenfabriek (NSF) also started broadcasting services in Hilversum, using a license for manufacturing and testing. The content of the broadcasts was provided by the ʻHollandse Draad-looze Omroepʼ, the precursor of the Dutch broadcasting membership-associations (AVRO, KRO, NCRV, VARA, VPRO, each inspired by a particular world view). Philips was a major sponsor of these broadcasts. In 1935 the Dutch government consolidated all broadcast transmitters under a new entity, NOZEMA, which undertook renewal of the transmitters in a new location – Lopik.⁶ (Vogt, Reference Vogt1958; De Boer, Reference De Boer1969; Vles, Reference Vles2008)

In 1924, the first wire-based radio transmission was started by Bauling on a commercial basis in the town of Koog aan de Zaan. For the subscribers the service was cheap compared to owning a radio set: only a headphone or loudspeaker was required, hence the service appealed to the working class. Radio distribution licenses were granted on a non-exclusive basis. The industry was led by entrepreneurs, essentially in a ‘competition for the market’. A total of 800 private and municipally owned radio-distribution networks were built, with 35 networks in The Hague and 200 in Amsterdam of which only 9 served more than 1,000 subscribers.⁷ Some radio-distribution organisations operated as cooperatives, whereby volunteering members enabled lower tariffs. In some cities (e.g., The Hague) the telephone network was used for the distribution of the radio signals. Also 20 municipalities became radio-distribution operators exploiting a local monopoly position.

From 1931, these networks were directly linked to the radio stations by cable, thereby facilitating ‘high quality’ transmission.⁸ Radio-distribution peaked with 51 per cent of total radio users by the end of 1932, a position unique in Europe. In 1940 under the Nazi regime, the radio-distribution systems were placed under the responsibility of the PTT. As upgrading towards TV distribution was not feasible, the decision to terminate radio-distribution service was taken in 1964, and the process was completed in 1975. (Schrijver, Reference Schrijver and Regeringsbeleid1983; Arnbak, Van Cuilenburg and Dommering, Reference Arnbak, Van Cuilenburg and Dommering1991; Bordewijk, Reference Bordewijk2004)

4.2.3 Television 1.0 and policy 1.0

Television broadcasting on a regular basis was introduced in the Netherlands in 1951. Shortly thereafter, the use of equipment for the central reception and distribution of the signal (CAI) was introduced, primarily by housing corporations in high-rise buildings. These systems improved signal quality and reduced the need for individual antennas. The use of these systems was legally the prerogative of the PTT, which had the monopoly on distribution of broadcasting signals. However, during the 1950s and 1960s the practice was condoned for practical reasons. In 1963 the government proposed a central antenna system (CAS) to be built by the PTT. Casema, subsidiary of Nozema, jointly owned by PTT and the broadcasting organizations, was set up to provide the service. Pilot projects were conducted in The Hague. However, this initiative was rejected in parliament in 1975. In the same year, the monopoly position of the PTT was replaced by a concession system, which allowed the legalization of the CAI-systems. Only one concession was made available per municipality and the municipality had a preferential position to exploit the concession or have the concession exploited by a third party.⁹ Hence, cable networks became predominantly owned by the local governments. The exploitation of the networks was often delegated to the local or regional energy company or to private firms; also Casema was positioned to exploit these networks.¹⁰ The future introduction of optical fibre to the home was recognized but considered largely incompatible with the cable systems and, based on a depreciation period of 10–15 years, it was not expected before the years 1990–2000. A research project DIVAC (digital subscriber access) was begun to explore the related challenges in 1982.¹¹ (Jelgersma and Titulaer, Reference Jelgersma and Titulaer1981; Schrijver, Reference Schrijver and Regeringsbeleid1983; Davids, Reference Davids1999; NLKabel, 2009)

4.2.4 Reflection on period 1.0

The important role of the municipalities in the development of the telephone and cable networks as described above should not come as a surprise, as municipalities have had an important role in infrastructure development in general. Consider, for instance, the development of the road system, the drinking water supply and the sewage system, as well as the local distribution networks for electricity and city gas (see for instance: Milward, Reference Milward2005; Van der Woud, Reference Van der Woud2007; and also the Stadtwerke in Germany, Chapter 8). In their objective of creating an attractive economic and social climate for business and citizens, good quality infrastructures play an important role. Moreover, competition between municipalities is to a large extent based on infrastructure supply: hence, their continued interest in telecom infrastructure, resulting in their current involvement in the roll-out of fibre.

4.2.5 Telecom 2.0 and policy 2.0

4.2.6 Cable 2.0 and policy 2.0

Already in 1981, the ‘Swarttouw Committee’ had recognized that the cable distribution networks could develop towards a broadband network by facilitating two-way communication. If these CATV networks were connected on a national scale, they could form a competitive threat for the PTT. As the telecom policy was based on a primary role for the PTT, the Committee suggested mitigating this threat by creating an integrated network (Davids, Reference Davids1999). This position by the Committee was subject to both support and critique. The Scientific Council for Government Policy (WRR) concluded that cable networks were not necessarily to be integrated with the telephone network. The Ministry of Economic Affairs strongly opposed a monopoly of the PTT on all infrastructural networks.¹⁶ The opposition resulted in the appointment of the ‘Zegveld Committee’ to investigate the issue. Their report, published in 1987, recommended that integration should take place as part of one future broadband fibre network for audio, image and data communication. In 1988, this option was included in an update of the Telecom Law. However, in 1992 it became clear that the cable operators, united in sector organisation VECAI, opposed the integration. In the same year the Parliament formally rejected the option. (Eenhoorn, Reference Eenhoorn1994; Davids, Reference Davids1999)

The EU-initiated Reform, which required the transposition of Directives into national legislation, changed the position of the cable operators vis-à-vis KPN. They were allowed to interconnect their networks, which was done mainly through optical fibre. The change in telecom policy brought an end to the concession system and, hence, municipally owned networks could be acquired by private entities.¹⁷ By 2001, the result was that 29 cable operators were operating 647 cable networks, in a total of 504 municipalities, serving 6,159,972 subscribers. The consolidation process continued during the following decade around three firms: (1) UPC – owned by US-based Liberty Global, which acquired the cable network in Amsterdam (A2000¹⁸) in 1998; (2) Essent Kabelcom, a division of the energy company operating mainly in the north-eastern and south-eastern part of the country, combined with Multikabel (operating under the temporary name Zesko); and (3) Casema operating in the western and central part. The latter two merged to become Ziggo in 2008, leading to two major cable companies, Ziggo and UPC, essentially RTV monopolies in their own territories sharing the market, with at the second tier Caiway as the main player operating in the western and central part of the country. (Van Bockxmeer, Poel, Hulshof and Rutten, Reference Van Bockxmeer, Poel, Hulshof and Rutten2002; De Leeuw, Reference De Leeuw2009; NLKabel, 2009)

4.2.7 Reflections on the period 2.0

This period provides a nice illustration of the policy formation process in the Netherlands, with a strong emphasis on committees preparing policy advice based on broad stakeholder interactions, a pattern that we can observe again in the context of polices regarding broadband development.

The period has been relatively successful due to a central government advancing the liberalisation agenda: for example, market entry now requires only registration, not a license, and local governments facilitate entry by applying a ‘liberal’ application of the rights-of-way policy.

While the telecom reform resulted in new players entering the PSTN sector, thereby increasing competition, it led to consolidation of the local and regional RTV-cable monopolies into two major regional monopolies and a few smaller ones. This set the conditions for a techno-duopoly to emerge with the introduction of the Internet: on the one hand, between KPN, the dominant player in the PSTN, with access-based operators such as BBned and Tele2, and, on the other, the RTV-cable operators UPC and Ziggo.

4.3.1 Leading networking role by academic community

Concerns regarding the impact of micro-electronic technologies on employment led to an investigation by the Rathenau Committee. This Committee emphasized the benefits of ICTs and advised the government to introduce a more aggressive policy on the use of ICTs in 1978. This led in 1984 to the publication of recommendations under the heading ʻInformatica Stimuleringsplanʼ, which included initiatives and funding to promote ICT usage in higher education. The academic community responded and obtained governmental support for the development of a multi-year project plan, released in 1985 by the Samenwerkende Universitaire Rekenfaciliteiten (SURF) coordinating committee. In 1986 the government allocated NLG 300 million (approx. EUR 140 million) for the initial period 1987–1990 of implementation by the SURF Foundation. The data networking services to the universities and academic research centres were to be provided in cooperation with the incumbent telecom operator PTT (now KPN), for which they created a new legal entity SURFnet BV, with 51 per cent of the shares owned by the SURF Foundation and 49 per cent by the PTT.²⁰ The fourteen universities in the Netherlands agreed to obtain their data networking services from SURFnet for the next four years. (Verhoog, Reference Verhoog2008)

On 25 April 1986 the management of the.nl domain name registration was granted to CWI. Piet Beertema became the Dutch equivalent of John Postel, until SIDN (Stichting Internet Domeinregistratie Nederland) as a foundation assumed this role in 1997. On 17 November 1988, EUnet, with its central node at CWI, received ʻInternet connected statusʼ with the NSFnet and the first connection from the Netherlands to the Internet was established by the academic research community of CWI in Amsterdam, using the first ‘.nl’ country domain name: cwi.nl. Kees Neggers assumed chairmanship of RIPE-NCC, the Network Coordination Centre for European IP organisations.²¹

In 1987 SURFnet1 started to connect all fourteen universities using Datanet1, the packet-switched network based on the X.25 protocol operated by the PTT.²² However, the ultimate aim was to develop one common network based on OSI-protocols that would meet the requirements of the user community. Hence, the deployment of SURFnet2, based on OSI and X.25, as of 1989.²³ In 1990 SURFnet2 was fully operational using 64 kbit/s links and also provided transport facilities based on the TCP/IP-protocol using a Cisco router, introduced following its popularity in the USA.²⁴TCP/IP appeared to operate much better than expected and by the end of 1990 the decision was taken to focus on TCP/IP and to create a multi-protocol infrastructure SURFnet3, with 2 Mbit/s connections; the first step towards broadband access. By the time of its completion in 1992, TCP/IP had become the dominant protocol. In a parallel effort, SURFNet together with NORDUnet²⁵ created ‘Ebone’, the first European IP-backbone to become operational, in 1992. In the same year Erik Huizer became a member of the IETF and Kees Neggers was one of the founders of ISOC. (Verhoog, Reference Verhoog2008)

As the Internet became more easily accessible (e.g., through the introduction of the Mosaic browser in 1993) SURFnet observed strong growing traffic volumes and decided to become the initial customer of KPN in the implementation of an ATM-based network under the name of SURFnet4 in 1994, initially providing up to 34 Mbit/s access connections and a 155 Mbit/s data rate in the backbone; to become an IP-based backbone network with data rates of 155 and 622 Mbit/s. (Verhoog, Reference Verhoog2008)

In 1994 the plans for the Amsterdam Internet Exchange were announced. Through the AMS-IX Internet service providers (ISPs) exchange Internet traffic, a service initially managed by SURFnet. In 1997 twenty participants created the AMS-IX Association to establish a Limited Liability Company (as a not-for-profit cooperative) to assume the exploitation and operational management of the exchange in 2002. In 2003, with 178 participants, AMS-IX became the largest Internet exchange worldwide. (Verhoog, Reference Verhoog2008)

In 2000 the pilot phase of SURFnet5 started with 80 Gbit/s DWDM on glass fibre in the backbone and access connections of up to 20 Gbit/s. The network was completed by early 2002, connecting 170 institutes with approximately 500,000 users as of 2008. Even higher capacity needs were foreseen; hence, a hybrid network was developed providing IP-services next to optical network services as part of SURFnet6.²⁶ Following experiments started in 2002, a point-to-point network of optical light paths ([lambda] s) was implemented in 2004 with a node in the Netherlands called ‘NetherLight‘ connecting to ‘StarLight’ in Chicago. The first 10 Gbit/s lightpath was delivered between the University of Groningen and SARA²⁷ in Amsterdam in support of the Lofar project in early 2006.²⁸ By the end of 2007, lightpath #220 was realized, the demand being much higher than anticipated, in particular for the Optical Private Network (OPN) feature. ‘NetherLight’, the node located at SARA in Amsterdam, became the European node in the GLORIAD network, a research network initiative to connect the USA, Russia and China through a global optical ring. (Verhoog, Reference Verhoog2008; SURFnet, 2009)

In a parallel development, student housing complexes also were connected to SURFnet to provide students the same facilities as they enjoyed at the university campus. In 1990 ninety students of the University of Nijmegen were connected using the RTV-cable network and in Wageningen a pilot project was also set up. In 1998 the project ‘Student Online’ was aimed at connecting as many students and university staff as possible using the telephone network. In 1998–9 a thousand people were connected to SnelNet in Amsterdam, a collaboration between KPN, SURFnet and NOB Interactive Media providing 2 Mbit/s ADSL connections with a first trial of delay TV.²⁹ In 2000 further implementation pilot projects started with ADSL in Amsterdam, Delft and Twente, as well as the ‘Fibre to the Dormitory’ project which connected 2,600 student apartments in Delft with data rates of 10–100 Mbit/s. This project was followed by similar projects in the university cities of Groningen, Nijmegen, Enschede, Utrecht and Wageningen. Also in 2002, in the ‘Freeband Impuls’ project commissioned by the Ministry of Economic Affairs, SURFnet started trials to provide wireless access across multiple university sites for students and staff, leading to the now globally available eduroam service supporting users worldwide. (Verhoog, Reference Verhoog2008) SURFnet continues to drive the technology agenda, for instance in cooperation with mobile providers in the integration of 4G and Wi-Fi as part of the GigaPort3 project.

4.3.2 Internet access for the business user

In the business community, data networking between IBM computers was facilitated by V-net, a host-to-host file transport network, using synchronous data link protocols (SDLC) developed from 1975 onward. By 1979 it included 285 nodes in Europe, Asia and North America serving some 1,400 computer systems. The early connections operated on dial-up telephone lines supporting data rates of 1200 to 2400 bit/s. (Van de Ven, Reference Van de Ven1984; Wikipedia, 2009d) In 1989, NLnet started to provide Internet access to enterprises, using dial-up connections. (Verhoog, Reference Verhoog2008) This evolved into dedicated Internet access and IP-networking services provided by specialized companies, such as BBned, Easynet, Eurofiber, Tele2, UNET and the local branches of global operators such as AT&T, BT Global Services, Colt and Verizon.

4.3.3 Internet access for the residential user

Anticipating growing demand for end-to-end digital services the PTTs under the auspices of the ITU agreed to create a new, fully digital, circuit-switched system called ISDN (Integrated Services Digital Network). The ISDN would build upon the already digital backbone network (IDN – Integrated Digital Network) and involve upgrading the access network to digital, providing narrowband access at 2×64 kbit/s (basic rate) to the residential user and 2 Mbit/s (primary rate) to the business user. The elaboration of the ISDN recommendations in the CCITT required three four-year study periods to be completed by1988. The development of the European standard by ETSI was completed in the early 1990s. Testing ISDN in practice started with a pilot project in Rotterdam in 1987.³⁰ Making the telephone switching systems ISDN-ready and having ISDN compatible terminals available led to a general introduction of the service in 1993. (Kaasschieter, Reference Kaasschieter1985; Ekkelenkamp, Verstraaten and Wijbrands, Reference Ekkelenkamp, Verstraaten and Wijbrands1992; Tanenbaum, Reference Tanenbaum1996)

The same year the Mosaic browser was introduced to unlock the Internet for the wider public and the Hack-Tic³¹ community established XS4ALL, the first dial-up Internet access provider open to the general public. (Hack-Tic, 1993)³² In January 1994, in cooperation with De Balie³³, followed the inauguration of FreeNet ʻDigital Stad Amsterdamʼ aimed at bringing together citizens and policy makers in an online community (Meerman, Reference Meerman2004). VuurWerk Internet³⁴ started in 1996 as the first entity to provide web hosting and domain name registration services. (Wikipedia, 2010b)

While the notion of integrated services such as telephony, telefax, textfax, videotext, teletext, video and teleworking, based on circuit switching was not a huge success, ISDN did provide Internet access at 64 kbit/s at the price of a telephone call. By October 2000 17 per cent of internet access connections were using ISDN.³⁵ (CBS, 2001)

In 1996 CAI Westland³⁶, a cable network operator in the western part of the country, was the first cable operator to provide Internet access, using the DEMOS-1 system of DeltaKabel Telecom at a data rate of 115 kbit/s, using a flat rate. (Verbree, Reference Verbree1997; Wikipedia, 2010a)³⁷ Shortly thereafter, the major cable network providers became Internet access providers: UPC using the Chello brand name, Casema under the French Wanadoo brand, and Essent Kabelcom using the @Home brand.

The uptake of the Internet was directly related to the diffusion of PC to homes and businesses. Hence, the tax-break provided for the purchase of a ‘home PC for business use’ (PC privé project) should be mentioned as a factor that positively contributed to the adoption of the Internet.

4.3.4 Industry development and dynamics

The emergence of Internet access and service providers in the early 1990s was followed by consolidation in the late 1990s and early 2000s, driven by competition on margin and increasing needs for investment capital. The incumbent telecom operators were the main buyers, extending their ISP footprints and services portfolios. In many cases brand names were retained after an acquisition, representing a different service positioning and different market segments being targeted, including residential users and business users. In 1997 NLnet sold its commercialized Internet provision activities to UUNET, a subsidiary of WorldCom. The XS4ALL foundation was transformed into a corporation, XS4ALL Internet Ltd, in 1996 which in 1998 became a subsidiary of KPN Telecom. In 2006 XS4ALL acquired hcc!net and in 2007 Demon NL. In 1997 Planet Internet, owned by KPN, merged with WorldAccess, the ISP established by Videotex. IPS acquired WorldOnline and was subsequently taken over by Tiscali in 2000. (Meerman, Reference Meerman2004; XS4ALL, 2007; NLnet, 2010) In 2004 the Top-10 ISPs by size was composed of: XS4ALL, Demon, Planet, Bbeyond, BabyXL, VuurWerk, Tiscali, Lycos, PSINet and Wanadoo. (Meerman, Reference Meerman2004)

Internet service provision is tightly coupled to the underlying Internet access provision. For instance, cable network operators provide access only to their own ISP, while KPN provides access to multiple ISPs, both wholly owned subsidiaries and, through unbundling of its infrastructure, to independent ISPs.

4.3.5 Reflections on the emergence of the Internet.

An important engine of Internet infrastructure development in the Netherlands has been the academic community through SURFnet. Supported by additional funding from the Ministry of Economic Affairs and the Ministry of Education, Culture and Science. SURFnet has pushed the operators, KPN and Telfort, and equipment providers, such as Cisco, Lucent Technologies and Nortel, into the deployment of leading-edge technologies; as well as opening up the market for dark fibre connections. SURFnet was instrumental in the set-up of the AMS-IX, thereby creating a highly attractive up-to-date communications environment in the Netherlands. Via its ADSL and fibre to the dormitory projects, SURFnet has been instrumental in developing a market demand for broadband access at home.

More generally, the tax-break provided for the purchase of a ‘PC at home for business use’ (PC privé project) has also contributed positively to the adoption of the Internet.

4.4.1 Competition in internet access

In 1996 through a change in the Media Law, cable network operators were allowed to engage in competition. In 1997 also the provision of telecommunications services over cable networks was permitted. (MinOC&W, 2000) This allowed the cable sector to diversify its service portfolio to include Internet access and to provide voice/telephony service. This required significant investments to upgrade the cable network from one-way analogue RTV-signal distribution to two-way digital communications services provisioning. Given the broadband nature of the coaxial cable-distribution network, the cable sector took the lead in providing end-users with high data rate Internet access, up to 30 Mbit/s.³⁸ By the end of 1999 some 130,000 digital cable modems were in operation.³⁹

In 2000 the first ADSL connections were provided by Demon using the network facilities of BabyXL Broadband (since 2002 part of Tiscali) and by XS4ALL, providing data rates of 512/64 kbit/s (downstream/upstream) and 1024/256 kbit/s. KPN started testing ADSL as part of the SnelNet project in 1998–9 and the roll-out started later in 2000. (MinV&W, 2000a; XS4ALL, 2001; Meerman, Reference Meerman2004)⁴⁰ In the following years ADSL gained significant market share.

In the ‘competition for the market’, data rates were increased step by step, resulting in a tit-for-tat competition between KPN, the wholesale-base operators BBned and Tele2, and the RTV-cable operators; only Essent Kabelcom with @Home followed a strategy of providing the highest data rate possible. (Van den Berg, Reference Van den Berg2004; Albrecht and Achterberg, Reference Albrecht and Achterberg2008)

By March 2004 the cable operators collectively provided 1 million homes with Internet access, up from 250,000 in the year 2000. ADSL services were led by KPN with 913,000 subscribers, Wanadoo 350,000 (which includes cable access), Tiscali 140,000 and Zonnet 127,000. In 2004, 50 ISPs served a total of 4.5 million online households, approximately 65 per cent of total households. (Meerman, Reference Meerman2004)

By 2008, fixed broadband connections appeared to have reached a saturation level with approximately 5 million connections. The market share of ADSL was stabilizing, reaching 45 per cent (Kool, Arno, De Munck and Huveneers, Reference Kool, Arno, De Munck and Huveneers2008).

The competition in Internet access resulted in a top-two position in the penetration of broadband within Europe; whereby broadband was taken as downstream capacity equal to or higher than 144 kbit/s. (EC, 2008a) In terms of the broadband performance index, again, the Netherlands assumed a top-two position, measured as a composite index of socio-economic context including the uptake of advance services, data rates, price, competition and coverage. (EC, 2008b)

With the market reaching apparent saturation, the ‘competition in the market’ increased in importance, and the ‘race on speed’ continued, with, for instance, UPC increasing the download data rate to 120 Mbit/s for its premium subscription offer in Amsterdam (NRC, 2008), and Ziggo announcing the upgrade from DOCSIS 2.0 to 3.0. (De Vries, Reference De Vries2009)⁴¹

4.4.2 Competition in voice

Competition between the CATV and PSTN networks started with Internet access. However, following the end of the PTT monopoly on telephony, the cable companies were allowed to provide voice services. Initially various techniques and standards were used, with very limited success. From 2004 onward the standard technique became Voice over IP (VoIP), supported by DOCSIS 1.1. The flat-fee-based offering was most attractive. In 2005 KPN responded with the introduction of VoIP based on ADSL, also on a flat-fee basis.

In 2004 Caiway started offering telephone service over the cable network. (Wikipedia, 2010a) In 2005 XS4ALL introduced VoIP, including a Webphone application as of July of 2006. (XS4ALL, 2007)

4.4.3 Competition in RTV-distribution

While the Internet protocol supports the distribution of video signals, the Internet is not ideally positioned for broadcasting TV signals in real-time. To facilitate infrastructure-based competition on the basis of Triple-play⁴² offerings – television, telephony and internet combined – a consortium of KPN with broadcasting entities obtained a license for nationwide terrestrial distribution of digital radio and television signals (DVB-T). The number of channels was limited to 23 TV-channels, plus 19 radio channels. In 2003, the service was launched under the name ‘Digitenne’ with country-wide coverage becoming possible after the shut-down of analogue TV broadcasting, the so-called Digital Dividend.⁴³ Meanwhile, KPN obtained 90 per cent of the shares and is now the sole provider of digital terrestrial TV broadcasting. The content is provided by the NPO (National Broadcasting Organisation) and by KPN. (KPN, 2009; Wikipedia, 2009b)

Also in 2003, KPN started to offer Video-on-Demand (VoD) via the Internet. From 2008, the company has added DVB services to handheld devices (DVB-H). Subsequently this services was terminated as TV-over-LTE is the more future-proof alternative.

From 2005, the major broadcasting stations also started to use the Internet as an alternative distribution channel, to offer delayed viewing of their programs (e.g., www.uitzendinggemist.nl).

OPTA, the Dutch National Regulatory Authority, concluded that, despite competition from Digitenne and satellite operators, the CATV operators in their respective service areas had significant market power, and, hence, regulation was deemed necessary (OPTA, 2009a). This new regulation replaced the 2006 ruling requiring access for RTV content providers to transfer programs using the RTV transmission service provided by the cable operators (OPTA, 2006a). By the end of 2009, the cable networks were scheduled to open-up for competition through wholesale access: that is, alternative service providers may access the CATV network for the transfer of RTV-signals. However, KPN was excluded from access to the cable networks, as cable operators are excluded from access to the network of KPN. (Hijink, Reference Hijink2008) This decision was challenged in court and was rejected in August 2010; the court did not support the market analysis on the regional level. During 2011, proposals were submitted in Parliament in a renewed attempt to open-up the cable market. The new (national) market analysis by OPTA concluded that the market for digital television was sufficiently competitive and analogue television was going to be replaced by digital so there was no need for further regulation of the cable market. The regulator has retained this position for the regulatory period 2012–2014. (ACM, 2013e)

4.4.4 Reflections on broadband competition

Starting with a multitude of municipal cable operators in the 1970s the sector has been consolidating until very recently. In 2007 a major merger took place as the owners of Casema, the private equity firms Cinven and Warburg Pincus acquired Multikabel and Essent Kabelcom, to be merged into a new entity called Ziggo. The Dutch cable market is now served by two major players UPC and Ziggo, plus a few niche players.

With Ziggo owned by private equity firms, the endgame started with an IPO in March 2012. Through Barclay’s Bank as issuing bank, Liberty Global, the parent company of UPC, acquired 12.7 per cent of the shares of Ziggo in March 2013. At that time rumours about a take-over of Ziggo by Liberty started to appear in the press. By July Liberty had extended its position to 28.5 per cent. In October 2013 Ziggo confirmed the receipt of a tentative take-over offer by Liberty, which they rejected as inadequate. Ziggo confirmed it was in discussion with Liberty Global in December 2013. Early in January 2014, Bloomberg reported that the take-over could be concluded shortly.⁴⁴

KPN is the leading provider of ADSL under multiple brand names, in part acquired through consolidation (Het Net, Planet, XS4ALL). Moreover, through unbundling, other ADSL providers make use of the local loop owned by KPN (e.g., BBned/Tele2). The market is segmented as not all users require the highest possible data rates.

The market for broadband has become characterized by a techno-duopoly, consisting of KPN and its retail providers, providing broadband services on a national basis using the PSTN, and two major cable providers, Ziggo and UPC, each operating in a different part of the country, collectively providing close to national coverage, using the RTV network. There is strong competition in the market, based on price and data rates provided.

The question that can be raised is whether two competing firms are enough for a healthy development of the sector. (OPTA, 2006b) Clearly, the incumbent players are optimizing the utilization of their networks through an evolutionary approach. While fibre plays an important role in this evolution in the backbone and metropolitan networks, as well as in the feeder network to the street cabinets (FttC), the transition to FttH is apparently not (yet) attractive enough or considered not (yet) necessary from their business perspective. Third parties, in particular municipalities, were claiming ‘market failure’ and hence taking initiatives for the development of Fibre-to-the-Home – considered as the (more) future-proof trajectory.

This claim of ‘market failure’ has been investigated by the CPB, the Dutch bureau for economic policy analysis, following questions raised in Parliament. (Van Dijk et al., Reference Van Dijk, Minne, Mulder, Poort and Van der Wiel2005) The core question was whether the then-current developments would be sufficient to match the increasing demand for broadband.⁴⁵ The investigation concluded that ʻmarket failures in the Netherlands are limited to market power and, to some extent, to an increase in spill-over effects of knowledge. Current regulation by OPTA (the national regulatory agency) adequately deals with the issue of market power.ʼ The investigation suggested that ʻthe markets for broadband in the Netherlands function well in terms of competition (static efficiency) and innovation (dynamic efficiency). Hence, the best policy is to rely on market forces.ʼ

The central government’s policy remained focused on facilitation of, rather than intervention in, the development of broadband markets in general and the deployment of FttH in particular. But it did not prevent or block local governments from becoming involved, to a degree.

4.5.1 Towards policy 3.0

In this section the role of the government in broadband development in general and FttH in particular is reviewed, starting with the Lisbon Agenda formulated in 2000. (EC, 2000).

In the translation of the Lisbon Agenda into national policy, the Dutch government emphasized access to a high quality broadband infrastructure, whereby it assumed a technology-neutral position, driven by the opportunities these networks provide, rather than the technology or data rates being deployed. In short, the market should do the job. (MinEZ, 2004)

Nonetheless, the market can be assisted by providing information and reduced uncertainty. A number of investigations and reports by the Ministry of Public Works or the Ministry of Economic Affairs⁴⁷ were published around the year 2000, all aimed at broadband development: ʻDe Digitale Delta: e-Europa voorbijʼ (the translation of the eEurope Action Plan into national policy actions), ʻBouwstenennotitie breed-bandʼ (Building blocks for broadband), ʻSlim Graafwerkʼ (Clever trenching). Also in support of the municipalities wishing to contribute to the topic: ʻBreedband Internet voor/door Gemeenten – Stedenlinkʼ (Broadband Internet by and through municipalities) and the ʻGedogen mits…of een totaal andere wegʼ (Position paper on underground infrastructures and trenching rights) by the VNG (Cooperation platform of Dutch municipalities). (MinEZ, 1999; MinV&W, 2000b; Maclaine Pont and Van Till, Reference Maclaine Pont and Van Till2001)⁴⁸

In the report De Digitale Delta the Dutch government outlined its ICT-related policy objectives as follows:

1. (1) the availability of a first class, affordable, accessible and reliable telecommunications infrastructure;

2. (2) highly developed knowledge base and knowledge infrastructure and high degree of innovation based on close interaction between suppliers and users of ICT applications;

3. (3) professionals and citizens that have access to the latest electronic media and are capable of exploiting its opportunities;

4. (4) laws and regulations that facilitate the development and application of ICT; and

5. (5) optimal use of ICTs in the governmental sector, which may serve as exemplar of ICT-usage. (MinEZ, 1999)

In 2000, the Ministry also took the initiative to create ʻKenniswijkʼ (Knowledge Quarter), a real-life environment to ‘test the consumer market of the future’, aimed at (1) breaking the perceived deadlock between infrastructure development and service development; (2) to improve the competitive position of the Netherlands; (3) to share the knowledge obtained through ʻStedenlinkʼ (City link); and (4) to provide input for government policy. The all-fibre pilot project in the city of Nuenen, involving 7,500 households in an area representative of the Dutch population, was part of the ʻKenniswijkʼ project. (Kools and Serail, Reference Kools, Serail, Onderzoeksreeks and Zaken2003)

By 2003–2004 the focus of the telecom policy was on assuring the proper functioning of markets and consumer protection, with an emphasis on continuity, safety, trust and quality. (MinEZ, 2003) Based on the ʻBreedbandnotaʼ , which recognized the various perspectives of market parties in relation to the roll-out of broadband, the policies can be summarized as shaping favourable preconditions for broadband development. (MinEZ, 2004)

4.6.1 The FttH case of Amsterdam

In 2001, the City Council of Amsterdam, observing ad-hoc plans being made for the deployment of fibre to the business and fibre to the home by private parties, placed the topic of fibre networking on the political agenda. The Council was concerned about a possible ‘digital divide’ emerging in the city, which was considered economically and socially undesirable. A second reason for involvement was the wish to channel the related digging activities to reduce the level of inconvenience to the public.

To address the issues properly, the City Council commissioned an investigation into: (1) the economic and social importance of FttH; (2) the expected capabilities of the existing infrastructure; (3) the expected demand for broadband by small firms and citizens; (4) whether the City should exert any influence on the developments; (5) if so, what goals should be pursued and what instruments should be used; and (6) what might be the financial implications. The resulting report Amsterdam the Big Cherry?, issued in 2002, concluded that the market will most probably not deliver an open infrastructure, nor fibre connections to every home and enterprise in Amsterdam, within the next 15 years. Hence, considering the economic and social importance, the local government may wish to take the initiative and guide the developments towards the creation of an open fibre infrastructure. (Weeder and Nijland, Reference Weeder and Nijland2002) Based on the report, the City Council decided to investigate a more active involvement in the roll-out of a dense open-access fibre network. The next steps were investigated by a committee chaired by former Minister of Economic Affairs Andriessen. Meanwhile, the City Council decided to pursue fibre deployment in a new-built city area called Zeeburg. In 2003, the Andriessen report provided the necessary guidelines for the implementation of a city-wide FttH project, as a public-private partnership (PPP): the implementation of an open-access, passive network infrastructure, supporting service level competition. (Weeder, Nijland and Hotho, Reference Weeder, Nijland and Hotho2003) The recommendations were turned into an implementation plan ʻAmsterdam: breedband in bewegingʼ which was ready in 2004. (Weeder, Nijland and Konijn, Reference Weeder, Nijland and Konijn2004; Citynet, 2009)

By 2004, out of the top-30 cities⁴⁹ in the Netherlands 29 cities had (draft) plans for implementing fibre networks to connect governmental buildings, 23 had plans for Fibre-to-the-Business, and 15 had plans for Fibre-to-the-Home. (Linssen, Reference Linssen2005) One of the smaller but highly important initiatives was by the municipality of Appingedam. Early in 2004 it decided to commence the network roll-out, but had to interrupt the work within a few months, following a court decision in a case filed by the local CATV operator (Essent Kabelcom) against the city. The court decided that the initiative had to be presented to the European Commission for approval of (alleged) state support. The decision by the European Commission, issued in 2006, concluded that in this case financial support by the state was not allowed, given that the newly built network would represent unfair competition with the broadband networks of KPN and Essent. (Weeder, Nijland and Konijn, Reference Weeder, Nijland and Konijn2004; Schouten, Reference Schouten2006) Subsequently, Amsterdam and other municipalities filed their applications for EC approval, making sure their financial involvement was provided on the basis of the so-called ʻmarket economy investor principleʼ.

In 2004, the municipality of Amsterdam started the European tendering process for construction of the network, resulting in the award to the combination of contractors BAM/DRAKA in 2005. In a negotiated arrangement, the exploitation of the fibre network on a wholesale basis was awarded to BBned to provide open, non-discriminatory access to retail operators. Subsequently, for the implementation of the fibre network, the city established a limited liability company Glasvezelnet Amsterdam (GNA) in 2006, in which the municipality participates for 1/3, the four housing corporations for 1/3, and investors for another 1/3; each party contributing EUR 6 million. (Citynet, 2009) This enabled the first phase of implementation, 40,000 connections, to start in 2006, subject to approval by the European Commission which was granted in December 2007. (NRC, 2007) It should be noted that UPC, the regional cable operator, has contested the fibre network plans of the municipality in the courts, at national and European level, albeit in vain.

Nonetheless, the debate on the involvement of municipalities in telecom infrastructure development in the Parliament resulted in tighter rules for participation of municipalities (and housing corporations) in telecommunication projects, becoming part of the 2006 revision of the Telecom Law. (Doorn, Reference Doorn2006; NU.nl, 2006) However, most of these restrictions were removed in the 2010 revision of the law, as part of an overall revision of laws and regulations to stimulate the growth of the economy following the crisis.⁵⁰

4.6.2 The FttH case of Almere

Another municipality placing fibre high on the political agenda was Almere – near Amsterdam – a young and fast-growing city in the reclaimed Flevopolder. As a city with the ambition to leverage ICTs in all fields under its responsibility, including housing, employment, leisure and care, and inspired by the Stokab model in the City of Stockholm⁵¹ and information from a fact-finding mission in the USA by the Ministry of Economic Affairs, it commissioned a study ʻAlmere op glasʼ in 2001. (Stratix, 2001) This study suggested an active role for the local government to implement an open-access ʻFirst Mileʼ infrastructure in a new city development, Almere Poort. It proposed a demand-driven community-based network, whereby the municipality leverages its role given by the Telecom Law to coordinate infrastructure developments. The city would take the lead in creating the civil infrastructure as a public domain activity. A separate entity (net-co) acting on behalf of the users would be responsible for the implementation and exploitation of the network, subject to a tender. The end-user services were to be provided in competition. As such a maximum degree of unbundling would be achieved. By the city taking the lead in the realization of this network, the need for operators to create another fibre network in parallel would be pre-empted, but not excluded. By leveraging the use of the network by government institutions, including city offices and schools, effective demand aggregation and efficient roll-out of the network would allow prices for fibre to be similar to those of ADSL.

In the next phase three parallel developments became important. The City Council preferred as a first project connecting an existing city area that required an economic boost as it was suffering from a relatively high degree of unemployment. At the same time the City wanted to attract high-tech industry and institutions, such as the Institute for Information Engineering (linked to the Technical College in Amsterdam) and a division of SARA. These institutes require high data rate connectivity and thus connections with AMS-IX were established in 2001. This in turn made it attractive for BT and later Global Crossing, KPN and COLT also to set up facilities in Almere. Interactions with the existing operators showed that they were reluctant to participate in the pilot project due to conflicts of interest and a lack of willingness to assume risk. This combined result provided the momentum for the pilot phase of the project to be directed by the City, ultimately targeting 1,700 homes and 450 businesses in business parks, to be connected in 2003.⁵² The project was subsidized with EUR 1 million by the Ministry of Economic Affairs. A special entity, the Almere Fibre Company Ltd (AFCo), was established for the implementation and maintenance of the civil works, with the City as the owner. In the subsequent tender process the newly established First Mile Venture (later to acquire the ISP Unet and to assume this name) was selected as the infrastructure operator. (GemAlmere, 2004; Jansen, Reference Jansen2004; Halsema, Reference Halsema2012)⁵³

The FttH part of the project was not successful because the willingness to pay was too low, resulting in a too low uptake but the FttB was successful and that became the focus of Unet’s business. In 2006 the City followed up with a demand-aggregation project among local institutions, administrative buildings, schools, health centres, etc., leading to an FttInstitute project to connect 200 locations with fibre. The tender was won by Unet. This was followed by an FttB-project expanding the backbone to connect all business parks by fibre. The FttH initiative was rejuvenated by the idea of modelling it on the Amsterdam arrangements with 1/3 participation by the City, 1/3 by housing corporations and 1/3 by private investors. Initially KPN decided not to participate in the venture, but at the end of 2007 it joined the project with Reggefiber. Based on the evaluation of the pilot, the City made internal rearrangements such that implementation approvals could be obtained in 1–2 weeks rather than 13 weeks. This facilitated the connection of 60,000 homes within two years. XMS and Lybrand became new service providers, in addition to KPN and XS4ALL. By taking a leading role and being prepared to kick-start the project with pilot funding, the City of Almere assured that the private sector is actively deploying Fibre-to-the-Home in new and existing parts of the city. (GemAlmere, 2004; Jansen, Reference Jansen2004; Halsema, Reference Halsema2012)

4.6.3 Reflections on municipal initiatives

The involvement of so many municipalities in fibre initiatives, large and small, suggests that infrastructure supply is considered very important for economic development; the case of Almere is a clear example. The awareness of limited bandwidth available on copper and perception of fibre as future-proof explains their early involvement. While there were many initiatives, not too many have ultimately led to the actual deployment of municipally-led fibre. Issues around alleged state-aid, requiring upfront clearance, and the issues of funding and managing have been too cumbersome in many cases. Nonetheless, the push by municipalities has forced earlier deployments although in the end it have been commercial parties which became involved in the implementation and operations of municipal fibre networks, in part or in whole. See also Section 4.8 on Reggefiber and Chapter 7 for the municipal fibre deployments in Sweden.

4.8.1 Policy 2.0 applied to telecom 3.0

As to be expected, the decisions by the NMa/OPTA were challenged in court, by cable providers UPC and Ziggo, alternative network providers BBned and Tele2, as well as ISPs Scarlet and Online, a subsidiary of T-Mobile (Emerce, 2009). Notwithstanding, the adaptation of the regulation by OPTA to accommodate the deployment of fibre to the cabinet and fibre to the home remained. The principle of open-access fibre networks, initiated by Reggefiber, became part of the regulation, as KPN became a shareholder in Reggefiber. The use of an actual fibre business case as starting point for setting the wholesale tariffs was novel. Therefore, we may conclude that so far Telecom 3.0 did not require a new policy but, rather, an extension of the existing Policy 2.0 to Policy 2.5.

4.9.1 Fixed broadband

Following a stabilization in broadband uptake around 2005, the increasing use of video on the Internet has driven the competition for higher data rates. ADSL2⁺ is being replaced by VDSL and the capacity of VDSL is boosted through vectoring and bonding. While the maximum data rate on the copper twisted pair was 8/0.5 Mbit/s (download/upload) for ADSL and 24/1 Mbit/s for ADSL2, with the combination of FttC and VDSL data rates of 50/5 Mbit/s can be provided and with vectoring 80/8 Mbit/s becomes feasible. KPN started deployment of vectoring in September 2013. On the CATV network with the combination of FttC and DOCSIS-3 data rates in excess of 100 Mbit/s can be provided by reducing the number of users sharing the capacity, with further upgrades possible. (TNO, 2012)

By mid-2013 the total broadband connections stood at 6.7 million or 88% of households (total: 7.6 million).⁶² DSL connections stood at 3.2 million or 48% and CATV connections at 3.0 million or 45%, while fibre accounted for 0.46 million or 7%. This implies a decrease 6.2 percentage points for DSL compared to 3Q2011 and an increase of 3 percentage points for CATV connections. The growth is small: over the last eight quarters the percentage of additions was 4.23% and exceeds the percentage of disconnects of 3.37% by only 0.9%.

Unbundling plays an important role with 3.2 million unbundled lines (44% of total) of which 78% is fully unbundled and 22% shared.

Broadband is increasingly subscribed to as part of a triple-play bundle, up from 39% to 56% over the period from mid-2011 to mid-2013. Broadband-only subscriptions dropped from 28% to 18%. The total number of bundled subscriptions reached 5.5 million by mid-2013, or approximately 75% of households.⁶³

By mid-2013 broadband by download data rate provides the following picture:

Relative to the Digital Agenda targets for 2020, the target 100% of households having access to up to 30 Mbit/s is largely accomplished, with 7.4 million or 98% of homes connected to the cable network. With 2.5 million homes (33%) subscribed to up to 30 Mbit/s and 0.4 million (5%) over 30 up to 100 Mbit/s, achieving the subscription target of above 100 Mbit/s is still further away. Nonetheless, progress around 10% over the last two years in the category ≥ 30 Mbit/s – < 100 Mbit/s is promising. See Figure 4.1.

Figure 4.1 Broadband connections by data download rate as percentage of total, the Netherlands, 2011–2013

Data source: ACM, 2013a.

Moreover, the FttH deployment by Reggefiber is picking up pace. See Figure 4.2.⁶⁴ According to Stratix Consulting, the deployment of FttH in the Netherlands had reached the level of 1.9 million homes passed (25.1% of the total) and 626,000 homes subscribed (8.3%) by September 2013. This represents a conversion rate of around 30%. (Stratix, 2013).

Figure 4.2 FttH deployment in the Netherlands, 2006–2013

Data source: Stratix Consulting, 2013.

Voice-over-Broadband reached 4.9 million connections, or 68% of total retail fixed telephony connections, with 9% ISDN and 23% PSTN (grand total: 7.2 million – 95% of households) by mid-2013. The average disconnect rate of 3.31% over the last eight quarters is slightly lower than the connect rate of 3.56%. Between end-2006 and the end-of 2012, fixed telephony revenues dropped by 30%.

RTV subscription revenues amounted to EUR 1.62 billion for 2012, this represents 7.4 million subscriptions. The share in revenue of non-cable operators has grown from 5% in 2006 to 24% in 2012 and, in terms of subscriptions, to 35%.

The importance of bundles increased to reach 50% for TV as part of triple-play and 9% for TV as part of dual-play with broadband by mid-2013. TV-only dropped from 57% to 40% between mid-2011 and mid-2013. Telephony as part of triple-play increased from 34% to 52% over the same period. Telephony-only dropped from 46% to 33%.

TV services are increasingly delivered as IPTV and hence the number of active connections on the cable network has dropped to 4.8 million or to 64% of the 7.4 million homes connected (98% of homes)⁶⁵ by mid-2013. Digital TV by technology stands at: 55% CATV; 16% DSL; 9% FttH/B; 9% terrestrial; and 11% satellite. (Telecompaper, 2013a)⁶⁶

4.9.2 Mobile broadband

Although the main body of this case study has been dedicated to the development of fixed broadband, the current status of the mobile broadband market is included in this section to provide a complete market overview, in the light of the realization of the Digital Agenda for Europe. A short summary is provided on the development of the mobile market, with a focus on the number of players as an important parameter in determining mobile broadband dynamics. (MinEZ, 2010; Anker, Reference Anker2013; TNO, 2013)

Following the introduction of GSM by the PTT – now KPN – in 1994, the wireless cellular market became competitive in 1995, when Libertel – now Vodafone – won the beauty contest for a second GSM license.

In 1998, the auction of GSM-1800 MHz (called PCS at the time) introduced three more mobile players: Telfort – as a joint venture between the Dutch railways and BT; Dutchtone – backed by France Télécom and Deutsche Telekom (DT) and two major Dutch banks, Rabobank and ABNAMRO; and Ben – a collaboration between Belgacom and Tele Danmark. Shortly thereafter DT sold its interest in Dutchtone and BT acquired a 100 per cent interest in Telfort.

In 2000, the mobile broadband era started with the auction of five licenses for UMTS frequencies. All five operators plus Versatel participated. All existing operators managed to acquire UMTS licenses. Vodafone provided the first commercial service in 2004, with a data rate of 384 kbit/s. KPN followed in the same year. In 2006, T-Mobile launched its service including HSDPA, providing initial data rates of 1.8 Mbit/s, to be increased over time to 3.6 Mbit/s, 7.2 Mbit/s and later to 28.8 Mbit/s.

From 2000, a number of ownership changes followed but the number of operators stayed at five, each operating a network with (almost 100%) national coverage. In 2000, France Télécom acquired the UK-based Orange and rebranded its mobile telecommunications as Orange, hence Dutchtone became Orange. In 2001, BT divested its mobile activities to become mmO2 PLC with O2 as its brand name. Thus Telfort became O2. In 2002, O2 sold its Dutch interests to Greenfield Capital Partners, which re-introduced the Telfort brand. In 2003, T-Mobile had acquired all shares in Ben and the brand disappeared, to resurface again as a T‑Mobile brand for SIM-only subscriptions in 2008.

From 2005 a consolidation wave started: Telfort was acquired by KPN, which retained the brand and returned the second license to the government. In 2007, Orange NL was acquired by T‑Mobile and the Orange brand disappeared. Three main players resulted: KPN, Vodafone and T-Mobile.

In 2010 followed the auction of the 2.6 GHz band to expand mobile broadband capacity, with new entrants Tele2 and Ziggo4 – the consortium of the cable operators Ziggo and UPC. In 2012 followed the so-called multiband auction enabling 4G/LTE roll-out. It included the re-auctioning of GSM900 and GSM1800 spectrum, the new ‘digital dividend’ in the 800 MHz band and new, smaller blocks of spectrum in the 2.1 and 2.6 GHz bands; including the spectrum that Telfort had occupied. KPN, Vodafone, T-Mobile and Tele2 obtained licenses and thereby assured the continuation of their operations, while Ziggo4 did not obtain additional spectrum. KPN announced national coverage of LTE by the end of 1Q2014, T-Mobile aims at an 80⁺ per cent coverage by the end of 2014, while Vodafone has launched 4G in a few major cities, but did not communicate any further targets. Tele2 and Ziggo4 combine the use of their new spectrum with existing MVNO arrangements with T-Mobile and Vodafone, respectively, to provide national coverage. Tele2 also entered into a passive network-sharing arrangement with T-Mobile. Ziggo and UPC also push Wi-Fi as part of their mobile broadband strategy. In January 2013, KPN announced its collaboration with FON, to extend its Wi-Fi strategy based on hotspots.⁶⁷

The three main MNOs plus 52 MVNOs serve the Dutch market through 70 brands, generating EUR 5.8 billion in revenues in 2012. Data revenues (excluding SMS) have grown from just under 25%, reaching 35% by mid-2013. SIM-cards in the market reached 22.1 million in 3Q2012, with KPN (including its MVNOs) as the market leader with 47% of SIMs; Vodafone followed with 29% and T-Mobile with 24%. MVNOs accounted for 20.3% of total SIMs and 15% of revenues. ARPUs per operator ranged from EUR 21.70 to EUR 25.50. Broadband connections (combined and broadband-only) reached 52% and M2M an additional 5%. On 6 January 2014, the LTE coverage by KPN was 71% by population and 45.6% by area; Vodafone 30% and 12.5% respectively; and T-Mobile 21.7% and 6.5%, respectively, while Ziggo/UPC and Tele2 had established only the minimums required by their coverage obligations.⁶⁸

Since 2Q 2010, more calls are made on the mobile network than on the fixed network. The number of mobile calls over the period 2010–13 remained flat. As of 2H 2010 the number of SMSs started to fall, to stabilize at 50% in 2013, while the revenues dropped by close to 30%. The data volume doubled over the period 3Q 2011–2Q 2013 and the retail revenues by a factor of 2.3.

5.1.1 The development of the copper network

5.1.2 The development of the cable network

5.1.3 Case focus: the Flemish initiatives leading to the founding and survival of Telenet

In Flanders, the project ʻMultimedia in Vlaanderenʼ (Van Batselaer et al., Reference Van Batselaer, Lobet-Maris and Pierson1997) was aimed at developing an information society in Flanders². This project was presented by the Flemish government in April 1996 and based on domestic studies as well as conclusions from European and international studies. The objectives of this project were multi-fold (Van den Brande, Reference Van den Brande1996), see also Figure 5.1:

1. – conversion of the Flemish cable infrastructure into an interactive broadband network offering broadcast, telecommunication and multimedia services;

2. – creation of a stable regulatory institution and legislation based on competition;

3. – promotion of research and development in the sector of information technology;

4. – assignment of an exemplary role to the Flemish government in the set-up of pilot projects;

5. – creation of additional jobs.

Figure 5.1 Impact diagram representing the important causes for theTelenet Vlaanderen project

The objective of this project was to create new applications and to lay the foundation of an information society in Flanders, as well as to create a united cable company as a competitor to Belgacom in Flanders. The policy plan of 1996 gave some examples of future applications, some of which are now commonly used: e-learning, e-working, e-shopping, e-administration, video-on-demand, e-health and many more.

The major consequence of the project was the foundation of Telenet Vlaanderen, set up to interconnect and potentially unite the independent Flemish cable companies to achieve an interactive broadband network offering broadcast and telecommunications, as well as multimedia services to all citizens. This foundation provided a good fit within the project ʻTelenet in Vlaanderenʼ, the major sub-project of ʻMultimedia in Vlaanderenʼ.

The reasons for this project were twofold. On the one hand, the infrastructure was insufficiently used, although Flanders possessed one of the densest cable networks in the world. The legal restrictions imposed upon the cable networks (every action in the cable sector must be approved by the RTT) limited the application possibilities. At the time, there were many small cable companies, each covering their own region with their own infrastructure but lacking interconnection between the different regions and infrastructures. On the other hand, with the liberalization of the telecom market of January 1998 (authorized by the European Commission) on the horizon, a need arose for competition with the incumbent Belgacom and other possible international entrants.

A thorough feasibility study performed by the GIMV (Gewestelijke InvesteringsMaatschappij voor Vlaanderen, a European venture capitalist with public funds and experience in private equity (GIMV, 1998)), estimated the investment costs at BEF50 billion (about EUR 1.25 billion). A starting capital of BEF17 billion (about EUR 420,000) was provided by a consortium of four large partners in the Telenet Holding: MediaOne (25%), the Flemish inter-municipal corporations (35%), GIMV (20%) and a financial consortium with representatives of the Flemish media and several financial groups (20%). A major part of this total investment was invested in rolling out an extensive fibre backbone of about 635 km in order to connect the individual networks of the various cable companies, which was completed in July 1997. Through the use of switching centres and by making the amplifiers in the coaxial network bi-directional, two-way traffic became possible, thereby enabling many additional applications, as such creating a playing field for innovation.

As a consequence of the ʻMultimedia in Vlaanderenʼ project, a set of new enterprises arose from the merger of the former local cable companies. Apart from the foundation of Telenet itself, 1996 was characterized by the creation of UPC Belgium (United Pan-Europe Communications Belgium), founded as a joint venture of Philips and UIH (United International Holding), with US-based Liberty Global and UPC Broadband as major stakeholders. Its network took over the existing networks in seven communes of Brussels (Etterbeek, Ganshoren, Jette, Koekelberg, Schaerbeek, Berchem-Holy-Agathe and Forest) and three communes of Flanders (Heverlee, Kessel-Lo and Leeuwen). Another cable company, Interkabel Vlaanderen, resulted from the merger of three local ʻintercommunalesʼ: Interelectra (province of Limburg and the Antwerp city of Laakdal), PBE (region of Hageland in the province of Vlaams-Brabant), WVEM (regions of Diskmuide en Wevelgem in the province of West-Vlaanderen, region Halle-Vilvoorde in the province of Vlaams-Brabant and the cities of Beerse and Vosselaar in the province of Antwerp) and Integan (region of Antwerp and city of Essen). Interkabel Vlaanderen offered broadband services, cable television and INDI, a digital television platform.

Although the business plan for Telenet was ambitious, the first Internet access offer, branded Pandora (launched in August 1997), was too expensive (about EUR 50 per month), leading to a delay of the expected boom in uptake and very high debts. Furthermore, Telenet (Pandora) only received revenues from offering additional services; the network revenues stayed with the network owners (privately owned and geographically divided ‘intercommunales’). In 2002, the newly appointed CEO, Duco Sickinghe, changed the strategy of the company. By buying into the intercommunales (UPC Belgium in 2006³, Interkabel Vlaanderen in 2008⁴) in return for a share in the larger Telenet, he could refinance the entire operation, using the ensured revenues from the network (cable TV revenues previously destined for the intercommunales) as collateral. Through the takeover of about 850,000 customers of UPC and Interkabel, Telenet acquired the fourth and final important part of the Flemish cable network. This strategic move prevented Belgacom from entering the cable infrastructure business, and therefore made Telenet a full competitor on the infrastructure level using its HFC (Hybrid Fibre Coaxial) architecture.

Apart from the strategy change, the Belgian NRA, BIPT, had an important influence on the business case for Telenet: that is, an interconnection agreement signed with Belgacom on the application of asymmetric regulation on termination fees to the advantage of Telenet meant a large extra income for Telenet, and saved the company to some extent (as without this measure, Telenet might have gone bankrupt). Some even considered this asymmetric regulation as an indirect subsidy from Belgacom to Telenet, justified in the context of creating a level playing field in telecoms in Flanders. Although this justification may be subject to discussion, it led to the current duopoly situation where Belgacom and Telenet compete for broadband customers.

5.2.1 Belgacom in a competitive setting

Starting 1 January 1998, the Belgian telecom market was completely liberalized, whereby Belgacom was obliged to open up its network to new entrants. Local Loop Unbundling (LLU) was legislated in October 2000, allowing access to the network of the incumbent from 1 January 2001 onwards. Belgacom first issued a Reference Unbundling Offer (RUO) in December 2000, which was reviewed and rejected by the BIPT several times. The publication of the documents was found inadequate and the standard contracts contained a number of clauses that formed a serious obstruction to new entrants (as stated in the review of Belgacom’s Reference LLU Offer, 12-03-2001 (BIPT, 12/03/2001; 14/03/2001). The final BRUO-proposal (Belgacom’s RUO) was approved by the BIPT in March 2001 and included rental possibilities on three levels: full LLU, line sharing and sub-loop unbundling. Bitstream access was granted a few months later through a separate reference offer. The Belgian market became not only officially but also practically open to OLOs (Other Licensed Operators).

Because of this newly introduced competition, Belgacom had to improve its applications and services in order to maintain its strong position on the telecom market. Figure 5.2 provides an overview of the most important events that characterized Belgacom’s history. In the following paragraphs we will discuss these events.

Broadband Internet: network upgrades

In order to stay competitive, a prime requirement is to have a competitive network. After emergence of the public Internet and dial-up use in 1995, the real story of commercial broadband Internet using the PSTN network of Belgacom begins in 1998, with the execution of a pilot project for the testing of ADSL (asymmetric digital subscriber line), ʻthe high-data rate access roads to the information highwayʼ. The project included 1,000 customers in several large Belgian cities (Antwerp, Brussels, Leuven, Liège, Mechelen, Ghent and Charleroi). Belgacom predicted that, by using this new technology, transferring information at data rates up to 8 Mbit/s (download) and 600 kbit/s (upload) per user would become everyday reality in the near future.

Because of the success of the pilot project (80% of the test public approved of the possibilities of the service and agreed to recommend it), Belgacom Skynet was one of the first operators (world-wide) to introduce ADSL commercially in April 1999. Soon, ADSL covered 30 to 35% of Belgacom’s telephone customers. The intention was to achieve coverage of 70% by the end of 2000. This objective was exceeded when in November 2000, ADSL services were available to 75% of Belgian population. By the end of 2002, the coverage was almost complete when 98% of the inhabitants had access to ADSL services (Belgacom, 15/01/2003), making Belgium the leader in Europe. In 2010, DSL coverage had reached 99.85% of the Belgian population. In terms of uptake, Belgacom had 517,000 subscribers by the end of 2002, an increase of 290,000 since the end of 2001. This number doubled again in the next two years, reaching the milestone of 1 million in 2004 (Belgacom, 25/02/2004). By June 2011, Belgacom had 1,835,000 subscribers.

Not only did technical measures give rise to network upgrades and expansion of uptake; at the end of 2008, the acquisition of Scarlet, a Dutch telecom company specialized in low-cost Internet offers, increased the market share of Belgacom by 5%. Both the BIPT and the OLOs reacted quite negatively to this acquisition; however, IBPT approved the acquisition on the condition that Scarlet’s backbone was sold. Not only did this take-over increase Belgacom’s market share but it also allowed Belgacom to offer a low-cost brand (although Scarlet is fully owned by Belgacom, the name is retained) (De Tijd, 09/09/Reference Tijd2008).

In 2002, Belgacom launched SDSL (Symmetric DSL), which provides a fast Internet connection with the same maximum available data rates in both directions (up and down). SDSL was primarily developed for business customers. The need for higher data rates made Belgacom explore the opportunities of fibre deployment within the scope of the Broadway project that was launched in 2004. This project aimed at upgrading the network to a combined copper and fibre network. The goal was to connect the central offices to the street cabinets using optical fibre (the so-called Fibre-to-the-Cabinet, FttC) and to roll out a VDSL platform between the street cabinets and the end users. The VDSL technology was introduced commercially on 2 November 2004 and by investing EUR 103 million in 2006, VDSL coverage of 45% was reached by the end of that year. This VDSL technology offered data rates up to 8 Mbit/s download and 400 kbit/s upload by the beginning of 2005. Because the equipment vendor Alcatel Lucent no longer supported the particular VDSL technology (incompatible with ADSL and too much spectral noise) and because the need for higher data rates was urgent due to the introduction of digital television, Belgacom implemented ADSL2+ in 2005 as a temporary solution until the new VDSL2 standard would be ready at the end of 2007. By 2009, a total investment of about EUR 500 million had led to the deployment of 14,000 km of fibre, connecting 17,000 ROPs (Remote Optical Platforms) representing a FttC coverage of about 70%. Thanks to the early investments in the Broadway project and the high ranking of the VDSL2-coverage (second place in Europe in 2009), Belgacom received the 2009 Innovations Award from Global Telecommunications Business (Belgacom, 29/10/2004; 09/09/2009).

Because of the increasing demand for higher bandwidths and in order to stay competitive vis-à-vis the cable operator Telenet, Belgacom decided to invest further in applications using optical fibre. The first tests concerning Fibre-to-the-Home (FttH), bringing the optical fibre into the living room of the customer, were executed in Rochefort in 2008 and extended to Sint-Truiden and La Louvière in 2009 (Belgacom, 09/09/2009) but no commercial deployment was envisaged yet.

On the one hand, Belgacom was deploying VDSL2 to all its customers, having reached a national coverage of 78.9% at the end of 2011, aiming for 85% by the end of 2013. On the other hand, Belgacom recognized the limitations of VDSL2, and therefore already communicated their ʻGet to fast, fasterʼ strategy (Belgacom, 27/09/2011). In a partnership with Alcatel-Lucent, Belgacom aims at maximizing the VDSL2 throughput by using the new state-of-the-art vectoring technology. VDSL2 vectoring is a noise-cancelling technology that will allow the use of VDSL2 at its theoretical data rates, which will allow rates of 100 Mbit/s and beyond to be transmitted on copper cables. Belgacom opted for this upgrade because it will bring high data rate broadband to the end-consumer in a fast and cost effective way. Tests of VDSL2 vectoring started at the end of 2012.

Apart from vectoring, other technological upgrades on VDSL are possible to boost bandwidth capabilities. In VDSL bonding, two physical twisted pairs to each customer are used instead of one, which almost doubles the data rate. However, as there are few spare copper pairs in Belgium, it is unlikely that this upgrade can be implemented for every customer. Phantoming adds a third – virtual – twisted pair, which would bring data rates up to 200 Mbit/s to each individual household⁵. Combining all options (bonding, phantoming and vectoring) would allow the DSL network to offer data rates of about 300 Mbit/s (Alcatel-Lucent, 2010, 2011).

The graph in Figure 5.3 provides an overview of the most important changes concerning the available data rates for residential users.

Figure 5.3 Evolution of the data rates offered by Belgacom, Dec. 1999–Aug. 2013 (for residential customers)

Data sources: Belgacom, 10/09/2003, 3/05/2004, 11/01/2005, 09/01/2008, 17/11/2008, 28/05/2010, 27/08/2010.

5.2.2 Telenet: an infrastructure-based competitor with similar offers

Starting in 1998, a duopoly of Belgacom and Telenet now dominates the telecommunications market in Flanders. Both companies upgraded their networks and services gradually, in order to counter the competitive pressure. This paragraph focuses on the evolution of Telenet, the Flemish cable operator. Comparison to the corresponding services offered by Belgacom is made where possible. An overview of the applications and services offered by Telenet, as well as a summary of the most important upgrades is shown in Figure 5.4.

Figure 5.4 Timeline showing the most important events in the history of Telenet

Broadband Internet: network upgrades

Figure 5.5 shows the evolution of the data rates offered by Telenet. To provide its services, Telenet uses a HFC network, which consists of fibre backbones and coaxial access in the areas served. To provide broadband access, Telenet uses the DOCSIS (Data over Cable Service Interface Specification) technology. The first version of this technology (DOCSIS 1.0) was developed in 1997 by CableLabs for the US market. To match the European cable frequency spectrum, the standard was adapted and the Euro-DOCSIS technology was born. The first upgrade (DOCSIS 1.1) was launched in 2001. Now, Telenet uses the European version of DOCSIS 3.0, which allows them to offer a download data rate of 100 Mbit/s and more (Boonefaes, Reference Boonefaes2005–2006; CableLabs, 2011).

Figure 5.5 Evolution of the data rates offered by Telenet, Dec. 1999-Aug. 2013 (for residential customers)

Data sources: Telenet, 14/12/2005, 20/06/2007, 21/05/2008, 08/02/2010, 7/06/2012.

Over the years, Telenet launched various new products, each time increasing the data rates. The first telecom product was launched in 1998 under its ‘Home’, ‘Do’ and ‘Young’ packages. Telenet Internet XL came to the market in September 2000, providing a downstream rate of 1 Mbit/s and an upstream rate of 256 kbit/s, thereby doubling their previous offer. In 2002, Telenet Internet XL offered download data rates up to 4 Mbit/s, while Euro-DOCSIS 2.0, introduced in 2003 as part of Telenet ExpressNet, enabled an upgrade to 10 Mbit/s. Download data rates increased to 20 Mbit/s in the fall of 2005. In 2010 followed the introduction of DOCSIS 3.0 with a download rate of 100 Mbit/s and an upload of 4 Mbit/s.

In March 2010, Telenet launched ‘Digital Wave 2015’, a project aimed at improving the existing network over the next five years, with an investment of about EUR 30 million every year. Through the roll-out of an extensive fibre network, Telenet wants to promote the transformation of Flanders into a digital and networked economy. Along with the transformation of the network itself, Telenet aims to develop a number of new applications, such as remote medical services, call centre video assistance over the Internet, 3D television, video on demand, mobile television, new file-sharing and synchronization and e-government (Telenet, 3/03/2010). Telenet keeps upgrading its network by gradually reducing the size of the service areas (SAs). An SA is the part of the network that connects the end-customers to the first local collection point through the use of coaxial cables. From the collection point onward only fibre cables are used.

Since the available data rates offered by Telenet nowadays are much higher than those of Belgacom, the motivation to upgrade towards FttH is weaker. Telenet, however, did intend to keep up the current level of investment in their network, within the scope of the Digital Wave 2015 project. They are currently expanding their DOCSIS 3.0 technology (using channel bonding to increase data rates to over 100 Mbit/s) to all customers, and keep on increasing the available bandwidth by reducing the size of their service areas. These projects allowed Telenet to increase the data rates again in June 2012, to a maximum download rate of 120 Mbit/s and upload of 5 Mbit/s for residential customers (Telenet, 7/06/2012). Plans for deploying FttH have not been announced so far.

5.3.1 Marketing strategies: focus on own strengths

From the previous section it is clear that, although there are some smaller niche players present, the fixed broadband market in Flanders is dominated by Belgacom and Telenet, each having a close-to-100% coverage with their networks (DSL and DOCSIS, respectively). However, since the available data rates on those networks differ significantly (maximum download data rates for residential users are 30 Mbit/s for Belgacom versus 120 Mbit/s for Telenet), both players market their offers using different key services and promotions.

Since mid-2012 Telenet focuses its marketing strategy on data rates and simplicity, for both the fixed and mobile markets. In July 2012, Telenet launched King and Kong, two straightforward tariff schemes for its mobile customers. Both are bundles including voice, SMS and mobile data for a fixed amount per month. To attract or convince fixed Telenet customers, discounts are given. In June 2013 Telenet launched a similar offer for its fixed services: Whop and Whoppa bundles including digital television, VoIP telephony and fixed and Wi-Fi Internet through their Wi-Free hotspots. Their advertising emphasizes the high data rates reached (Whoppa: ʻfor whom fast isn’t fast enoughʼ) and the inclusion of all services into one simple and transparent bundle.

Belgacom, as a quadruple-play operator, goes one step further in bundling its services. Because Proximus is a subsidiary of Belgacom, they are able to offer both mobile and fixed Internet in one package: the ʻGeneration packʼ. Various options exist, for different download limits, voice and SMS usages, etc. Telenet could follow the same strategy by making use of its full MVNO contract but has not taken this path so far.

Another field of competition is the broadcasting of football matches. Although Belgacom long had the monopoly on the Jupiler Pro League Football (the main football competition in Belgium), this right is not exclusive anymore. Both Telenet and Belgacom now try to attract customers with their football channels, Telenet even promising 2,222 live goals on its Sporting Telenet channel for the upcoming season, or a refund of all subscription fees.

5.3.2 Regulatory setting: unbundling obligation and the new telecom law

Unbundling obligation

1998 was the year of the liberalization of the European telecom market. The incumbent Belgacom was obliged to open up its network to new entrants. Belgacom remained owner of the network but had to provide network access to other operators. The consequence of this liberalization was the rise of many new OLOs (Other Licensed Operators) in the following years (e.g., British Telecom, MCI, Colt, Versatel, Coditel, Tele2, Dommel, EDPnet, Mobistar, Eleven, Scarlet). This fragmentation of the market in the field of the copper network stood in great contrast to the intense concentration of the market in the cable network. This is an important observation: while Belgacom was obliged to open up its network, the cable companies were asked to form a united cable network.

However, the fragmentation of the market remained limited, as the new entrants had to combine the significant investments attached to the start-up of a new company with the need to offer low prices to attract customers. New entrants had to offer cheaper and/or better products than the incumbent, because they had to overcome customer loyalty for the existing brands. More firms in the market made the competitors play each other off by using aggressive pricing to attract more customers. This on-going price pressure made it hard for the new entrants to survive, forcing some to end their activities (e.g., Eleven) or to sell their activities to another operator (e.g., Versatel, Tele2, Scarlet).

Other operators managed to develop a customer base of significant size. An excellent example here is the operator Schedom, which provides a cheap Internet connection for urban subscribers under the brand name Dommel. Its success is implicit in the selection of the market segment. By focusing only on urban, densely populated areas (including many students and gamers), Schedom can keep its line prices low (Dommel, 2011). Other operators, like Scarlet, were taken over by Belgacom, although Scarlet remains an autonomous subsidiary (Scarlet, 2011).

The impact of the unbundling obligation remained marginal until the BIPT lowered LLU prices in 2006.⁷ The scale of the alternative operators was by then large enough to climb further on the ladder of investment and to invest in unbundling. These investments were stopped in 2008 when the economic crisis hit and when Belgacom announced that it intended to close down 10% of its MDFs (Main Distribution Frames) as a consequence of the move to an all-IP infrastructure and VDSL deployments in the sub-loop. Because the planned closure impacted 40% of the unbundled lines, the BIPT intervened with additional obligations⁸ to guarantee a fair return on investments for the alternative operators, but the investments in unbundling never recovered. See also Figure 5.7. Carrier resale has gained in volume due to take-overs (Scarlet by Belgacom, LLU network Base by Mobistar) and to the move from ADSL2+ to VDSL2 (no subloop unbundling).

Figure 5.7 The use of wholesale access, Belgium, 2H 2003 – 1H 2013

Data source: BIPT, 2013a.

Although Telenet long enjoyed a monopoly on the cable network and services, the CRC (Conference of Media Regulators) and the Belgian NRA (BIPT) published a proposal for opening up the cable network in 2010. This proposal was transformed into a formal decision on 1 July 2011, in which the authorities regulated Telenet and the other cable operators and created an obligation to provide wholesale access to analogue television and broadband Internet services, as well as opening up the digital television platform. Telenet responded stating that regulating analogue TV is not useful because of the declining number of subscribers and that regulating digital TV is not necessary, as in that market there is enough competition from different platforms (cable, IPTV, satellite, DTT and Internet TV). They furthermore argued that regulating the most important competitor to the SMP incumbent would definitely not enhance the infrastructure-based competition the European Commission favours (Telenet, 18/07/2011, 21/06/2011, 21/12/2010, 26/05/2011). Notwithstanding the arguments Telenet put forward, they were all rejected by the regulators. The regulators and operators are now defining the implementation of the decisions; a final decision from the Court is expected for end 2013 (Belgacom, 2013a). The CRC has already approved the qualitative elements of the decision, calling for Telenet to provide the access within a six-month period after the official request (either through impress payment or signed letter of intent) of the alternative operator (CRC, 2013). However, a final outcome of the court case is not expected before mid-2014 (Telenet, 2013).

Concerning the current LLU and bitstream access regulation of Belgacom’s network, the BIPT is also re-evaluating its models for calculating the price caps, following the European guideline to stabilize the prices for copper lines between EUR 8 and EUR 10 per month, as such increasing the flexibility for the deployment of fibre-based networks. As mentioned above, these caps were set in 2010, and remained unchanged until 2012.

The new telecom law (2012)

In July 2012, a new telecom law was voted in Parliament, focusing on transparency, competition promotion and (universal) service obligation. The law requires customer protection by obliging the operator to be transparent in their communication with their clients and by allowing customers to change operators for free. Every contract can be modified or terminated by the client after a period of six months, without needing to specify a (legal) reason. Finally, the law includes a quality of service obligation, whereby the BIPT can enforce a minimum quality level when certain requirements are not being met. With this law, the Belgian telecom market should comply with, if not exceed, the requirements set by the European Commission.

The effect of the telecom law became visible quickly after it was put into force at the beginning of October 2012. The number of ported mobile numbers rose from 71,000 in September to 160,000 in October of that same year (see Figure 5.8). Although the effect slowed down later in the year (136,000 in November, 151,000 in December), it remains significant and demonstrates the impact of telecom regulation in Belgium. This strong increase in ported mobile numbers can be fully attributed to the new telecom law, but certainly also finds a cause in the launch of the competitive King and Kong subscriptions by Telenet, strategically positioned two months before the implementation of the telecom law.

Figure 5.8 Ported mobile numbers per month, Belgium, 2012–2013

Data source: BIPT, 2013a.

The effect on the fixed market was lower, but still significant, with a net increase of about 9,000 ported numbers (an average of 29,038 ported numbers per month before versus an average of 38,131 after the implementation date).

5.3.3 External influences: blocking of 4G by Apple

Besides the regulatory influences, there are some other external influences affecting the telecom market in Flanders and Belgium, including the competitive interplay between the market parties. One salient example of this kind of influence is the impact of Apple blocking 4G access on iPhones. Although the recently launched iPhone 5 supports 4G functionality, it has not been activated in Belgium. Belgacom communicated the following: ʻ4G functionality is currently not activated on these devices making it impossible to use the Proximus 4G network. Apple will decide when 4G will be available for their devices on our network. This also applies to the 4G network of other operators.ʼ (Belgacom, 2013b).

In January 2013 Apple updated the iOS operating system, allowing iPads to surf on the 4G-network but ignored the iPhone. Proximus, the mobile brand of Belgacom, was not aware, nor notified of Apple’s plan and believes it to be a pure commercial decision (Stevens, Reference Stevens2013). Mobistar, on the other hand, is listed as a provider allowed to offer 4G on the iPhone, but they have no 4G network operational yet.

6.1.1 Historical framing and overview of developments

In contrast to most other monopoly telecom markets, prior to liberalization there was not one national operator in Denmark. The Danish telecom market was indeed a monopoly market, but the monopolies were regional. In the late phase of the monopoly era, there were four regional operators, and it could be claimed that there was a kind of yardstick competition between these operators. In combination with a relatively high GDP per capita, this contributed to a well-developed penetration of fixed-line telephony in Denmark compared to most other countries.

In 1990, a holding company named Tele Danmark was formed in order to unite the regional operators and in 1995 the regional structure was entirely abandoned. Before the establishment of Tele Danmark, the fixed-line penetration rate had reached approximately 57 per 100 inhabitants, and when fixed-line telecommunications was liberalized in Denmark in 1996, the penetration rate had increased to approximately 62 per 100. Up until 2001, the penetration rate of fixed-line telephony kept on increasing, reaching its peak at approx. 72 per 100 – and has decreased ever since, being no more than around 25 per 100 in 2012; these figures include ISDN. PSTN-based telephony actually started decreasing even before 2000. Regarding broadband, the point is that the fixed-line telephony access infrastructure in Denmark is very well developed, which is important for the strength of DSL.

Tele Danmark (to be renamed TDC in 2000) was established with the explicit purpose of strengthening the Danish position in what was seen as a soon-to-be international telecom market. The liberalization trends, developing in the early to mid-1980s and receiving a strong platform at the European level with the 1987 Green Paper (CEC, 1987), were to a large extent seen – maybe not as a threat but at least as a serious challenge by the majority of politicians and the telecom administrations in Denmark. It was thus not the liberalization of the Danish telecom market which was the primary goal of Danish telecom policies at the beginning of the 1990s but, rather, the defense and strengthening of a Danish ‘national champion’. Hence, in the European context, Denmark was not among the front runners with respect to telecom liberalization at the beginning of the 1990s. However, this changed in the mid-1990s, as it became clear to the Danish government that liberalization would eventually be the result of international developments. In 1995, the Danish government issued a policy strategy document entitled Bedst og billigst hennem reel konkurrence (‘Best and cheapest by way of real competition’) (Forskningsministeriet, 1995), which inaugurated a swift change in Danish telecom policies. First, value-added fixed-line services were liberalized, and in 1996, eighteen months before the EU’s 1998 deadline, the provision of fixed-line infrastructure and fixed-line telephony services were also liberalized.

Ahead of this development, customer premises equipment (CPE) had been liberalized in 1990, and two companies had been licensed in 1991 to operate 2G mobile networks in the 900 MHz band. One of these was the Danish incumbent Tele Danmark and the other was a new Danish company called Sonofon, which was later taken over by the Norwegian incumbent Telenor. In 1997 licenses for GSM 1800 were awarded, which resulted in four mobile network providers operating GSM networks (900 and 1800 MHz): Tele Danmark, Sonofon, Telia (the Swedish incumbent), and Mobilix (later Orange). Further in the process, in 2001, a new mobile infrastructure provider, 3 (Hutchison Whampoa), entered Denmark in connection with the 3G licenses. Upon Telia taking over Orange in 2004, there are four mobile infrastructure providers (TDC, Telia, Telenor, and 3) in Denmark at present, of which two of them (Telia and Telenor) have started sharing their mobile infrastructures. All four operators offer 3G as well as LTE (‘4G’), allowing for mobile broadband.

In 1997, the Danish state sold its shares in Tele Danmark, resulting in the American operator Ameritech (later SBC, following the merger between Ameritech and SBC) becoming the dominant shareholder in Tele Danmark. This came as a real surprise to the Danish public, as hitherto the main thrust in the international liberalization of telecommunications had been the creation of a strong Danish operator; and what would then be the result of a US company taking over? In reality, it did not result in any major changes and was, to a large extent, an expression of different confluent developments:

1. (1) the Danish telecom market had quickly become one of the most liberal markets in Europe;

2. (2) the Danish market was one of the most advanced telecom markets, technologically;

3. (3) the Danish market being advanced with respect to penetration, it was seen as a kind of test market; and

4. (4) last but not least, Ameritech paid a vast sum of money for the Tele Danmark shares, which alleviated the financial situation of the Danish state.

The Ministry of Finance was the main driver behind this development.

With the Danish state selling its shares in Tele Danmark, the incumbent operator became totally privatized. The state did not even maintain a ‘golden share’ or any other special right of intervention. And further down the trajectory, the vast majority of shares in the company were sold to a group of private equity funds. This also created a degree of unrest: it was one thing to sell the company to another telecom company with strategic interests in the telecom field but quite another to sell the company to equity funds without any special interest in telecom and whose primary raison d’être is to extract as much money as possible from their acquisitions.

As in most other European countries, until the mid-1990s the Internet was used almost only in universities – and only by a small minority at those universities. Tele Danmark gave priority to trials and marketing of videotex solutions, and it was not until 1995 that the final attempt with videotex was given up in Denmark and Tele Danmark started focusing on the Internet. ADSL was introduced in Denmark in 1999. The first to offer ADSL was a small Danish company, Cybercity (later to be bought by Telenor), and Tele Danmark followed quickly thereafter. Cable modems were also introduced in Denmark in 1999, first by Stofa (which at that time had been taken over by Telia) and shortly thereafter by Tele Danmark. The uptake especially of ADSL but also of cable modems grew very quickly, and broadband penetration in Denmark has for many years been among the highest in the OECD countries. DSL and cable modem solutions constitute the mainstay of fixed broadband access in Denmark.

In addition to DSL and cable modem solutions, fibre in the access networks also started to develop. However, fibre penetration is not near the level attained in either Japan or Korea, and also Sweden and Norway are somewhat ahead of Denmark. But fibre has started growing, mainly promoted by energy companies. In the past few years, TDC has also given more attention to the fibre access market by acquiring the fibre operation of DONG, the largest energy provider in Denmark, in 2009 and the broadband access company ComX in 2013. The energy companies have been the promoters of fibre in the access networks. One of the energy companies also offered a WiMAX solution for a number of years but had to close it down in 2012 because of lack of profit and future development possibilities. In 2012, another energy company, Syd Energi (SE), made a bold move and bought Stofa, making SE the second largest cable modem access provider in Denmark after TDC.

The result is that the broadband access market in Denmark is based on DSL and cable solutions as well as fibre access – in addition to the very fast-growing mobile access market. TDC is the dominant player in the DSL market. The cable access market has two main players, TDC and Stofa, with TDC being the dominant one. In the fibre area, the energy companies with their cooperation in the company Waoo! are the largest players, followed by TDC. And in the mobile market, the dominant players are the four mobile infrastructure providers: TDC, Telia, Telenor, and 3, with TDC being the largest one. In addition to these players, there are a large number of service providers of various kinds. In Denmark, companies do not need a license to operate in the telecom field, except for mobile infrastructure providers. New service-based companies are being established constantly and can grow to a considerable size in terms of subscriptions, especially in the mobile area. However, almost as quickly as these companies grow in the market, they are taken over by the main infrastructure providers. The business models of some of the newcomers actually aim at being taken over by the larger players.

As can be seen from this brief overview of operators, the Danish incumbent TDC is present on all platforms. In most other European countries with cable TV networks, the incumbent telecom operators can have no cable operations – in the interests of promoting infrastructure competition. In Denmark, however, this has not happened, and the result is that TDC is the all-dominant player in the Danish broadband market.

In this contribution, this issue is further discussed, as are the possible implications of equity funds having owned the incumbent operator TDC. Furthermore, the implications of the use of different kinds of infrastructures are examined. Moreover, the paper aims to shed light on the importance of the interplay between service development and infrastructure development. For the period leading up to the introduction of broadband via ADSL in Denmark in 1999, the most important developments seem to be the establishment of a national incumbent in 1991, the full and relatively early liberalization of all telecommunications in 1996, and the great emphasis that has been put on service developments from the policy side.

6.1.2 Structure of the chapter

First, development in the policy framework is described in Section 6.2. This is followed by an introduction to the key actors in the telecom area in Section 6.3. In Section 6.4, investments are examined. In Section 6.5, the coverage issue is addressed. Section 6.6 analyses how competition has evolved, and Sections 6.7 and 6.8 look at how the market has performed from a demand side perspective. Section 6.7 thereby looks at the uptake of broadband, while Section 6.8 goes into other aspects of the EU Digital Agenda targets regarding the use of services by private citizens as well as private companies. Finally, the main actors in the development of broadband – operators, political and administrative institutions – are discussed in Section 6.9, and conclusions and reflections are put forward in Sections 6.10 and 6.11.

6.2.1 The first phase – reluctant liberalization and focus on industrial policy

The first phase began in the 1980s resulting in the liberalization of customer premises equipment (CPE) in 1990 and the licensing of two competing mobile operators in 1991. Telephone subscribers had already begun to buy their own phones in super markets and attach them to the network in the 1980s, and at that time mobile telephony was not seen as a central revenue generator for telecom operators. In this early phase, and in the beginning of the 1990s, the Danish government was reluctant to infringe upon the powers and privileges of the incumbent (formed by merging the former regional monopoly operators and a state-owned international operator in 1990), and the liberalization process was mainly driven by the requirements from the European Commission.

In addition to the reluctant adaptation to the new upcoming liberalization, there was also a recognition that other types of communications were on their way, requiring more bandwidth than existing kinds of communications. An initiative to invest in broadband facilities was taken in this phase – in 1985, when it was decided to build a so-called hybrid network. This network had a dual purpose: (1) to provide private households, connected to a cable TV network, with TV and radio channels; and (2) to provide private companies and public institutions with high-speed data channels (mainly for the purpose of two-way video-transmission). Four out of twelve fibres in the system were dedicated to radio and TV broadcast. Another four fibres were allocated for the high-speed network, and the remaining four fibres were reserved for future use. The project was initiated as a public investment but was foreseen to be self-financing within six to seven years. In the first planning phase, an estimate by the Ministry of Public Works was that the total investment would be approximately DKK 4 billion (equalling approx. DKK 8.5 billion in 2013) including telecom company investments and investments by cable antenna communities (Østergaard, Reference Østergaard1986). The real investments were only a fragment of this, so the ‘grand’ project was never realized to the originally anticipated extent.

The idea with this shared infrastructure was to pave the way for a fully integrated broadband network by sharing the costs with cable TV subscribers. In addition, the project was seen as a support for Danish industry. In the initial phase, the network was based on the Danish DOCAT (Digital Optical Community Antenna Trunk) standard, and the large Danish producer of cables, NKT, lobbied strongly for the project. Thus, the hybrid network would create more demand for Danish-produced telecom equipment and, at the same time, provide Danish industry with access to new advanced communication services.

However, the number of companies using the hybrid network as a communication facility remained extremely low and in reality the network became a cable-TV network. In 1995, the network was closed down as a separate network and integrated into the Tele Danmark (later TDC) networks. Today, the most important remaining impact of this hybrid network is that it allowed Tele Danmark to engage in the provision of cable TV services – a market they were denied by regulation until then.

Another important initiative in this period was the formation of Tele Danmark. Up to 1990, the Danish telecom market was served by four regional operators, all with dominant public ownership. Long distance and international communication was under the auspices of the department for Post and Telecom. The reason for this merger was to prevent the emerging competition among the regional operators and to create one strong Danish player in the international telecom market.

While the hybrid network and the creation of Tele Danmark both strengthened the government-controlled monopoly, a more liberal approach was taken in the area of mobile communications. Two operators were licensed to operate mobile networks in Denmark in 1991. The incumbent received a license, and the other was assigned to a new company named Sonofon, which was later taken over by Telenor. Sonofon was established as a joint venture between the Danish telecom company Great Northern Telegraph Company and the American Bell South. The reason for the mobile exception to the rule was that mobile, at that point of time, was not considered an essential part of the telecom area, as fixed line telephony constituted the cornerstone of the telecom business. Furthermore, the initiative could strengthen the Danish position before the expected liberalization of the telecom markets.

6.2.2 Second phase – promotion of real competition

While Denmark in the late 1980s and the first part of the 1990s ‘dragged its feet’ with respect to reforming its telecom market, the government, led by the social democrats at that time, saw ‘the writing on the wall’ with information society plans and suggestions for reforming the telecom area coming out of the US and from the European Commission. The decision was made to take a proactive stance on the potential of information and communication technologies and to reform the telecom market. This was in a sense an innovation of the Danish policies in the area – a radical innovation – which was subsequently followed by a continuous stream of incremental additions.

This fundamental change started with the creation of new ministry of Research and Telecommunication and the publication of ‘Informationssamfundet år 2000’ (Forskningsministeriet, 1994). The report aimed at presenting a vision/project for how ‘modern information technology will bind together public institutions and businesses and constitute a proposition to the citizens’ (Forskningsministeriet, 1994). One year later, the Ministry of Research issued a report entitled ‘Bedst og billigst gennem reel konkurrence’ (‘Best and cheapest by way of real competition’) (Forskningsministeriet, 1995), which came to function as a kind of blueprint for the process of liberalizing the Danish telecom market.

In spite of various qualifications made since then, this overall policy strategy is still valid. Competition on the Danish market was to be promoted through an early liberalization combined with a rather strict competition regulation. The idea was that Denmark should be among the first countries in Europe to liberalize its telecom market and, thereby, attract foreign telecom investments.

Therefore, the Danish parliament made a decision aiming to complete the liberalization process for services and infrastructure eighteen months ahead of the 1998 target set for liberalization of the European telecom markets. Since then and until recently, it has been a priority to stimulate competition and Danish telecom legislation has been ahead of the requirements set by European framework regulations in most areas.

In this period, France Telecom chose Denmark as the destination for their first major foreign investment in Europe. They established a new mobile operator, Mobilix, and contributed to making the Danish mobile market one of the most innovative and competitive in the world.

Whereas there was infrastructure competition in the mobile area, the emphasis was on service based competition in the fixed network market, and a key issue was to ensure open access and low interconnection charges in order to enable new entrants to compete with the incumbent operator, Tele Danmark. The Interconnection Act was passed, stipulating inter alia that operators with a significant share of the market (that is to say more than 25%) were to grant interconnection to other operators at cost-based prices and on terms that were objective, transparent, and non-discriminatory.

Another initiative taken in order to promote competition was the unbundling of the local loop. This was introduced in 1998, when Mobilix wanted to introduce ADSL before this service was available from Tele Danmark.

The pricing principles for unbundling introduced in 1996 were based on historical costs. The principle was to allow the inclusion of the total extra costs related to the provision of the service plus a reasonable margin. However, if the dominant operator (the incumbent carrier) had a market share of more than 80%, only 30% of the operating costs should be included. This implied that the incumbent operator should bear a part of the costs until the new entrants had obtained a reasonable market share. In this way, the incumbent was required to subsidize its competitors until some of them had established themselves on the Danish market. This approach was compatible with the investment ladder strategy later described by Martin Cave (Reference Cave2006).

To ensure low interconnection rates, the historical cost approach was supplemented by a best-practice clause in 1998, enabling the national telecom regulator to reduce interconnection charges to the international level for best practice even if Tele Danmark was able to document that the actual costs were higher. The definition of best practice was subsequently changed after several debates between the telecom agency, Tele Danmark and the new entrants. ʻBest practiceʼ came to be defined as the average of the interconnection rates in the three countries with the lowest interconnection rates. It was also possible for the regulatory agency to reduce rates if they were lower in just one country. However, in this case, corrections for country-specific conditions should be made in advance. By definition, the best practice clause kept the Danish interconnection rates among the lowest in Europe and the charges were reduced to one-third within five years (Falch, Reference Falch2002).

Stimulating competition was not the only policy created to facilitate infrastructure development. In order to ensure coverage in rural as well urban areas, a universal service obligation was adopted in 1996 and Tele Danmark, as the incumbent operator, was appointed as the universal operator in all markets. However, Denmark went further than demanded by the EU legislation, as Tele Danmark had an obligation to provide ISDN and leased lines in addition to provision of ordinary phone services.

6.2.3 The third phase – technology-neutral regulation

In 1997 the EU issued a Green Paper on the convergence of communication technologies (CEC, 1997). The markets were increasingly developing from vertical to horizontal structures, and different infrastructures were starting to deliver essentially the same services. One of the implications was a greater emphasis on promoting different infrastructures in order to create ‘real’ – meaning infrastructure-based – competition instead of just service competition. In Denmark, this found an expression in the political aim of ‘several pipes to the home’ announced in a political framework agreement signed in September 1999. The government added to the ‘best and cheapest’ policy the goal of promoting public access to the network society. Fostering competition was seen as the principal means to achieve this goal, through initiatives aimed at stimulating the creation of competing access routes for consumers.

This policy has been termed by the government as technology-neutral and market-oriented. Such a characterization is obviously self-congratulatory to some extent, as the overall policy and regulatory framework favors some kinds of infrastructure developments over others and specifically favours certain technology developments in the wireless broadband area. However, the characterization is correct in the sense that public money is not allocated for the development of specific types of network infrastructures.

Following the political agreement of September 1999, much of the existing legislation for the telecom sector was consolidated into the Act on Competitive Conditions and Consumer Interests in the Telecommunications Market in July 2000. By the adoption of the new legislation, it was decided that interconnection rates in the future should be based on LRAIC (long run average incremental costs). The regulatory agency, Telestyrelsen, therefore initiated the construction of a cost model for the Danish telecom networks, building on the LRAIC concept. Interconnection charges, set on basis of the LRAIC model, took effect from 1 January 2003. Ordinary interconnect charges were reduced by 17–36%, but the LRAIC price calculated for raw copper was 24% above the current price at the time (Falch, Reference Falch2002). Therefore, it was decided to implement this price increase on raw copper over a period of seven years. As a result, interconnection charges have increased from among the lowest in Europe to a little above the EU average. See Table 6.1.

Table 6.1 Average total cost per fully unbundled loop (€ per month), Denmark and EU, 2002, 2009 and 2011

Data sources: EU: Telecommunications Regulatory Package – VIII Implementation report – Annex I – Corrigendum March 2003; Telecommunications Regulatory Package – XII Implementation report – Annex I, 2010.

It can be argued that the LRAIC approach for calculating the interconnection rates is objective and scientific. Still, there is plenty of room for discretion in the price-setting. In Denmark, the LRAIC model was elaborated in collaboration with TDC as well as a group of new entrants. The new-entrant group was dominated by mobile operators and their primary interest was to ensure low rates for fixed termination of mobile calls. Furthermore, an increase in prices for raw copper was in accordance with the ‘several pipes to the home’ strategy, as it would presumably promote infrastructure investments by alternative operators. It would, therefore, be in line with an investment ladder strategy, as increasing interconnection charges were supposed to incentivize new operators to invest in new infrastructure instead of relying on old infrastructure owned by the incumbent.

At present, the price regulation of wholesale broadband services has become less strict than at the turn of the century. However, the regulation in other areas regarding open access is broader in its scope than in most other European countries. The telecom legislation does not distinguish between different types of fixed communication networks. This implies that the same regulation applies to copper wire, coaxial cable, and optical fibre networks and an SMP operator can be required to open any network regardless of the technical platform used.

The inclusion of cable and optical networks is relatively new. In December 2009 the regulatory agency required TDC to open their cable network, enabling other operators to use this infrastructure for the provision of broadband services. The implementation of this, however, has caused problems. An LRAIC model has been developed, but no competing service providers are currently able to offer broadband on TDC’s cable network because of the technical frequency allocation of the network.

Since June 2011 TDC has also been required by the regulator to open its fibre-based broadband access network to other operators and service providers. This applies not only to the current subscribers of TDC, but also to homes passed. Potential broadband subscribers can demand to be connected to TDCs optical network if the network passes their home in a distance of less than 30 m. The costs of connection, which are estimated to be around EUR 2,000, must be paid by TDC – even if the customer wants to subscribe to a competing operator.

One can argue that the Danish telecom regulation is in line with the technology-neutral principle recommended by the European Commission. However, the Commission does not recommend including fibre and cable networks in the definition of the market for broadband access (market 5). Denmark is, indeed, not the only country that applies a definition that deviates from the EU recommendation on this point. Cable access is included in market 5 in eleven other EU countries (Valcke et al., Reference Valcke, Hou and Stevens2011). However, Denmark is the only country, where this has been followed by a request to an operator to provide open access to cable as well as to fibre networks. The best explanation for this state of affairs is that TDC has a stake in, if not dominates, all fixed infrastructures.

6.2.4 A fourth phase of regulation?

Lately, the contours of a new phase seem to be appearing internationally. This phase puts less emphasis on stimulating competition and more emphasis on infrastructure construction. The reason is that it has turned out to be difficult to stimulate infrastructure competition if it did not already exists on the basis of a legacy cable network, and that the companies investing in new infrastructures (fully or partly optical networks) complain that there is no incentive to invest if competitors can obtain access to the new networks on equal terms. TDC has also argued along these lines.

In 2010, a new policy initiative was presented in Denmark. This initiative was launched against the background of the general international surge of broadband plans in connection with the financial and economic crisis starting in 2008. In Denmark, a High Speed Committee was established to examine whether initiatives were needed to support broadband development. The results of the deliberations of the committee illustrate the general political priorities in Danish ICT and broadband policies with a focus on the demand side. There were no suggestions for using public money for broadband development.

However, the irony of the matter was that, even though this was fully in line with longstanding Danish ICT and broadband policies, the minister responsible for the ICT area at that point apparently felt that something more had to be done. Therefore, a policy statement was issued saying that by 2020 ‘100 Mbit/s or more for all households and enterprises’ would need to be available. There was no economic backing for this goal. It was merely an expression of intent for the market players to realize. The only tangible political measure to implement broadband was concerned with requirements regarding coverage by operators in connection with license auctions for mobile broadband frequencies. In addition, the minister announced a constant ‘political attention’ to broadband development – which is rather intangible but nevertheless puts pressure on operators to make available higher broadband data rates to everyone.

In 2011 a new government came into office. One of their first initiatives was to dissolve the national IT and telecom regulator, the IT and Telecom Agency. This move might not be as drastic as it seems at first glance, as most of the activities have been continued in other departments, but public attention on ICT as a specific policy area was reduced. In their recently published plan for broadband and mobile infrastructures (Erhvervsstyrelsen, 2013), the policy initiatives are in line with the initiatives suggested by the High Speed Committee but with slightly more active participation by the public sector. For the first time, the government will be engaged in direct economic support to public infrastructure developments. It granted EUR 6 million to a specific region (Bornholm), and encouraged public-private partnerships in other regions. Even though the amount is very small, it marks a change from the former policy. And more initiatives are demanded by various actors, including fibre operators, Telia and Telenor, organizations of business and residential users, and municipalities. In an address in October 2013 to the minister of Business and Growth, these organizations asked for re-opening of the political agreement from 1999. Their precise aims were not clear but appear to demand a new approach to Danish ICT policies.

6.2.5 Demand-stimulating policies

While regulatory policies have undergone continuous incremental changes since the early liberalization, the policy on the demand side has remained basically unchanged since ‘The information society year 2000’ report of 1994. It is based on the understanding that infrastructure and service development initiatives are mutually interdependent. However, the Danish policy has, until now, been to intervene in this interdependent relationship by way of promoting the service side.

This includes the promotion of G2G e-government: i.e., the use of ICTs internally in public administrations and between public administrations in different areas and at different levels. But, more importantly for broadband development, it also includes the promotion of electronic communications between public administrations and citizens and businesses. An example of such initiatives is that public administrations no longer accept paper based invoices from companies with which they do business. Also, all-electronic tax reporting between the tax authorities and citizens and companies has become the norm as of 2012, although other means of communicating with the authorities remain possible for people who do not use the Internet, such as the elderly. Also from 2012, all students applying to higher educational institutions must do this electronically. Another example is the cooperation of a public-private-partnership between the state and financial institutions regarding identity management in electronic communications: the same identity management system is used in electronic communications between citizens and public administrations and for electronic banking. The implications of such initiatives are that there is increased pressure on citizens and companies to communicate electronically and, therefore, to use broadband connections.

A characteristic example in this context was the broadband policy initiative presented by the High Speed Committee in 2010 (Højhastighedskomiteen, 2010). Seven areas of policy initiatives were given priority. The first two were digitalization of the public sector, and using the public sector as a platform for innovation. This was followed by suggestions regarding the development of cloud computing and information and IT competency and by initiatives regarding the environment, climate and green IT, as well as research and development in the IT area. This policy input is perfectly in line with policy initiatives since the mid-1990s.

6.3.1 TDC – the incumbent operator

The incumbent operator – first named Tele Danmark and later TDC – was created by a political initiative in 1990 as a holding company owning the four regional monopolies and one international operator that existed at that time. A large share of the total stock, resulting in a controlling interest, was later sold to a foreign company, Ameritech, which later merged with another American operator, SBC. In 2006, a group of private equity (PE) companies took over 87.9% of the TDC stock – a share that has been reduced gradually, with a final PE exit in September 2013.

As the incumbent operator, TDC is by far the largest supplier of physical broadband infrastructure. TDC is the only provider of a copper based infrastructure covering close to 100 per cent of all households. TDC also possesses the largest cable network, covering more than half of all households (homes passed). And, since 2003 they have laid down fibre tubes whenever a new building has been constructed, and they now own an extensive fibre network of 45,000 km.

TDC follows a strategy where gradually copper-based facilities are replaced by optical fibre and the DSLAMs are moved closer to the customer. In this way, it becomes possible to provide DSL services at higher data rates without providing a full FttH solution. Through the use of vectoring technology, TDC will be able to provide 100 Mbit/s and eventually up to 250 Mbit/s using the existing copper cables. TDC expects that the market share of copper-based solutions will decrease, although copper will still have a role to play in 2020.

TDC provides FttH as well. In 2009, TDC acquired DONG Energy’s 7,000 km fibre network with 15,000 end users in the capital region. In 2013 TDC bought the broadband provider ComX, also in the capital region. With these networks, TDC now has a fibre network, which covers more than 200,000 households (homes passed), and this number is expected to increase to 500,000 in 2020, or approximately 20 per cent of total households.

The role and impact of PE ownership has been subject to extensive discussion (Nyrup, 2007). While the data rates offered to consumers have not developed as fast as predicted (ITST, 2011), it is not obvious that the PE take-over changed the course of development significantly. For telecom users in Denmark, the most significant implication is that TDC has continued the course which was laid in the Ameritech/SBC period: i.e., to be a second mover with respect to new network solutions on the Danish market and not to use the Danish market as a test bed for wider international initiatives.

On the other hand, PE ownership has had severe consequences for the international presence of TDC, as the PE owners sold off most of the international operations and turned TDC into a company operating primarily in the Danish market. During the past decade, the two other large incumbent Nordic operators, TeliaSonera and Telenor, have become sizeable international players, while TDC has shrunk to become basically a national player.

6.3.2 The largest competitors

Next to TDC, the three largest competitors on the Danish market are the two other Nordic incumbents, TeliaSonera (under the brand name of Telia in the Danish market), Telenor, and the operator 3. Telia and Telenor are operating in fixed as well as mobile markets. However, both companies focus on mobile communications, as it has been difficult to penetrate the fixed market, while 3 is a mobile-only company.

Telenor Denmark, with a turnover of EUR 765 million in 2012, is the second largest provider of broadband and telephony in Denmark. Telenor entered the Danish market in 2000 through the acquisition of a majority share of Sonofon. In the Danish market, Telenor is first of all a mobile operator, but has also achieved a position as major provider of broadband services through the acquisition of Cybercity in 2005. Cybercity was established in 1995 and had for a short while, up to 2000, a higher DSL market share than TDC, on the then relatively small DSL market using unbundling of TDC’s network. In 2012, Telenor had around 175,000 DSL customers and a market share of approximately 15 per cent. However, the company does not own a copper infrastructure and depends on TDC’s access network for the provision of DSL services.

Telia Denmark is part of the Swedish/Finnish incumbent telecom operator TeliaSonera and the third largest telecom operator on the Danish market with a turnover of around EUR 683 million in 2012 when they had approximately 75,000 DSL customers (6.3% market share). Until 2010, they had their own cable infrastructure, as they owned the Danish cable operator Stofa. Telia was the first operator to introduce LTE services on the Danish market in late 2010 and their strategy is to market mobile broadband as an alternative to fixed broadband. The selling of their cable business in 2010 can be seen as part of this strategy.

In 2013, Telia and Telenor started offering mobile broadband jointly, using infrastructure sharing. The purpose is obviously cost savings, but it also indicates the severe competitive pressure for competitors to TDC in any technology area on the Danish market.

The operator 3 (Hutchison-Whampoa) entered the Danish market in 2003 in connection with the launch of 3G in Denmark. 3 was the first on the Danish market with 3G and the company was the largest 3G provider and promoter for a long time, while the other providers held back their 3G operation because of lack of services requiring 3G capacity. 3, however, was forced to enter the 3G market as soon as possible as they, unlike their MNO competitors, did not hold a 2G license as a fallback. The market share of 3 fell when the other MNOs started launching 3G, and 3 held approximately 10 per cent of the mobile market in 2012.

6.3.3 Syd Energi and other energy companies

In addition to DONG Energy, which sold its network facilities to TDC, other energy companies have invested in fibre networks since 2000, and their networks pass around 700,000 households (approx. 25 per cent of total households). Approximately 200,000 customers are connected or are about to be connected. Since the economic downturn started in 2008, the energy companies have scaled down their investment activities and are focusing instead on obtaining more customers to be connected to the existing infrastructure.

The most important player among the energy companies is SE (Syd Energi). SE is the energy provider in South Jutland. This region is one of the least densely populated regions in Denmark but, due to SE, it is also the region with the best coverage of optical fibre.

In February 2013, the Competition Authority approved the acquisition of Stofa by SE. In the cable access market, Stofa is the second-largest provider of broadband via cable after TDC. In 2011, 414,000 customers were connected to the Stofa network, approximately 20 per cent of the CATV market. Out of these, 151,000 were supplied with broadband. Stofa was formerly owned by Telia but was acquired by the Swedish private equity firm Ratos in 2010. Their infrastructure covers major cities especially in Jutland and Funen. There is little overlap with TDCs cable network and hence the direct competition between the two is limited.

In September 2010, the energy companies formed Waoo! in order to promote broadband services via their fibre infrastructure. With 155,000 customers (4.5 per cent of households), Waoo! is the most important provider of broadband via fibre.

6.3.4 Other infrastructure providers

Another player in the Danish market is COLT Telecom. They operate their own fibre network in Copenhagen and provide wholesale and retail services to business customers. GlobalConnect is yet another company and has been on the market since 1998. They provide dark fibre on a wholesale basis.

Fixed wireless infrastructures do not play a significant role in the Danish broadband market. Two licenses have been issued, one to TDC and another to Skyline which provided approximately 40,000 fixed wireless connections in 2010, but went bankrupt in June 2012.

In the early phase of broadband development, an interesting phenomenon that contributed to market expansion was the emergence of a few local community-based alternative network providers. These operators were established in order to serve a market need in primarily rural areas, where services provided by the major operators were either inadequate or very expensive. By far the largest example of such a network provider is Djursland.net. Djursland.net took off in year 2000, when the development of a wireless broadband network covering a population of 82,000 began.

This network was once one of the largest alternative broadband operators in Europe. The organization was dissolved in 2009, but the network facilities are still in operation. At present, this type of alternative networks is less important, as coverage by competing solutions has improved. Moreover, some of these networks have been taken over by Dansk Bredbånd, which later was bought by Waoo!, the broadband company owned by a cooperation of local energy companies. However, the alternative networks have played a role in establishing early coverage in rural areas.

6.3.5 Other service providers

There are many providers of retail broadband service on the Danish market – in the DSL field as well as the mobile area (as MVNOs). The development pattern, which has been seen repeatedly, is that service providers open with cheaper and more attractive pricing for users than the existing operators. They grow to a considerable size, running with an increasing deficit, but with a quickly developing number of subscribers. Finally, they are bought up by one of the large network operators, interested in both acquiring the subscriber base and preventing these operators from being taken over by one of their competitors.

6.9.1 Operators

In the DSL field, the national incumbent, TDC and its affiliates, is by far the largest provider. TDC owns the copper access infrastructure in Denmark and has almost three-quarters of the DSL market. In the cable area, TDC is also the largest broadband provider with approximately two-thirds of that market. The other major cable provider is Stofa, which is now owned by Syd Energi (SE), one of the energy companies that hitherto have concentrated mainly on fibre access. The mobile broadband market is more evenly distributed between the four mobile infrastructure providers (TDC, Telia, Telenor, and 3) and their affiliates. The fibre access market is dominated by the energy companies, with TDC trailing behind, seeking to follow the fibre market in order to expand its fibre activities when needed.

With respect to broadband, in none of the four access technology areas (DSL, cable modem, mobile, fibre) has the incumbent TDC been the first-mover. Cybercity (later taken over by Telenor) was the first-mover in the DSL area; Stofa was the first cable modem provider; 3 was first in 3G-based mobile broadband; and the energy companies are the leaders in the fibre market. All the while, TDC has, maintained a strategy of being a follower and moving quickly into the dominant role in a particular market segment when challenged by the first-mover – with fibre being an exception at present. In contrast to other European countries, the incumbent operator has been permitted to be present in all four broadband market segments.

In addition to the four main technology areas, there have been two interesting broadband initiatives involving other technology solutions. One is a self-organized grass-roots initiative in Djursland (the ‘nose’ of the Jutland peninsula) where activists built a local broadband access network based on Wi-Fi to serve the rural areas of Djursland.³ This initiative was started in 2000, when the broadband offers of the traditional Internet Service Providers were lacking or considered insufficient. The Wi-Fi network still exists (in 2013) and serves approximately one third of broadband connections in Djursland, which has a population of approximately 80,000 inhabitants. But it has remained a relatively isolated initiative and has not had any broader influence on Danish broadband developments.

The other initiative is concerned with WiMAX technology. In 2005, two licenses for providing WiMAX were awarded – one of the licenses to TDC. However, TDC announced right away that they did not intend to use the license but had acquired it in order to prepared: i.e., to see whether this market would develop. The other license led to actual deployment of WiMAX by a succession of companies with Skyline, owned by an energy company, as the last one. Skyline went broke in 2012 and had to abandon approx. 40,000 customers in rural and semi-rural areas in Jutland. Some of them have complained that they subsequently lacked sufficient broadband connectivity.

An important dynamic regarding broadband pricing relates to retail providers having entered the market offering a better deal to customers and acquiring a sizeable market share. After a while, when a sufficiently large number of subscribers have shifted to the new service provider, the company is taken over by one of the large market players, primarily TDC, Telia, or Telenor. These players actually have a mutual understanding that this is an untenable situation, undermining the profitability of the markets, but it has been difficult to act accordingly when new service providers have attracted a large number of customers and are up for sale. The business model for such service providers is to attract as many customers as possible, running with a deficit, and then to be bought up by one of the large network operators. This business model is especially prevalent in the mobile area (entrant MVNOs).

In the fixed network area, the incumbent TDC is very dominant. The affiliates of other Nordic incumbent operators, Telia and Telenor, have tried to break into the Danish fixed broadband market. This applies especially to Telia with its former ownership of the cable company Stofa. However, they have not really succeeded and have lately retreated to prioritize the mobile area, where they have joined forces with Telenor in the joint operation named TT. By mid-2013, TT offered the best LTE network in Denmark, and Telia and Telenor put all their strengths and ambitions in Denmark into mobile broadband.

In public debates on the fixed network markets, the main dividing line in Danish telecommunications has been between the Danish incumbent TDC on one side and all the other operators on the other. This includes the incumbents from other Nordic countries that have generally required better conditions for service providers on the Danish market – conditions they would often not grant to competing service providers in their home markets.

The only fixed network area where TDC is really being challenged is in fibre access. In the energy industry, Denmark has a number of local and regional companies. As these companies have rights of way and often dig up the ground in order to reach energy customers, they also started to offer fibre access a decade ago.

Hitherto TDC has not given much priority to fibre access. Their strategy has been to deploy fibre ever closer to the end-users but to use copper for the last stretch as part of their VDSL strategy. One result is that the energy companies dominate the fibre access market, and another is that the highest bandwidths offered in Denmark are generally in the provincial regions in Southern and Mid-Jutland (on the peninsula), where the energy companies have been most active in fibre deployment. In most other European countries, fibre is mainly offered in densely populated urban areas.

In order to contain this situation, TDC in 2009 bought the largest fibre operation in Denmark. This fibre operation had been running with huge deficits and DONG Energy had, therefore, tried for some time to sell it. Eventually, TDC bought it, but has not used this acquisition to boost their fibre operation or sought to attract new end-users to their fibre acquisition. Apparently their strategy is to wait with fibre-to-the-home and push fibre ever closer to the end-users as part of their VDSL roll-out.

The energy companies offering fibre joined forces in 2010, creating the joint marketing company Waoo!. In 2012, a new development took place when one of the energy companies, SE, acquired the cable company Stofa and thereafter left Waoo!. This suddenly made SE a far larger broadband provider than the other energy companies. How this will influence the strategic actions of the energy companies in the coming years is difficult to say. During the past few years, there has been a virtual political and market struggle between TDC, on the one hand, and the energy companies on the other. The energy companies have been seeking to present themselves as the companies that will deliver ‘true’ high-speed broadband to households and companies, while TDC argues that they upgrade their broadband networks when the actual demand is there.

In 2006, private equity funds took over ownership of TDC. The equity funds acquired 87.9% of the TDC shares but could not get hold of another 2.1% to reach 90%, which according to Danish law would permit them to force the remaining shareholders to sell their shares. Nevertheless, the equity funds totally dominated TDC, which raised some concern although not from the political authorities, nor the regulator. The concern was that equity funds are focused on short-term profitability rather than on long-term investments in the business they are entering. What has happened was that the international operations of TDC have been sold off.⁴ While TDC as well as Telia and Telenor had started internationalizing when the telecom markets were liberalized around the world, TDC has become a company focusing primarily on the Danish market and to some extent the Nordic market, while Telia and especially Telenor are international players.

The question is then whether the ownership of TDC by the equity funds has undermined the development of telecommunications in Denmark, by causing a lack of investments in infrastructure and new services. This, however, does not seem to be the case – although the counter-factual development cannot be produced. The overall result has been that TDC has focused primarily on the Danish market, but not that the Danish market has been starved of investments. Through the periods of different kinds of ownership, TDC has maintained a relatively conservative and reactive market strategy – following other market players and responding once they have gained momentum and become a possible threat.

6.9.2 Political-administrative institutions

In all countries, the national political-administrative institutions including the telecom regulator play an important role for the development of broadband. In the case of Denmark, there are at least two issues which are important to address.

One is the great emphasis that the policy system has put on the development of services and applications to be delivered on broadband, including the role of state institutions in promoting the uptake of broadband-based services and applications. Until very recently there has been a general agreement in the political system that no national public money should go into general telecom infrastructure investments. The focus has been on how to push and develop the use of the infrastructural resources, while the infrastructural investments have been left entirely in the hands of the operators.

The other issue is the development of the regulatory institutions. Because of the historical structure with regional operators in Denmark partly privately owned, there were actually separate telecommunication regulatory institutions in Denmark through major parts of the 20^th century. However, with the then-upcoming liberalization in mind, a new independent regulator was created in 1990. This regulator was very active in the 1990s, especially in the period following the full liberalization in 1996 and in the first years following the year 2000. Nevertheless, the regulatory institution lost momentum thereafter in connection with the lack of political attention and initiatives from the policy side. With the coming to power of a new center-left government in 2011, the regulator was split up into different parts, and the traditional telecommunication regulatory assignments were transferred to the Danish Business Authority under the Ministry of Business and Growth.

There has been an overall agreement in the Danish political system concerning telecom and ICT policies, and there have been no significant policy shifts with changing governments. This applies to the whole period since the liberalization in the mid-1990s. However, lately there have been some disturbances in the Danish telecom policy arena. These disturbances relate to the expansion of broadband networks.

A political goal was set that all households and companies in Denmark must have access to at least 100 Mbit/s downstream by 2020. By the end of 2012, 65% of households and companies already had access to 100 Mbit/s or more. However, the difficulty is reaching the remaining households.

What the results of the political process will be is difficult to predict, but there are signs of unrest in the policy environment that have not been observed since the policy agreement regarding telecommunications in the parliament in 1999. And it is the broadband issue which is causing this unrest. In 2009, the government established a High Speed Committee, which in 2010 presented its deliberations and suggestions. The suggestions by the committee centred on the promotion of services and applications but, shortly after the presentation of the High Speed Committee report, the government issued its policy goal of 100 Mbit/s for all. At first, this did not create much alarm but as the implementation deadline comes closer, the political debate has started to focus on how this policy goal will be met.

So far, the policy focus has been on advancing services and applications. The various governments and their administrations have time and again issued reports and taken steps to promote the use of broadband-based services. This has entailed initiatives, such as promoting the internal use of broadband-based services in governmental institutions and also requiring those who wish to communicate with public institutions to use electronic means. In 2012, a small number of public services (central and municipal) started the transition to full digital provision. For the coming years, a wide range of public services will be transitioned to digital provision. These administrative goals are well in line with the original Danish policy approach in the area.

6.9.3 Expectations based on theory

In the introduction we addressed the ‘ladder of investment’ concept (Cave, Reference Cave2006) and posed the question whether the market has responded as intended to the application of LRAIC-based access pricing. The crude answer is ‘no’ or ‘maybe to some extent’. If climbing the ‘ladder’ from resale to full unbundling is considered, there has been no such development in Denmark since 2008. If new infrastructures are considered, the answer is ‘maybe’. New operators obviously have not built alternative PSTN infrastructures, as TDC has such a network with full national coverage. Telia tried with the cable operation of Stofa to become an alternative infrastructure provider but it sold Stofa to a private equity firm, later to be bought by the electricity company SE. Today, the strategic focus of Telia as well as Telenor in the Danish market is on mobile broadband. As for the electricity companies, which have become the primary drivers of fibre in the access networks, they never considered offering broadband with service provision via the PSTN of TDC. The overall conclusion, therefore, must be that actual infrastructure and service competition is driven much more by strategic moves by operators than by a theory of gradual growth towards infrastructure provision. The increasing interconnection and access prices must be inputs to the strategic considerations by alternative operators. However, it is not clear that this has been the main driver.

In the Danish market, there is a more even distribution of the different access technologies than in most other European broadband markets. Nevertheless, DSL plays the central role. An important reason is the national coverage of the PSTN for many years. DSL is, therefore, a natural technology trajectory. This indicates that path dependency (David, Reference David2007) is an important element in broadband developments.

Another central issue is the relationship between competition and innovation (Aghion et al., Reference Aghion, Bloom, Blundell, Griffith and Howitt2002). The Danish case clearly illustrates a positive correlation between competition and innovation in infrastructure developments. The Danish incumbent has been a ‘second mover’ regarding new infrastructures. TDC was neither the first on the Danish market to launch ADSL nor the first to offer cable access. The same applies to fibre and mobile broadband. How these access technologies would have developed without competition is impossible to say, but it is clear that alternative access offers have put pressure on TDC to advance their operations in the different technology areas.

Last but not least, the question regarding the interrelationships between service and application provision and infrastructure provision deserves mentioning. The question is what comes first: the demand and provision for services and applications or the demand and provision for infrastructure. This is, of course, a chicken and egg discussion. However, the important point is that Danish broadband policies have given priority to stimulating service and application developments. The extent to which this differs from other countries can be discussed. However, the political priorities in Denmark have been clear since the new policy direction was announced in 1994–1995 – with one interesting exception being the sudden policy priority of reaching 100 Mbit/s downstream coverage by 2020, which was clearly inspired by broadband initiatives taken in a number of other countries following the financial and economic crisis starting in 2008.

7.3.1 Swedish broadband policy

The current broadband evolution started in 1999–2000 with a government bill that was aimed to support national and local initiatives and investments. The headline was ‘Trust to IT, Access to IT and IT competence’. The trust included security and technology. Access implied broadband provision with national, regional and local network support, including SMEs and schools. The financial support included SEK 2.5 billion (or roughly EUR 280 million) to a national operator-neutral backbone, SEK 3.2 billion (in the form of tax breaks) to municipalities to develop access and SEK 2.6 billion to regional networks and to create local infrastructure plans.

This was an important political declaration that ICT and broadband are important for society, and all local politicians and authorities started to discuss and plan broadband deployments. However, after the so-called IT bubble burst, these activities cooled down somewhat.

The current Minister for Information Technology and Energy has restarted a program to speed up broadband development: she launched a Swedish Digital Agenda and formed a broadband commission (Bredbandsforum) together with the regulator (PTS) and the industry to promote broadband development. Bredbandsforum has started several programmes; for example, fibre to the farms, which is a broadband cooperative model.

The Swedish Government’s broadband policy is documented in an official publication, Regeringskansliet (2009). There one can read that the overall objective is for Sweden to have world-class broadband. A high use of IT and the Internet is good for Sweden in terms of growth, competitiveness and innovation. It contributes to the development of a sustainable society, and also helps to meet the challenges of increased globalization, climate change and an ageing population in a sparsely populated country. A prerequisite for meeting the challenges set out in the Digital Agenda is access to high-speed broadband across the country. Swedish targets are: 40% of households and businesses should have access to at least 100 Mbit/s by 2015 and 90% in 2020. It is important that Swedish businesses and households in all parts of the country can take advantage of the opportunities offered by powerful broadband access. Then traditional work practices can change, new services and business models can evolve and new behaviours can emerge.

The central tenet is that electronic communications and broadband access should be provided by the market. The government should not control the market, nor the development of technology. Their task is to create good conditions for the market and remove barriers in the development of broadband by ensuring there is relevant regulation in place.

7.3.2 Regulation

Municipal planning responsibilities are clarified in the Planning and Building Act as reinforcing the connection to the infrastructure for electronic communications. The government has also initiated a so-called Broadband Council for collaboration and dialogue on the deployment of broadband. Furthermore, the Swedish NRA – the Post and Telecom Authority (PTS) – is assigned the task of investigating how radio-frequency bands for electronic communications can be used to improve accessibility in areas which lack broadband access, or with low-capacity and low-quality broadband. The universal service obligation is revised to reflect a minimum level of functional Internet access.

While FttB/FttH networks have been deployed massively in Sweden for over a decade, regulation impacting fibre access networks (be it directly or indirectly) has started to appear only recently.

In May 2010 the PTS passed decisions regulating the wholesale market for (physical) network infrastructure access (including LLU and shared access) as well as the wholesale market for broadband access. The Decision on the market for physical network infrastructure access (LLU) includes both copper and fibre lines. In addition to the provisioning of fibre access, it also requires the incumbent to deploy new fibre infrastructure in existing ducts if requested by an operator who is willing to pay the cost of the investment. However, this decision has been appealed in some parts and a court ruling on this issue is still pending. Concerns have been expressed that the reference offer of the incumbent does not correspond to the Decision of the NRA, as it does not allow the purchase of access to fibre between the incumbent’s Metropolitan Point of Presence and the network termination points connecting buildings, so-called ʻFibre to the Buildingʼ, FttB¹.

In August 2009, the government introduced an amendment to the competition law (which came into force on 1 January 2010) aimed at preventing the state, municipalities or regional governments from engaging in commercial activities in a way that is harmful to competition. The new legal provisions came as a response to the numerous complaints received by the national competition authority (NCA) regarding the problems encountered by private operators when competing with public ones. Certain operators still consider that some housing companies owned by the municipalities are trying to eliminate competition by allowing only one electronic communications provider to offer services in their blocks of flats.

In 2009 the PTS was assigned the task of studying open networks and services, and reached the following conclusions:

1. 1. Openness creates the prerequisites for innovation and competitiveness but must be balanced against other interests worthy of protection, such as incentives to invest and network security.

2. 2. Openness is promoted by securing non-discrimination and effective competition.

3. 3. Openness is of great significance and it is therefore important that suppliers in their marketing activities and in applicable terms and conditions provide clear and specific information with respect to lock-in periods and restrictions relating to Internet access and access to services.

The report that concluded the study states that one essential challenge to openness today is restrictions in access to passive infrastructure (e.g., dark fibre). A second major challenge can be traced back to insufficient consumer mobility due to long lock-in periods, high transition costs and other lock-in effects². One challenge which affects several levels in the value chain relates to openness when managing electronic communications over the Internet, network neutrality. The report suggests several measures aimed at securing openness, measures which take all interests worthy of protection into consideration – especially incentives to invest and network security. PTS suggests stronger principles for equal treatment when building new infrastructure, increased access to existing infrastructure, mandated information to consumers regarding possible pitfalls, the importance of openness and, finally, increased transparency regarding the existence of potential limitations of Internet traffic such as prioritization of traffic and blocking of services.

Somewhat controversially, PTS has recently decided to introduce price regulation for TeliaSonera, recognizing its position of significant market power (SMP). The price regulation is based on four broad and complex models. These models calculate the cost (or would-be cost) for Telia to install fibre throughout the entire country to every household and business. A price per connection is then calculated and, because Telia currently only has x% of the broadband market, the price that Telia can charge is x% (currently roughly 50%) of the calculated price per connection. While this regulation only applies to TeliaSonera, this puts a de facto price regulation on the whole market that is below the actual price of dark fibre. If another actor – a municipality network, for instance – charges the actual deployment cost, it risks losing customers to TeliaSonera. If it tries to compete with TeliaSonera to retain customers and applies TeliaSonera’s regulated price, it risks be taken to court for illegal state aid for setting a price below production costs (which the PTS model effectively requires). This in turn leads to a re-evaluation of municipal networks, which can affect municipal finances and thus provide fewer opportunities for fibre investment.

It is interesting that, against a certain widespread understanding in Sweden, the position of the Swedish regulator (PTS) is that staying on the passive level is no guarantee that the market is not distorted. For instance, if a service is currently provided by other actors on commercial terms and on a competitive basis, installing a fibre infrastructure which allows others to come onto the market at potentially lower prices is per se distorting the market. Naturally, that would not apply if the fibre network is responding to a need which the market (with the current infrastructure) cannot provide. This can be seen as an academic question. In any case, PTS’s mantra is that the market should take care of NGA deployment.

7.3.3 Regulatory trends

The price regulation of TeliaSonera’s fibre access introduced by PTS has generated a lively and somewhat bitter debate. It is believed that the price regulation represents a de-facto price standard which is going to make any fibre investment uneconomical and even politically unattractive (due to the risk of incurring state aid allegations). This position is widely held in the market and there is a significant degree of uncertainty on the future of this contested regulation, especially after a report commissioned by the Swedish Association of Local Authorities and Regions (SKL) (Deloitte, 2011). The report’s goal was to examine the impact of the price regulation and its results pointed to several risks; it concludes that there is no justification for claims a regulated charge for dark fibre will benefit end-users through lower prices and better choices.

SKL points out that many municipalities have invested heavily in fibre networks to assure access to modern broadband for their inhabitants. Municipalities have built up those networks because private investors were not taking such initiatives. Now those public investments are at risk of being frozen completely. SKL is discussing a scenario in which the market will develop towards a single, dominant player: i.e., a return to the monopoly of twenty years ago. TeliaSonera is backing the requests that this regulation be phased out, or the calculation model modified, claiming that claim this regulation and the specific model used are based on past circumstances no longer present and that, as the market has changed dramatically, so should the regulatory agenda.

Given this controversy, it can be expected that these rules are going to be amended, if not repealed altogether.

7.3.4 Government support and state-aid issues

The government bill ʻ99/00ʼ identified those areas where the market was too weak for a commercial roll-out. Hence, any such region or municipality could apply for financial support to deploy broadband networks. All municipalities received money to develop a local ICT infrastructure plan. This plan, which must include the connection of public buildings, such as schools, hospitals, etc., is the basis on which to apply for additional government funding for the actual network deployment. A crucial prerequisite to qualify for funding is that the network must be operator-neutral. This was widely viewed as being compliant with EU state aid rules in the sense that intervention was made in situations of market failure and acted upon to support the telecom market with a fibre infrastructure for all, rather than to compete with market actors. This was confirmed, as no claims of illegal state-aid were filed against the municipal initiatives.

7.4.1 Drivers for FttH deployments and broadband strategies

As is the case elsewhere, there is an underlying belief accompanying these fibre access network projects in Sweden: a political conviction that increased broadband penetration leads to social, environmental and economic benefits. Broadband infrastructure, primarily fibre, is increasingly seen as the fourth utility, and it is considered the task of the public administration to make sure that this utility is put in place one way or another. This, combined with the lack of commercial operators willing to invest in broadband, provides the major driver for the regional and municipal deployments described in this study. All the cases analysed concern predominantly rural areas, meaning that the population density is very low. This implies on the one hand that the investment costs are high and, on the other hand and more importantly, the potential revenues per deployment are significantly lower. This breaks not only the traditional FttH business case on the basis of traditional triple-play, already not a very strong case in very dense areas, but also the upgrading of central offices with DSLto connect only a few end-users. This is typically the case in the areas analysed in this study.

This is also a familiar situation in other European regions in which, due to the topography of the area, the access to broadband services was scarce and many areas were not being equipped with ADSL, which prompted the local governments to intervene⁴

However, once the infrastructure is deployed and activated, it is important to get the end-users on-board to use the network. According to the experiences from the projects analysed here, an important driver for service uptake has been the availability of high-speed Internet service. However, a second wave of subscriptions is now anticipated to be driven by TV-services (especially HD) which can be provided cheaper, with higher quality and flexibility of use and with broader choice.

A very important driver to generate interest and support in the deployment phase and high uptake in the operations phase is the presence of local fibre champions and a positive ICT culture. This is of course linked to historic factors (e.g., Hudiksvall had LM Ericsson and the Fiber Optic Valley, which helps to explain why it was so much ahead of the neighbouring municipality, Nordanstig), but the lack of such a catalyst can be compensated to a great extent by visionary politicians and by running extensive and continuous information campaigns among the population.

7.4.2 The open access business models

The dominant business model among Swedish municipal fibre network operations is an open access network model. The four cases described in this chapter reflect different flavours of such a model; therefore this section describes the model and reviews experiences obtained in Sweden.

The open network model, in which services are provided on a fair and non-discriminatory basis to the network users based on a shared infrastructure, is enabled by conceptually separating the roles of service provision and network provision. See Figure 7.2. Due to the different technical and economic characteristics of the different layers of the network, different roles and actors can be identified. A fibre access network broadly consists of a passive infrastructure (including right-of-way acquisition, trenching, cable duct construction, fibre cable installation and connections to home and office premises), and active equipment (transponders, routers and switches, control and management servers). The passive infrastructure is typically characterized by high CAPEX, low OPEX, low economies of scale (individual connections to homes and offices), and is highly local, hard to duplicate and hence subject to regulation. The active equipment is characterized by lower CAPEX, higher OPEX, higher economies of scale, and is subject to less regulation if provided in competition. These factors allow a further role separation between a physical infrastructure provider (PIP), which owns and maintains the passive infrastructure (typically real estate companies, municipalities or utilities), and the network provider (NP) which operates (and typically owns) the active equipment (incumbent operators, new independent operators or specialized broadband companies).

Figure 7.2 The open network model and typical open access value chain

Source: Adapted from Forzati et al. (Reference Forzati, Larsen and Mattsson2010).

Depending on which roles different market actors assume, the network will be open at different levels and different business models will arise, as illustrated in Figure 7.3. A single actor may act both as PIP and NP (7.3, a), in which case the network is open at the service level. If the roles of NP and PIP are separate (b, c and d), then openness at the passive infrastructure level is achieved. Generally, one PIP operates the passive infrastructure, while one or several NPs can be allowed to operate the active infrastructure, typically over a fixed period of time, at the end of which the contract may or may not be renewed (in which case a new NP is designated and active equipment may need to be replaced). Most often, economies of scale make it impractical to have a truly multi-NP network (although larger networks may assign the operation of different geographical parts of the network to a different NP). Independently of the specific model, however, the NP should offer different service providers (SPs) access to the network on non-discriminatory conditions. The end users typically purchase services directly from the service providers. The NP receives revenue from the SP and pays a (one-time) connection fee to the PIP for network access.

Figure 7.3 Access network business models

Source: Adapted from Forzati et al. (Reference Forzati, Larsen and Mattsson2010).

If the NP also acts as SP (b) the network cannot be described as really open according to the definitions applied, but it is still more open than the conventional vertically integrated FttH model in (g), which most incumbent operators follow today. In case of local loop unbundling (LLU, Figure 7.3e), a vertically integrated operator is still present, but there can be multiple actors working as combined NP and SP. In case of bitstream access the vertically integrated operator assumes the role of NP, but there can be multiple SPs offering their services using wholesale bitstream access.

Note that some of the roles shown in Figure 7.3 can be divided into several others and that the delineation between the various roles is not always as clear as the figure suggests. However, the figure provides a fair idea of which types of business models with respect to open access are used in Sweden today. Although the degree of ‘openness’ will vary depending on the type of actor and the layer in the network that is considered.

The first optical access networks in Sweden were built around the turn of the twenty-first century by Bredbandsbolaget, TeliaSonera and a number of municipalities. While Bredbandsbolaget and TeliaSonera are vertically integrated operators, owning the whole value chain from fibre infrastructure to services, for the municipality networks the situation was different. In most cases a municipal company both owned and operated the network (PIP + NP). Although situations with only external SPs existed, in many cases, the municipalities also acted as SPs, sometimes in competition with other SPs. This became a default as it had been difficult to attract external SPs for several reasons:

1. (1) limited number of subscribers;

2. (2) connection procedures (both technical and administrative) varying from network to network, thereby hindering economies of scale for the SPs;

3. (3) a small number of SPs, much fewer than today;

4. (4) lack of business actors taking the role of NP and handling the contacts with the SPs, etc.

The market has matured considerably during the past ten years. The process of connecting SPs is now much simpler, technology has improved, there are network providers handling SPs as wholesale clients, the number of SP is now much higher, etc. Also, many municipalities have come to the conclusion that they should focus more on providing infrastructure to their citizens, rather than competing with commercial companies in the services market.

Although the market is not yet fully mature, in general municipal network operations have moved downwards in the value chain – increasingly closer to just owning the fibre infrastructure (pure PIP role) – while network operation (NP role) and service delivery (SP role) are left to other players (other municipal utilities, independent operators, or even large telcos such as TeliaSonera). Despite the clearer roles, many smaller municipal networks struggle economically, due to a reduced revenue share in a marketplace with many competing players. This has led to a consolidation process where municipal networks either have been acquired by competitors or have started to collaborate closely. This consolidation process is still ongoing.

7.4.3 Services and uptake rate

Services present on Swedish FttH networks are predominantly voice (both stand-alone IP and over-the-top), Internet and TV, which make up the classic triple-play. Different data rates for Internet service are generally offered, from 1 Mbit/s to 1 Gbit/s. The most commonly offered and commonly subscribed services are 10 Mbit/s and 100 Mbit/s, for which the retail prices range between EUR 10 and EUR 30⁵. Telephony is generally quite cheap but it is not a driver of uptake. Internet service is provided almost exclusively over an Ethernet point-to-point solution; thus, there are two alternative configurations regarding the customer-premise equipment (CPE):

1. For FttH, a CPE consisting of an O/E (optical/electrical) converter and an integrated switch (sometimes a router) which functions as a service separator using VLAN (virtual local area network) tags;

2. For an FttP + in-building LAN, if only Internet-based services are provided, a simple O/E converter and a switch are placed in the basement of a multi-dwelling unit: CAT 5 or CAT 6 copper cables are then drawn to each apartment; hence, end-users do not need any special CPE and can connect the Ethernet CAT 5/6 cable to their computer or home Wi-Fi router; otherwise, a solution similar to the FttH case is followed.

TV over the FttH connections are usually provided using a VLAN, and the service is then separated at the CPE. TV packages usually include must-carry channels (around eight national and regional channels), plus bundles of four to six channels for a cost of EUR 5 to EUR 10 per bundle. Both traditional proprietary TV platforms are present (with a cost of roughly EUR 15 for the basic package) and now there is also the so-called open-TVplatform, using a Sweden-wide standard open platform, which allows the end-user to change TV-provider without having to change the set-top box. Prices for basic packages with the latter are as low as EUR 3.50 per month.

Other services such as video-on-demand, home-security (video-surveillance), cloud services (back-up, remote hard drive, antivirus, etc.) are starting to appear, although these are for the moment few in number and the added value compared to equivalent over-the-top services (e.g., running over the Internet connection) is not always clear. Several e-health pilot projects have been run but, for the time being, the big breakthrough is still to come (although SMEs like Open Care are very active). When it comes to e-government services, these are for the time being limited to over-the-top, web-based basic services also available elsewhere, but arguably of better quality through FttH.

A service that is potentially of interest is teleworking, including teleworking centres. While teleworking is an increasingly popular phenomenon in Sweden, and one which is greatly encouraged by fibre networks, teleworking centres never really picked up. One reason for this may be the generally very good home connections, enabling distance working without the extra cost for office space.

The uptake of services in the Swedish municipal networks varies. As an indication, the average uptake on the networks operated by Zitius (currently active in ten municipalities) is roughly 35 per cent. The trend appears to be similar to the uptake of cable-TV, and that is generally gradual. As further reference, the Deutsche Telekom led research project OASE⁶ works on different scenarios: a ʻconservativeʼ one in which a 60% uptake is reached in 20 years (in areas with high ADSL density and relying on a traditional triple-play-driven, vertically integrated business model) and an ʻaggressiveʼ scenario in which the same uptake is reached in 8.5 years. Interestingly, even this is conservative in some cases: some rural, single-dwelling-unit areas in Hudiksvall and Säffle have uptake rates of 80% or more only a few years after deployment. Queues of households wishing to be connected to fibre (for an entry fee of EUR 1500) are now the case in Hudiksvall.

7.4.4 Observed socio-economic benefits

One of the first effects that was observed in FttH municipal networks in Sweden is a saving of 30% to 50% of the total municipal data and telecommunication costs (see also Forzati et al. Reference Forzati2012). This is partly due to increased efficiency (reduced equipment, energy consumption, and footprint per unit of transmitted information) and partly due to the fact that the high-capacity fibre network allows for more competition between service providers and thus lower prices.

In 1996, the city of Stockholm started purchasing telephony services from the open market, which was unique for public organizations in Europe at that time. The city had recently connected its operational sites and offices with its own fibre network. This allowed it to procure its telephony in full competition and to drive down costs. The city’s external telephony cost was between EUR 15 and EUR 20 million. The competition made available on its fibre-optic network resulted in a savings of 30%. However, this is a conservative estimate, and savings have probably grown larger over time, says Per-Olof Gustavsson, who at that time was active in the City of Stockholm’s city council office. In Jönköping, where the fibre connection has been less extensive, the savings figure was around 10 to 15 per cent.

Similar savings have been observed at regional administrations as well. The Stockholm Regional Council (Stockholms läns landsting) reduced its data and telecommunications costs by 50%, equivalent to roughly EUR 8 million, thanks to the fibre network. In Norrbotten, a fibre network has been installed linking 5 hospitals, 33 clinics and 34 dental clinics, which reduced communication costs also by 50% while providing fifty times faster communication. Service providers have been able to create solutions for digitized medical records, transmission of digital radiography, digital recipes, video-conferencing and IP telephony.

Another interesting fact is that the Swedish tenants’ association (Hyresgästföreningen) has agreed with property owners’ associations and housing companies to an increase in the rent of around EUR 5 per month. This can be considered a conservative evaluation of the perceived added value of fibre connection for the end-user because it makes available high-quality services at lower prices, especially entertainment and communications, the ability to work remotely and more free choice of work and housing, improved individual health, reduced need for hospitalization, simpler and more transparent interaction with public services, etc.

Indirect effects that were observed included the reduction in migration to larger cities (a problem among rural municipalities in Sweden) and improved employment, due to the availability of an ICT infrastructure that prevents businesses from moving to the cities, creates new business opportunities and allows people to stay in or move to rural areas with higher quality of life, thanks to a healthy local business environment and the possibilities offered by distance working. See for details the four cases described in Section 7.5.

7.4.5 Successes, challenges and future directions

Fibre deployments in Sweden have been successful in general when it comes to public support and end-user uptake. In some cases, requests for fibre connections exceed the roll-out capacity and queues have been forming. This success has been visible especially in rural areas where alternatives for Internet connections are lacking, and where bottom-up approaches (e.g., in terms of co-operatives to build village networks) are often used to gather support for centrally driven initiatives. Sometimes the municipality only needs to connect the existing village networks and build out the metropolitan section of the network.

In the urban areas it is often more challenging to gather significant uptake, especially among privately-owned multi-dwelling units (MDUs). Publicly owned MDUs are usually inclined to sign up, thanks to the involvement of the housing organizations, which see fibre not only as an infrastructure upgrade but also as a tool to simplify the provision of traditional services (the fibre network they connect to is usually an open network, therefore leaving the choice of Internet and TV providers to the tenants, which is seen as highly desirable) and more advanced services, such as estate management tools (remote surveillance, etc.).

Among the challenges being faced by municipal fibre network owners is that of attracting service providers, especially in the smaller and more remote municipalities. This has been to some degree solved by forming regional networks, which are a type of federation of small municipality networks⁷. These regional networks provide the scale, visibility and ‘single-interface’ towards the service-provider market, and have proven to be very successful in bringing service providers to all member municipalities. Another positive effect is that they have made it possible for small networks to rent out dark fibre to the wholesale market (this currently accounts for 30% to 70% of municipal network revenues).

The financial situation is generally quite positive, with municipal network companies showing positive results after ten years (with less than 25% of them currently showing negative results)⁸, although different business models and different situations have led to different outcomes, also depending on how depreciation and one-time connection fees are accounted. In general, companies that only invest in passive infrastructure (PIP role), and contract out the network management (NP role) tend to have a leaner organization, an easier business case (similar to a municipal utility), and a better medium- to long-term financial situation. For these fully functionally-separated networks, negotiation on price in the NP contract is a delicate balancing act, although market prices are now starting to appear. Some companies, still vertically integrated, tend to be in a relatively good financial situation, although in most cases their fibre deployment is limited to the nodes for xDSL backhauling (FttN). Extending fibre to the premises (FttH/FttB) will require extensive investments, which will significantly affect the cost-revenue balance. Another delicate factor is the accounting, in particular the choice of the depreciation period for the investment and the distribution of the revenue from the one-off connection fees: fine tuning these parameters can in some cases make or break the business case.

A practical and important limitation of FttH in general is its reliance on power at the end-users’ premises, which makes it challenging to deploy critical services on fibre connections (such as nurse alarms for elderly and reduced mobility citizens). There are several solutions being investigated and tested but none is sufficiently reliable yet, so the issue remains for the time being.

Turning to future directions and trends, the most important one is probably a general move down the value chain. Encouraged by the SSNf (Swedish Association of Urban Networks) and SKL (Swedish Association of Municipalities and Regions), many municipal networks are migrating from vertical integration towards a layered model, in which the municipal network company only owns and operates the passive infrastructure. The rationale for this is that passive infrastructure is a familiar business for municipalities and their utilities⁹, while active equipment is better left to the market¹⁰. While in general this is seen as the right way to go, and there are no more municipalities choosing to build a new fibre network and operate the active layer themselves, some challenges remain with this layered model, which should be mentioned:

1. Some municipalities have moved away from the integrated PIP/NP model but are currently stuck in a hybrid model (whereby some connectivity services are still offered), which risks removing traditional revenue streams while leaving important residual overhead and running costs (see the Hudiksvall case).

2. The contracting-out of the NP role implies that specialized local technical staff is not needed anymore: while this reduces the costs significantly and brings about more specialized competence overall, it has the drawback of increasing the potential response time in fault-management (see also the Säffle case).

3. The open access business model has been a success in bringing freedom of choice and low prices, but also poses some challenges in the management of the responsibility areas (fault management and first-line support) between the service providers and the network provider (running the actual network connectivity). A standard for open network interfaces would probably help make things run smoother¹¹.

Among other trends and future directions, we note that redundancy (generally in the form of rings) is recognised more and more as a critical feature of the network, in order to attract businesses and public sector actors. Also, a progressive defragmentation of the municipal networks is taking place, partly thanks to the emergence and success of regional federations mentioned above and partly to a consolidation process whereby some municipal networks are either acquired or effectively merged into larger neighbouring networks, thanks to contracting out the NP role. Among the effects seen are a reduction of resource duplication, increased professionalism and general synergetic effects. However, to do that on a regional basis (e.g., by centralizing the tens or hundreds of technicians scattered in various municipal networks) would provide a great resource for all, although that is seen as too big a challenge at present.

7.5.1 Hudiksvall

The Hudiksvall municipality (Hudiksvalls kommun) is situated in Gävleborg County, on the coast of central northern Sweden, some 300 km north of Stockholm, 130 km north of Gävle (population: roughly 100,000) and 84 km south of Sundsvall (population: roughly 50,000), to which it is connected by the E4 highway and the East Coast Railway (Ostkustbanan). The municipal seat is in the town of Hudiksvall.

The town of Hudiksvall is located on the coast, and the municipal territory extends along the coast as well as for several tens of kilometres inland. The territory is largely rural but includes several towns and villages. Roughly 40% of the population lives in the main town, so the fibre deployment is of several very different characters: urban in town, ‘sub-urban’ in the villages, and sparsely rural in the largest part of the territory.

The largest employers are the municipality and the county council, with around 38% of the workforce. Although in decline during the twentieth century, the forest industry represents half of the industrial activity; the largest private employer is the Holmen paper product company, with about 10% of the workforce. The electronics industry has become an important sector (19% of the workforce) and is dominated by Ericsson Network Technologies, making optical and copper cables for telecommunication systems. This, together with efforts by the municipality and the leveraging of European structural funds, has led in more recent times to the development of fibre research and speciality production¹³.

Business model

Fiberstaden operates a hybrid between a vertical integration and an open access business model. (See also Figure 7.4.) The role of the NP is currently given to Zitius in Hudiksvall, but Fiberstaden still sells wholesale connectivity services. In Nordanstig, Fiberstaden acts as an integrated PIP+NP. In both cases, the networks are owned by municipalities directly, not by Fiberstaden. In Hudiksvall, Zitius offers a broad range of services, through a portal called bredbandsväljaren.se (ʻthe broadband chooserʼ), where all services carried over the network (currently offered by twenty SPs) can be ordered in one place. However, once a service is purchased, the end-user pays directly to the service providers and all communication takes place directly between SP and end-user. The service providers are attracted to this model thanks to the nationwide presence of Zitius (so an SP only needs a point of presence in Stockholm to serve a large number of end-users across the whole country).

Figure 7.4 The business model for the Fiberstaden fibre network, Sweden

Source:Authors.

Ten service providers have switches in Hudiksvall and five SPs have switches in Nordansting. Fiberstaden also offers point-to-point transmission capacity to companies and, increasingly, dark fibre (e.g., to Tele2 to connect 4G masts; TTC; Norrsken; and Telia). Currently, 30% of Fiberstaden’s revenue comes from dark fibre and wholesale connectivity services.

As part of the retail business model, the media converter at the customer’s premises is owned by Fiberstaden. Fiberstaden receives EUR 4.50 per month per active customer from the NP Zitius. Troubleshooting is taken care of by Fiberstaden; Zitius has no local staff.

In theory, the end-user has no need to know about the existence of the NP and indeed the bredbandsväljaren.se portal is accessible directly from the municipality’s website and from the websites of the individual SPs. At the municipal website the PIP is presented as a municipal service, just like water, electricity or sewage. Zitius is not even mentioned in bredbandsväljaren.se, despite its being managed by the company. However, challenges remain, notably in fault management, as a clear demarcation of responsibilities is missing, leading to end-users being sent back and forth between SP, NP, and PIP when faults occur.

The price to connect a single home to the property border is EUR 1,500 (the actual cost is estimated to be roughly double), while property owners take responsibility for the digging across the property.

Examples of the socio-economic impact of fibre

Hudiksvall started installing fibre in 2002. Since then, the number of firms in the Hudiksvall municipality has increased by 6% to 14% per year between 2004 and 2009. Representatives of Hudiksvall’s municipality say that fibre has a positive impact on the municipality’s overall business climate. Moreover, there are examples of companies moving from Hudiksvall town towards more rural areas of the municipality thanks to cheaper rent.

Hudiksvall previously had, like many rural northern municipalities, a negative population growth, but since the installation of its fibre-optic network, the municipality stopped losing residents. In December 2002 its population had fallen to 37,048 (down 4% since 1994). The decline halted in 2002 when a significant number of households were connected by fibre. See Figure 7.5. This suggests that the effect of fibre investment may have a fairly rapid impact on population trends, although other factors may also have contributed. Similarly, thanks to fibre, in Nordanstig one can now see people returning to the municipality after having left years ago.

Figure 7.5 Population evolution in Hudiksvall, Sweden, 1995–2012

Data source: Statistics Sweden (SCB).

There are many examples of the effects of fibre on individuals in the local communities: for instance, in the village of Lindefallet, between Hudiksvall and Söderhamn, where 98% of the residents are connected to fibre. Several people decided not to move, thanks to fibre. Peter Engstrom, 34 years old, and Linn Sjoberg, 30 years old, chose to build a house in Lindefallet because of fibre, which is seen as a prerequisite for living close to the forests and nature without feeling completely isolated. Peter and Heather Nilsson moved from California to raise their four children in a safe environment, and journalist Marie Sandberg and author Georg Johansson moved there from Brussels. Several local companies have managed to assert themselves in an increasingly competitive environment thanks to fibre, and even farms are connected to fibre.

Since fibre was deployed in 2004, the population increased by 7.5% in a village that has neither a school, nor a health clinic, nor an industrial area. The investment in fibre, in combination with active associations, makes Lindefallet a lively village and thus a village that people want move to and to invest in.

7.5.2 Säffle

The Säffle municipality (Säffle kommun) is in Värmland County in west central Sweden. Its seat is located in the town of Säffle. The municipality covers a peninsula in Lake Vänern (Värmlandsnäs), and includes a large freshwater archipelago.

In the town of Säffle, the economy is largely based on industry. Säffle has continued to grow as the wood pulp industry has expanded in Sweden. The pulp mill in Säffle has been a major driver of the local economy. In the remainder of the municipality farming is important. The Värmlandsnäs peninsula is very significant to the economy through pork production. The area supplies more than 200,000 people with pork. The archipelago off the peninsula (which also has old rune stones and other ancient monuments) is a tourist attraction.

The territory is largely rural with smaller localities of few hundred inhabitants each. Two-thirds of the population are concentrated in the main town, so fibre deployment is of three very different types: urban in town, covering the largest portion of the population, ‘sub-urban’ in the villages, and rural in the largest part of the territory.

In its broadband strategy, the Säffle municipality declared that every household should have access to a high-speed network. Due to the lack of commercial operators ready to invest in broadband in the municipality, it decided to start a fibre deployment project through a municipality-owned company called Säkom. All investments have been made on commercial terms and have received financing partly by EU structural funds and partly through commercial loans.

Business model

Säkom operates a ‘pure’ open access business model (variant (c) in Figure 7.3; see also Figure 7.6). The NP role was taken by Telia by agreement in autumn 2009. This was later amended and Telia is now allowed to offer services. Säkom is a small organization, which the CEO likes to define as a group of ʻqualified purchasersʼ. Säkom needs to be financially independent in the long term, but does not need to generate profits.

Figure 7.6 The business model for the Säffle fibre network, Sweden

Source: Authors.

The separation of the PIP and NP role makes it viable for operators to join and provide connectivity and service provision, because long-term investment for the infrastructure deployment, which typically has a horizon of five to ten years, is no longer part of their business case. On the other hand, the municipality is in a position to accept a longer-term return on investment and assume the capital investment in infrastructure, while avoiding responsibility for technical issues in which it does not have enough competence. Moreover, the NP contract was awarded to a national operator which can rely on significant economies of scale, bringing down the operating cost of the network.

At the same time, the business case for service providers is enhanced by the existence of infrastructure and connectivity, so they can focus on efficient service provisioning (with cost reduction coming from know-how and economies of scale due to their national or sometimes international scale), customer care, marketing and product development. Efficient service provisioning is especially important for Internet service (increasingly seen as a commodity), where price and reliability are the major selling points, whereas the TV product offering is an important differentiating factor.

Telia was awarded the NP contract with the obligation to provide and connect at least three ISPs, two TV providers, and two providers of prioritized IP telephony (today there are five ISPs, three TV and two priority IP-telephony providers). Telia was not allowed to sell services initially. It was argued that if Telia had been allowed to sell services in the beginning, other SPs would have found it hard to establish a viable position. However, this decision was not very popular with the citizens. Once other SPs had the time to become established and Telia was not considered dominant anymore, this ban was removed.

As part of the business model, the media converter is provided and maintained free of charge by the NP (Telia) and, when buying services other than just Internet, the service separator is also provided and maintained by Telia. First-line support is taken care of by the service providers, although end-users frequently call Säkom (PIP), which nonetheless thinks that this first-line burden is manageable. With time, communication between PIP and NP has improved considerably. Telia as NP has concluded service contracts with outside firms for installation and maintenance (in Säffle and nearby Karlstad).

There are two types of service-level agreements (SLAs) in place: restoring services within twenty-four hours, if fewer than twenty-four users are affected, and restoring services within six hours, if twenty-four users or more are affected. The SLA was invoked when the TV service started to have problems last autumn; these then escalated, and on New Year’s Eve one TV service (from SP Serverado) was down until 3 January. The NP Telia offered three months of free TV as compensation.

The digging cost is EUR 4.50/m on average (varying between EUR 2/m to EUR 11/m in the rural area, and around EUR 35/m in town), which increases to around EUR 8/m to EUR 10/m if connection costs are included. Housing cooperatives and housing companies have to pay EUR 2,200 to connect to the fibre network, while the in-building network is their own responsibility. An agreement was reached whereby housing companies pay EUR 11 per month per apartment and the tenant pays EUR 2 per month to Säkom for 15 years.

In addition to the retail sector, Säkom is offering dark fibre rental to business users and large organizations. Among the current clients are ʻNet for Mobilityʼ (dark fibre rental for ten years), the municipality and the Swedish Church. Businesses that have expressed an interest in dark fibre include the ICA food chain, TDC, and others. Capacity is not really limited, as ninety-six fibres are installed in the backbone network and twenty-four fibres to the nodes.

Currently a price per-kilometre is charged, but other models like price per connection, and prices for specific fibre spans are being considered; some are more popular than others. Large firms are currently connected to the copper network and they are generally bound by long-term contracts. When these run out, they are likely to migrate to fibre.

7.5.3 Norrsken (central-northern Sweden)

The central-northern regions (län in Swedish) of Dalarna, Gävleborg, Västernorrland, Jämtland and Västerbotten represent mostly rural regions with middle incomes and a low population density (generally below ten per square kiometre on average).

The territory is largely rural, generally characterized by medium-size towns along the coast (from 10,000 to 80,000 inhabitants) in which by far the larger portion of the population is concentrated, as well as a large number of smaller localities of a few hundred to a few thousand inhabitants each, inland, at very large distances from each other and generally surrounded by farmland.

7.5.4 Skånet

Skåne is the southermost region (län) in Sweden, and includes thirty-three municipalities (kommuner), the largest being Malmö (301,000 inhabitants), Helsingborg (130,000), Lund (111,000 inhabitants) and Kristianstad (80,000 inhabitants).

Around 130 km long from north to south, Scania covers less than 3% of Sweden’s total area, but the population of approximately 1,230,000 represents 13% of Sweden’s total. Skåne is mainly a hilly territory; with moderate population density (110 inhabitants/km²), it is among the most densely populated regions in Sweden. Skåne is a relatively wealthy region, with intensive farming as well as a thriving high-tech and academic metropole.

Historically, Skåne has had a good penetration of broadband connections provided by DSL. A higher regional goal now applies: 100 Mbit/s to almost everyone in Skåne by 2020, meaning fibre to the home.

7.6.1 The Stokab model

Stokab owns and is responsible for the passive fibre network, while market players operate and deliver services over the network. Stockholm’s basic idea is that IT infrastructure should be available to the whole society – public sector, telecom operators, and other businesses alike. Therefore Stokab’s network is designed to facilitate competition: the fibre network is open to everyone on equal terms.

The aim of the network’s deployment is to create an ICT infrastructure that allows competition by giving telecom operators and other companies and organizations access to the infrastructure. This vision differs from the prevailing opinion in the rest of Europe, where fibre and broadband networks are often considered as networks for telecom operators. Stokab, however, has inspired several municipal and regional fibre networks throughout Europe and beyond, whereby the open access network model is becoming increasingly better appreciated. Stockholm is often cited as a world-class ICT city.

Aside from passive fibre lines, Stokab provides physical space in nodes equipped with power, cooling, etc. Stokab’s fibre network connects almost all multi-dwelling units and commercial properties in the Stockholm municipality: about 90 per cent of households and almost 100 per cent of enterprises have the possibility of signing up for a fibre-based connection.

An extensive backbone network connects industrial areas, all major healthcare facilities and urban centres in the region. The fibre network is available in all parts of the municipality and as an extensive interconnecting network throughout the region. With its 1.25 million kilometres of fibre (in 5,000 km of cables), Stockholm is one of the world’s most developed cities in terms of fibre.

Since the company’s inception in 1994, the passive network structure and the business model have been designed to enable all stakeholders to define their own network structures. The lease of the network can expand as well as shrink based on a player’s need.

In 2012, Stokab had over 100 telecom operators and more than 700 companies and organizations as customers. These can lease fibre directly from Stokab to deliver services in competition. Virtually all telecom operators in Sweden have facilities in Stokab network nodes. National and international fibre connections reach Stokab nodes so that all operators can gain access to links throughout Sweden and the rest of the world, through virtually any operator.

7.6.2 Socio-economic return on investment

Stokab had great importance for Stockholm’s businesses and IT-development. Without Stokab’s fibre network, science parks like Kista, north of town, would probably not have developed into what is today’s success: Kista Science City, for instance, has more than 1,000 ICT companies and around 24,000 employees, as well as 6,800 university students and 1,100 researchers within the field of ICT. It is an attractive environment for ICT companies and developers so it is not surprising that all major IT and telecom companies, as well as universities and research institutes like Swedish ICT, have offices in Kista.

The fibre network has also facilitated innovations and new enterprises such as Spotify and Skype. Media companies have also been able to produce television in a whole new way.

For the past twenty years, Stokab has invested an average of more than SEK 250 million per year, to total of SEK5.4 billion up to 2012. This investment has been possible thanks to the profits generated by Stokab. Break-even was reached in 2001 and the accumulated profit has now passed 1 billion SEK. Until now the profit level has been low compared with the investment level due to the heavy upfront investments required in the initial phase. From the year 2005, the returns have increased steadily, enabling further major investments. It is worth mentioning that the fibre network was built without public funding and was instead financed through loans and revenues.

Through this extensive open fibre network provided by a neutral player, telecom operators can lease and design their own fibre networks without having to make costly investments or having to pay expensive leasing fees to a competitor. Today, to lease fibre in Stockholm costs less (sometimes much less) than half as much as in other capitals around the world. This translates into lower costs, not only for operators but also for all enterprises that have a need for fast and reliable communications. Lower lease prices propagate down the value chain and stimulate entrepreneurship and new service developments.

The fibre network also delivers a wide range of indirect effects to society. It enables, for example, the more effective use of cloud services, videoconferencing, healthcare, distance education, and other bandwidth-hungry services like HD-TV, video on demand and other streaming media. Moreover, innovation power is unleashed when both small businesses and households have access to the same broadband connectivity that previously was only available to large companies.

As the telecom operators compete on equal terms, competition is fierce in Stockholm, which leads to lower prices for broadband compared to cities where competition is weaker. Savings due to lower broadband costs for companies are estimated at approximately SEK 75 million per year if compared to the capital city of Denmark, Copenhagen. The difference becomes even more significant when other, more expensive, European cities are used as the benchmark.

Stockholm’s city housing companies have had a major role in the development of broadband. Early on they adopted a broadband policy to connect their properties to Stokab’s network and to install fibre all the way to each apartment. They are also installing home networks inside the apartments with outlets in every room. Through collaboration models, they have inspired other property owners of multi-dwelling units to join Stokab’s network. The housing companies’ accumulated investment now amounts to nearly SEK 2 billion.

Building a property network also has other merits beyond the mere delivery of broadband services to the tenants. Since the property owners connect all parts of the property, the communications network can also be used for managing, monitoring and measuring the facilities. By connecting multi-dwelling units with fibre, property owners have been able to use control and automation services more effectively (electronic locks, surveillance, etc.), while at the same time being able to raise the rent as the fibre connection has given tenants an added value. Until now, fibre connectivity has led to an increase in use value for the tenants and a higher property value for municipal housing companies in Stockholm (nearly 100,000 apartments) of SEK 1.85 billion, as well as increased rental revenues of over SEK 30 million per year. These effects cover the housing companies’ investments almost in full and are expected to grow in the coming years.

Generally, when building a 4G/LTE network, 70–80% of the total cost derives from deployment of fibre infrastructure. In principle, each base station needs to be connected to fibre to sustain the high 4G/LTE capacity. Leasing the required fibre connections, instead of investing in a private backhaul network, can significantly reduce the deployment cost for 4G/LTE. The world’s first 4G/LTE network was installed in Stockholm. Net4Mobility (jointly owned by operators Telenor and Tele2) states that 4G/LTE would not have been launched in Stockholm if the necessary fibre had not been available to lease from Stokab. Today, four 4G/LTE-networks with extensive coverage are operating in Stockholm.

As Stockholm City and Stockholm County have been able to connect their premises with fibre, it has become possible to purchase data and telecommunications services on an open market. This has generated a cost saving for the municipality and for the county of about SEK 2 billion over the years 1996–2012.

Stokab procures its deployment, operations, materials, planning, etc., from the private market. The procurement process and the large investments made over the years have generated a direct economic activity that is estimated to be over SEK 5 billion in revenues for the supplier industry.

Several studies show that high-capacity broadband leads to growth and the creation of jobs beyond the direct economic activity generated in the supplier industry: e.g., through the development and use of advanced services and products, as well as higher ICT competence, which in turn leads to increased productivity and entrepreneurship. According to Acreo’s econometric model, the ‘job value’ which the fibre network has created in Stockholm is estimated at about SEK 7.7 billion.

8.2.1 Public policies related to broadband deployment

According to German constitutional law (Grundgesetz (GG), 2012), exclusive legislative competence in the field of telecommunications lies with the federal level (Art. 73 (1) Nr. 7 GG). Telecommunications are thereby understood as a technical process of signal transmission with the possibility of reproduction at the place of destination, including the provisioning of technical infrastructure and equipment at the beginning and the end of the transmission line (Nettesheim, Reference Nettesheim and Säcker2009).

Moreover, the federal level has an obligation to provide for blanket coverage of the state territory with basic telecommunications services of adequate quality (Art. 87f (1) GG) – universal service. In doing so, the state assumes responsibility as a guarantor of subsistence and will provide for a legal and regulatory framework allowing for a market supply of necessary infrastructure and services (Nettesheim, Reference Nettesheim and Säcker2009).

Content-related issues of transmitted signals, broadcasting and media law remain the domain of the sixteen German federal states (Bundesländer or Länder for short). This does not mean, however, that the Länder are prohibited from any activity that contributes to telecommunications infrastructure development. On the contrary, as long as they comply with competition law and state aid rules, the Länder and municipalities may take measures that complement and enhance telecommunications provision above the minimum level (Kühling and Neumann, Reference Kühling, Neumann, Inderst, Kühling, Neumann und and Peitz2012).

8.2.2 Definition of broadband in Germany

There is no legal definition of broadband in the German telecommunications law (Telekommunikationsgesetz, TKG). The analysis of various policy documents of the German government leads to the conclusion that broadband is a connection whose transmission rate clearly exceeds that of the Integrated Services Digital Network (ISDN): i.e., 144 kbit/s. Until 2009, for statistical purposes, the minimum transmission rate of 384 kbit/s downstream was used (Breitbandatlas 2009-02), but the current broadband strategy of the federal government introduced a more appropriate and up-to-date notion of ʻ1 Mbit/s broadbandʼ, defined as a minimum of 1 Mbit/s downstream and 128 kbit/s upstream, which is considered a basic level provision (Breitbandatlas 2009–02; BMWi, 2009).

8.2.3 The first phase: market liberalization framework and the beginning of broadband promotion

The German telecommunications law adopted in 1996 did not contain provisions directly dealing with the development of the broadband infrastructure nor mentioned the promotion and support for application of innovative and new technologies (TKG, 1996). Its objective was to promote competition and universal coverage with telecommunications services through sector-specific regulation (Art. 1). The only provisions of relevance for the development of infrastructure as an alternative to the copper lines of the Public Switched Telephone Network (PSTN) of the incumbent were those implementing the European Directive 90/387/EEC of 28 June 1990 (European Communities, 1990), on the establishment of the internal market for telecommunications services through the implementation of Open Network Provision (ONP). According to these provisions, a dominant provider of public telecommunications services and infrastructure had to ensure access to and interconnection with its network for its competitors and users under non-discriminatory conditions (TKG, 1996, Art. 33 and 35). Also, sharing of public conduits or ducts was foreseen in the case where the laying of new telecommunications cables was impossible or required disproportionally high expenses (Art. 51).

The initial roll-out of the broadband networks was complicated by the heterogeneity of ownership of various parts of the public telecommunications network. While the backbone telecommunications network belonged to the incumbent Deutsche Telekom, the ownership of the local loop was not uniform. The local access networks were owned (named in the order of diminishing ownership) by Deutsche Telekom, Deutsche Bank, Bosch Telecom and a number of small and very small network operators (Möschel, Reference Möschel2001). Due to this particularity, coordination of investments was necessary both for the backbone infrastructure and the local loop.

Following the European Commission Directive 96/19/EEC, amending Directive 90/388/EEC with regard to the implementation of full competition in telecommunications markets, the German telecommunications market was fully liberalized on 01.01.1998. Subsequently, Germany unbundled the copper local loop, in compliance with Regulation (EC) No 2887/2000 of the European Parliament and of the Council on Unbundled Access to the Local Loop (European Communities, 2000) and Commission’s Recommendation on Unbundled Access to the Local Loop (European Commission, 2000), thus enabling the competitive provision of a full range of electronic communications services, which includes broadband multimedia and high-speed internet.

Similarly to the European Union (EU) broadband policy (Cava-Ferreruela and Alabau-Muñoz, Reference Cava-Ferreruela and Alabau-Muñoz2005), the development of broadband infrastructure in Germany has been considered a policy measure aimed at the realization of the information society and, therefore, was promoted in subordination to the respective initiatives. This approach of ‘embedded broadband promotion’ was adopted in one of the first, but still relevant, policy measures of the federal government: the ʻInitiative D21ʼ. Initiative D21 is a partnership between the political actors at the federal, regional (Länder) and local levels on the one hand, and interested economic partners from all industry sectors on the other. Projects supported through Initiative D21 should be non-profit and practice oriented and contribute to overall societal development and economic growth.

Simultaneously, the German government led the adoption of broadband technologies creating (additional) demand and applying innovative applications and services. In 1999 an eGovernment initiative for local governance – MEDIA@KOMM – started with the aim to develop and apply multimedia in towns and communities (DIW, 2004). Within the initiative BundOnline 2005, launched in 2000, a modernization of the federal administration was carried out, whereby all governmental services suitable for such provision were transformed into online services (BMI, 2006). The measures appeared to be a success such that follow-up projects were launched. MEDIA@KOMM-Transfer was carried out in 2004–2007 for the promotion of regional and local eGovernment and relevant international cooperation. The ongoing MEDIA@KOMM-Innovation project aims to support the development of regional and local eGovernment networks and provides for the exchange of experience and best practices, as well as for harmonization and standardization of eGovernment solutions.

In March 2002, the Federal Government and the members of the Initiative D21 founded the Deutsche Breitbandinitiative, which is an open platform for dialogue amongst industry, science institutes and the government in order to enable networking and experience and knowledge transfer and to strengthen contacts between the decisive actors. It does not have its own budget, but serves as a cross-sectional project active in all the fields where broadband solutions are possible but not yet implemented (DIW, 2004). In June 2003, the Deutsche Breitbandinitiative adopted an action plan which was to complement the infrastructure development by supporting the creation of broadband services, applications and content, stimulating demand, enhancing security and raising users’ awareness (BMWi & BMBF, 2003). The promotion of an investment-friendly climate was intended to attract private investors, who were to assume the main financial burden of the broadband roll-out.

8.2.4 The second phase: greater consideration of information society requirements

The eEurope 2005 Action Plan for the development of the information society (European Commission, 2002), which focused in part on broadband as a necessary infrastructure to provide high-speed connectivity – and thus stimulate and enable use of advanced applications and services by public and private parties – required all member states to adopt national broadband plans. The German action programme Informationsgesellschaft Deutschland 2006 was adopted in December 2003 and continued the line of embedded broadband promotion (BMWi and BMBF, 2003). To realize the information society, the development of an appropriate infrastructure was deemed indispensable and telecommunications and broadband were identified as the very first spheres of activity of the German ICT policy. The action programme had the objective of making broadband the dominant access technology by 2005 and of reaching a broadband penetration level of over 50% by 2010 (BMWi and BMBF, 2003).

The responsibility for implementation was divided between the Federal Ministry of Education and Research, which deals with technological development, and the Federal Ministry of Economics and Technology, which is in charge of promoting the application of technologies (DIW, 2004). The federal action programme largely embraced the key steps proposed by the Deutsche Breitbandinitiative, notably the central role of private investments in broadband development. It renounced subsidies as distorting competition, making an exception only for financing from the EU structural funds – the European Regional Development Fund and the European Agricultural Fund for Rural Development. Furthermore, the action programme followed the strategy of stimulating the supply by promoting the demand. Internet usage and eGovernment initiatives were to be promoted as part of this programme. In this context, the programme DeutschlandOnline of 2003 became an umbrella project for federal, regional and local levels of governance (Thome, Reference Thome2006).

The new version of the German telecommunications law of 2004, implementing the revised EU framework, aimed at improving the general investment climate for infrastructure projects. Yet, the promotion of broadband and next-generation networks was not singled out as a priority. Rather, the promotion of efficient investments in infrastructure and the support of innovations were listed among the objectives of the law (TKG, 2004, Art. 2). Accordingly, a number of provisions were designed to secure the planning and investments for the undertakings. Article 21 of the TKG, which deals with the imposition of access obligations on dominant infrastructure providers, stated explicitly that, when considering this regulatory measure, the Bundesnetzagentur (BNetzA), the German telecommunications regulator, had to take into account the initial investments of the infrastructure’s owner as well as the necessity to create incentives for further investments in infrastructure in order to promote competition in the infrastructure market. Where access obligations were imposed on a dominant provider, the Bundesnetzagentur had also to regulate its interconnection prices (TKG, 2004, Art. 30(1)). When fixing the price, the regulator had to take into account infrastructure- or service-specific risks connected to the capital invested (Art. 31(4)). All adopted regulatory measures were part of the ex-ante sector-specific regulation, meaning that the intensity of the regulation depended on the existence of effective competition within the market at hand and of significant market power of the network provider in question (Christmann, Ensslin and Wachs, Reference Christmann, Enßlin and Wachs2005).

The subsequent action programme of the federal government iD2010 – Informationsgesellschaft Deutschland 2010, was adopted in 2006 in line with the EU i2010 strategy ʻA European Information Society for growth and developmentʼ (European Commission, 2005) and continued the promotion of information-society-oriented broadband development (BMWi, 2006) applying the two-prong strategy: the supply should be improved and the demand created. By that time, the supply had increased rapidly and the federal government announced that a new realistic objective was household penetration of 98 per cent by 2008, irrespective of the type of network. The creation of an interactive broadband atlas would permit a better assessment of broadband coverage and demonstrate broadband access potential for the users, industry and policy-makers. Further studies of broadband development and dialogue within the Deutsche Breitbandinitiative were to pursue the same objective. On the demand side of the market, a steady growth was predicted, such that a 50% level of uptake by German households was expected well before 2010. It could be enhanced by further innovation related to services and content and be supported by creating the appropriate conditions for it: the digitalization of information channels and the convergence of the media.

In 2007, the so-called ʻregulatory holiday’ provision was introduced as Article 9a of the TKG, intended to directly promote broadband development. Accordingly, new markets were not to be regulated unless there were sufficient grounds to assume that lack of regulation would hinder development of sustainable competition in the market in question in the long term. The regulatory holiday provision exempted Deutsche Telekom from having to grant competitors access to its new VDSL network. The new market definition encompassed those markets for services or products that differed substantially from services and products already available and did not merely replace them (TKG, 2004, Art. 3 Nr. 12a), thus embracing both various broadband and new generation technologies and services (BMWi, 2009).

The European Commission raised objections against this provision, claiming violations of Articles 6 to 8(1) and (2), 15(3) and 16 of Directive 2002/21/EC of the European Parliament and of the Council of 7 March 2002 on a common regulatory framework for electronic communications networks and services (European Communities, 2002b) and violation of Article 8(4) of Directive 2002/19/EC of the European Parliament and of the Council of 7 March 2002 on access to, and interconnection of, electronic communications networks and associated facilities (European Communities, 2002a), as well as a violation of Article 17(2) of Directive 2002/22/EC of the European Parliament and of the Council of 7 March 2002 on universal service and users’ rights relating to electronic communications networks and services (European Communities, 2002c). The Court of Justice of the European Union upheld the Commission’s claim, finding that the amendment to the TKG meant an undue legislative interference into regulatory competences of the national regulatory authority that seriously restricted regulatory discretion of the Bundesnetzagentur regarding the market definition and market analysis (CJEU, 2009). As a consequence, Art. 9a had to be removed in the new version of TKG in 2012.

8.2.5 The third phase: intensified and targeted policy of broadband promotion

After almost ten years of subordination of broadband development to the needs of the information society, the so-called broadband strategy of the federal government adopted in 2009 ‘broke ranks’ in terms of its approach: it deals exclusively with broadband development, leaving other information society topics outside its scope (BMWi, 2009). Another novelty is the differentiation between the broadband data rates (1 Mbit/s, 2 Mbit/s and 50 Mbit/s). The federal government set two objectives, which correspond to the measures to strengthen the German economy declared in the second recovery package approved in January 2009 (Pakt für Beschäftigung und Stabilität in Deutschland zur Sicherheit der Arbeitsplätze, Stärkung der Wachstumskräfte und Modernisierung des Landes, – Bundestag, 2009):

1. 1. to close the gaps (‘white areas’) in broadband coverage and to ensure an ubiquitous availability of 1 Mbit/s broadband by the end of 2010 and

2. 2. to provide a 50 Mbit/s broadband access to 75 per cent of all households by 2014. This will constitute a basis for further fast deployment of high-data rate connections for the whole territory.

To realize the objectives of the German broadband strategy, several measures were proposed, to be kept updated and adjusted depending on the progress of the strategy’s implementation:

1. For example, the existing infrastructure and facilities will be shared in order to avoid duplication of infrastructure.

2. A broadband atlas and a database on all relevant building sites and projects will be developed.

3. Furthermore, demand-oriented empty ducts will be laid and construction of telecommunications infrastructure will be undertaken in cooperation and coordination with other utility sectors (for instance, water supply).

4. The described measures will be carried out in close cooperation with the Länder and the local authorities and this teamwork will be strengthened through the establishment of working groups within the Federal Ministry for Economics and Technology, consisting of representatives of the Länder and the federal government.

In this context, financing through the so-called ʻjoint tasksʼ has been improved. Already in 2008, the financing of broadband infrastructure roll-out became possible for rural areas, where broadband development was not commercially profitable, within the Joint Task for the Improvement of Agrarian Structures and Coast Protection (GAK) of the Federal Ministry for Food, Agriculture and Consumer Protection. The Länder can propose their own measures to be included in the GAK framework at the federal level, containing principles and general requirements for integrated rural development. When approved, the framework plan is to be implemented at the Länder level (GAK-Gesetz, Art. 7-9). This plan expired at the end of 2013 and foresaw in Part B financial support for rural areas that lack or are short of sufficient broadband provision and can prove business and household demand (GAK-Rahmenplan, 2011). The respective local authorities or associations could apply for financing through subsidies, grants or loans from the federal government. With these means they can complement private and communal investments in projects for infrastructure roll-out and accompanying activities (planning, consulting, informational events, etc). In the new framework the laying of empty ducts has been rendered eligible for financing and the federal subsidy ceiling has been increased from 60% to 90% with the remainder provided by the Land. The limit of the subsidy per project was also raised up to EUR 500,000.

The laying of empty ducts is perceived to be a decisive measure for the development of high-speed communications networks. For its promotion, a special umbrella framework has been created at the federal level that provides general conditions for subsidies at the local level and eliminates the necessity to notify the Commission (Bundesrahmenregelung Leerrohre, 2011) of every single community measure. It was approved by the Commission in July 2010 and provides EUR 600 million over the period of five years (until 2015) for development of passive infrastructure in ʻwhiteʼ and ʻgreyʼ areas.

Additionally, since mid-2009 structurally weak areas, especially in eastern Germany, can rely on financing through the Joint Task for the Improvement of Regional Economic Structure (GRW) of the Federal Ministry of Economics and Technology for the development of a high-quality broadband network that provides a minimum service of 2 Mbit/s. Similar to GAK, a coordination framework defining eligibility for financing areas and measures was adopted to be implemented by the Länder (GRW-Gesetz, Art. 4-6). The conditions for a subsidy require a higher data rate (at least 2 Mbit/s) and evidence of an inadequate cost-performance ratio in comparison to adjacent urban areas (GRW-Koordinationsrahmen, Attachment 4, Number 3.2.4.).

It appears that the available funds have been actively used. The EUR 10 million provided annually by the federal state under the GAK was overdrawn in 2011 and 2012 (EUR 19.2 million and EUR 14.7 million, respectively) which became possible due to the remaining sums from the early years of the fund and thanks to excess payments from the Länder. Due to this success, the fund, originally to end in 2013, will continue its operation until 2018 (Goldmedia, 2013). Funding via GRW has had a more modest success (2012 with EUR 2.9 Million was the peak year) because GRW financial means are committed not to broadband development but to development of structurally weak regions in general and are used for various projects (Goldmedia, 2013).

Under the second recovery package, which expired at the end of 2010, additional financing was available for the cases not falling under the Joint Tasks (Börnsen, Reference Börnsen2009). The attractive features of this mechanism were the absence of the subsidy limit per project and the possibility to apply for regional projects not being carried out by local authorities.

One more source of financing has been established directly for undertakings: according to the Law on Future Investments (Zukunftsinvestitionsgesetz – ZuInvPG, 2009) the Länder can provide local authorities with finances which they in their turn can make available to undertakings developing or operating broadband networks. The Länder have the competence to decide whether and under what conditions financing provided by the Law on Future Investments will be used. For instance, undertakings may apply for guarantees by the federal government and/or Länder that assume a default risk of up to 90 per cent.

Besides the joint financing, some Länder (Bavaria, North Rhine-Westphalia, Rhineland-Palatinate, Schleswig-Holstein and Thuringia) have their own financing programmes that offer cheap long-term loans allowing 10 to 20 years return on investment (Goldmedia, 2013). It will also be noted that the aforementioned EU structural funds – European Regional Development Fund (founded in 1999) and European Agricultural Fund for Rural Development (founded in 2005) – present an additional funding possibility which is, yet, of a secondary character in relation to the national financing opportunities (BMWi and BMELV, 2009).

The latest amendments to the German telecommunications law of 2012 follow the line of the governmental strategy to strengthen the promotion of broadband development and declare the ʻacceleration of development of heavy-duty public telecommunications networks of next generationʼ as one of the major objectives of telecommunications regulation (TKG, 2012, Art. 2(2) Nr. 5). Having dropped the ʻregulatory holidayʼ provision, TKG 2012 relies strongly on enhancing security of planning and investment for undertakings, thus further developing the approach launched in TKG 2004, as well as on facilitating risk sharing and cooperation among investors (Körber, Reference Körber2011).

Over time, the regulatory framework has stabilized and become more precise, thus enhancing the security of planning and investment. The law aims at objective, transparent, non-discriminatory and proportional regulatory principles underpinning the work of the Bundesnetzagentur, which will be substantiated by maintaining the same regulatory concept over longer appropriate periods of time, to increase predictability of regulation, and by considering investors’ risks in case of access obligations, by allowing for risk-sharing agreements between undertakings, by prohibiting discrimination and promoting competition, as well as by recognizing the complementary and transitional character of regulatory measures in relation to competition (TKG, 2012, Art. 2(3)).

A more straightforward measure to improve the investment climate is the possible adoption by the Bundesnetzagentur of periodic administrative rulings (Verwaltungsvorschrift) describing the general regulatory requirements. To assess efficient investments and innovation regarding new and enhanced infrastructure, a methodology for risk determination, criteria for establishment of access conditions and prices for risk-sharing models, and examples of risk-sharing models will be considered by the regulator (TKG, 2012, Art. 15a(2)). On the basis of such administrative rulings and upon request of any provider of a public telecommunications network contemplating development of new-generation networks, the Bundesnetzagentur is to provide region-specific information about expected changes in the regulatory framework or measures (TKG, 2012, Art. 15a(4)). To further improve consistency and predictability of regulation, as well as the security of planning and investment for undertakings, the Bundesnetzagentur has the potential to expand market regulation periods from three to a maximum of six years, provided the European Commission has no objections to such prolongation (TKG, 2012, Art. 14(2)).

Next to the already mentioned provisions of Article 15a(2) facilitating risks sharing, the regulatory measures on the basis of Articles 30 and 32 of the TKG create additional incentives for this practice. When regulating the prices charged by dominant network providers for access (both services and facilities), the Bundesnetzagentur will consider initial investments and allow for an appropriate return (risk premium) upon the invested capital, as well as take in utmost account special risks and risk-sharing agreements between undertakings regarding next-generation networks (TKG, 2012, Art. 30(3) and 32(3) Nr. 3).

In addition to sharing investment risks, TKG 2012 seeks to promote actual infrastructure sharing between undertakings. Previous legislation dealt with only one aspect of these issues, namely with cooperation when laying ducts and wires. According to Art. 21(2), the Bundesnetzagentur, acting upon a request or ex officio, may impose on a dominant provider of public telecommunications networks, among other obligations, the requirement to ensure open access to certain network elements or facilities, including unbundled broadband access. At the same time, a number of obligations are to be imposed on a dominant provider under Art. 21(3), including access to passive network elements and open access to technical interfaces, protocols and other key technologies necessary for services or interoperability of services, as well as collocation and other types of sharing of buildings, ducts, poles and other facilities.

In order to achieve the targeted data rate of 50 Mbit/s and more, and to prevent signal degradation, it is necessary to at least partially upgrade the copper wire between the distribution frame in the local exchange and the subscriber terminal with fibre. To do this, the new Art. 77a (TKG 2012) guarantees access to the relevant network infrastructure on the premises in two ways: either by collocation of wires of competing providers in the same conduits or by sharing the existing wires if duplication of the infrastructure is economically inefficient or technologically impossible (Kind and Schramm, Reference Kind and Schramm2012). The Bundesnetzagentur’s order to co-locate or share the wires and conduits in question can be addressed not only to network providers but also to other owners of (passive) infrastructure who do not operate the networks. Moreover, the imposition of the obligation does not depend on the market power of the addressee but has the objective of achieving synergy effects in relation to the network infrastructure (TKG, 2012, Art. 77a(1); Kind and Schramm, Reference Kind and Schramm2012). The Bundesnetzagentur can create an inventory of all facilities that can be used for telecommunications purposes (for instance, ducts, poles, antennas, etc.) and can request the relevant information from telecommunications providers and other legal entities in possession of such facilities (TKG, 2012, Art. 77a(3)).

In order to further minimize the costs of the development of new broadband networks, TKG 2012 allows the collocation of telecommunications infrastructure with other facilities (water, gas pipes, power lines). Upon a written request from an interested provider of public telecommunications networks, undertakings and legal persons in possession of facilities that can be used for the development and laying of next-generation networks are obliged to make an offer to share such facilities (TKG, 2012, Art. 77b(1)). Additionally, Art. 77c-77e provide for sharing of roads, waterways and railway infrastructure belonging to the federal government.

The investment climate for broadband development is further improved by initial regulation of the deployment of innovative laying techniques. Commonly, telecommunications infrastructure will be laid in compliance with the standards set by the German Institute for Standardization (Deutsches Institut für Normung, DIN) in the General Technical Instruction for the usage of roads for cables and telecommunications lines (Allgemeine Technische Bestimmungen für die Benutzung von Straßen durch Leitungen und Telekommunikationslinien, ATB-BeStra) which prescribes, for instance, a certain depth and width of trenches for the wires. Art. 68(2) of TKG 2012 provides for an exhaustive list of requirements to be fulfilled for a request to be approved by the carrier of public easement allowing the requesting provider to lay fibre infrastructure with the help of micro- or mini-trenching techniques in deviation from the aforementioned DIN standards.

An important measure to promote innovative broadband technologies is the decision of the Bundesnetzagentur of Summer 2013 to slightly ease the incumbent’s obligation to provide access to the sub-loop segment of the network. Now Deutsche Telekom may refuse access to the sub-loop if it is necessary to enable implementation of vectoring at the street cabinet (by Deutsche Telekom itself or another company). All competitors will still be able to interconnect at the cabinet using optical fibre and implement vectoring if they offer an appropriate bitstream product under open access arrangements.

8.2.6 Broadband-related radio frequency spectrum measures

Like other national policies in the radio spectrum field, German policy is subject to the limitations imposed by decisions about radio frequency allocation and uses adopted at the international (International Telecommunications Union, ITU) and regional (European Conference of Postal and Telecommunications Administrations (CEPT)) and EU institution levels (Bundesnetzagentur, 2013a; TKG-Kommentar, 2013). Implementing the international allocation of spectrum frequencies, the federal government adopted a national frequencies regulation that allocates certain frequency bands for certain uses. On its basis, the Bundesnetzagentur draws up a detailed spectrum frequency plan which is then used to assign frequencies to individual operators (TKG 2012, Art. 52 ff.). Furthermore, physical characteristics of spectrum frequencies impose further constraints on their use for broadband provision. Due to this, measures in the field of spectrum frequencies are dubbed ʻsupporting policyʼ (ʻUnterstützende Frequenzpolitikʼ), meaning that they will be used in order to close the gaps in the provision of broadband in rural areas and in order to improve mobile use of ICT (BMWi, 2012).

Arguably, the relevance of radio spectrum for high data rate transmission has become apparent with the development of the Universal Mobile Telecommunications System (UMTS) standard and its commercial deployment. In contrast to the second-generation standard (Global System for Mobile Telecommunications, GSM), in addition to voice UMTS would allow transmission of data and video at a much higher rate of up to 2 Mbit/s, with all participants constantly sending data packets on the same frequency band. This led both operators and policy-makers to believe that in the near future UMTS would replace not only GSM but also parts of the landlines used for Internet provision (Petzel, Reference Petzel2006). These and other factors played a decisive role in the auction for UMTS licences in summer 2000, total proceeds of which were an exceptionally high sum of about EUR 50 billion paid by six winners. The licences included an obligation to deploy an own UMTS network by 2003 and to reach a theoretical supply rate of 25% of population by the end of 2003 and 50% by the end of 2005 (Petzel, Reference Petzel2006). The auction was strongly criticized by its participants, scholars and other observers and the ʻUMTS euphoriaʼ soon died away as the winners struggled – and ultimately failed – to fulfil their licence obligations (Virnich, Reference Virnich2003; Petzel, Reference Petzel2006).

Radio frequency spectrum policy has come to play an increasingly important role again during the third phase of broadband policy development due to the launch of the commercial use of the Long-Term Evolution (LTE) standard. It has been recognised that mobile and wireless solutions in general are instrumental in closing the gaps in the provision of basic broadband to rural and remote areas. Their effective application depends, however, on the availability of spectrum and its efficient use.

In May 2010, promptly following the requirements of the EU law (European Commission, 2010a) and in compliance with the broadband strategy, Germany auctioned the 800 MHz frequencies for mobile operators. Herewith, Germany was the first EU member state to make use of the Digital Dividend. As a result of the auction, in autumn 2010 new licences were issued for more than 11,000 locations to use LTE base stations. The licence conditions obliged the operators to provide mobile broadband in a particular priority order, starting with towns and communities short of infrastructure of less than 5,000 residents (Bundesnetzagentur, 2013a; Goldmedia, 2013).

Germany intends to stay at the forefront of the exploitation of the radio spectrum for broadband development. Thus, in the context of the discussions in the ITU about reallocation of frequencies of 700 MHz spectrum (Digital Dividend II), the federal government launched a discussion process, established a Mobile Media 2020 Forum and developed an ICT strategy in order to identify the demand for frequencies and their possible use (BMWi, 2010; 2012). In this context, in 2013 the Bundesnetzagentur launched public consultations on the basis of the discussion paper on the process of assignment of 700 MHz and 1.5 GHz frequencies as well as the 900 MHz- and 1800 MHz frequencies, of which the rights of use expire in 2016 (Bundesnetzagentur, 2013b). Besides, the regulatory authority drafted scenarios of possible future allocation of these frequencies and started a formal procedure for demand assessment for the time after 2017 (Bundesnetzagentur, 2011; 2012).

8.3.1 Background about the incumbent operator

The incumbent operator, Deutsche Telekom, belonged originally to the Deutsche Bundespost, a public institution created in 1947 which was in charge of postal and telecommunications services in the later Federal Republic of Germany. Before that time, the Deutsche Bundespost was called the Reichspost. In 1989 the Deutsche Bundespost was divided into three public companies: Deutsche Bundespost Postdienst for the postal service, Deutsche Bundespost Postbank for the postal banks and Deutsche Bundespost Telekom for the communications service. The Deutsche Bundespost Telekom became the Deutsche Telekom AG in 1995 and was subsequently privatized in 1996. Currently, the German federal government has a direct ownership of 15% of the shares of Deutsche Telekom and owns further 17% indirectly through the Kreditanstalt für Wiederaufbau (KfW) bank.

Traditionally Deutsche Telekom has been the company that invested in telecommunications assets. But, when comparing the investment made in fixed, mobile and cable networks which provide broadband, telephony and video services, the alternative operators combined have invested more than Deutsche Telekom in every year of the period 2002–2013. For the period considered, the alternative operators have contributed 53.7% of the total investment, whereas Deutsche Telekom has provided 46.3% (Dialog Consult-VATM, 2013).

8.3.2 Differences between the new and old Bundesländer

After the reunification of Germany in 1990, the government made important efforts such that the PSTN infrastructure in the new Bundesländer and in the old Bundesländer would become similar. But there are still differences between the regions as regards broadband infrastructure. For low-capacity broadband services, which have a data rate up to 1 Mbit/s and which are available through fixed and wireless networks such as Digital Subscriber Line (xDSL), Wireless Local Area Network (WLAN), LTE, UMTS, cable, Fibre to the X (FttX) and Worldwide Interoperability for Microwave Access (WiMAX), all the Bundesländer had a broadband household coverage of more than 98.5% in 2013 (Breitbandatlas, 2013).

However, for broadband services of more than 50 Mbit/s which can be provided through xDSL, Fttx, cable and WLAN networks, there are differences. In 2013, ten of the old Bundesländer had a household broadband coverage of 50 to 95%, whereas only one old Bundesland had a coverage of 10 to 50%. The five new Bundesländer had a broadband availability of 10 to 50% (Breitbandatlas, 2013).

8.3.3 Fixed broadband market

Competition in the fixed broadband market

The predominant network in the fixed broadband market in Germany is the copper wire-based access network. Figure 8.1 depicts the different types of fixed access used for the period 2002–2013. The figure reflects the xDSL, coaxial cable and fibre-based networks. Other types of access networks, such as power-line- and satellite-based networks, were not included due to the limited number of broadband lines deployed.

Figure 8.1 Fixed broadband connections by technology, Germany 2002–2013

Data sources: Based on Bundesnetzagentur (2013c); Dialog Consult-VATM, 2013.

In 2002 the only type of fixed broadband deployed was through copper lines. In 2013, 80% of broadband lines were based on xDSL. The average annual increase between 2002 and 2009 was 38%. The incumbent operator is the provider of the majority of xDSL lines, including the wholesale lines used by the alternative operators.

In recent years the cable networks have been playing an increasingly important role. For the period 2007–2013 the average annual increase of cable access lines was 35%. With only 0.3 million homes connected with Fibre to the Home/Building (FttH/B) networks, the fibre-based access networks represented only 1% of the broadband lines in 2013.

At the beginning of 2007, the xDSL lines amounted to around 95% of the fixed broadband access lines. There were six major DSL providers in Germany: Deutsche Telekom with 48.2%, United Internet with 14.9%; HanseNet (which later became Telefonica) with 13.3%; Arcor (which later became Vodafone) with 13.1%; Freenet (later 1&1) with 7.4%; and Versatel with 3.2% market share (DSLWEB, 2007).

In 2013, the total number of all types of fixed broadband lines was 28.3 million. Deutsche Telekom had still the majority of broadband access lines with a 43.8% market share, followed by 1&1 with 12.0%, Vodafone with 10.6%, UnityMedia Kabel BW with 8.5%, Telefonica with 8.1% and Kabel Deutschland with 7.1%. EWE Gruppe, Versatel, NetCologne and M-Net had less than 3% market share each (Dialog Consult-VATM, 2013).

Cable operators play an increasingly important role in broadband provisioning: in 2006 only 3.4% of the broadband lines were provided by cable operators, whereas in 2013 this has increased to 18.9% of the lines.

8.3.4 Mobile broadband market

8.3.5 Fixed versus mobile broadband

As a result of the deployment of the 3G and 4G LTE wireless networks, and the increasing use of smartphones, the usage of mobile broadband services has grown. As an illustration, Table 8.1 shows the development of the monthly volume of data on fixed and mobile access lines and the corresponding percentage of increase. While the volume of data transmitted through the fixed networks is much higher than that transmitted using the mobile network (15 GB vs. 261 MB in 2013), the growth in mobile data use is much higher.

Table 8.1: Monthly volume of data per user and annual increase, Germany, 2008–2013

Data source: Based on Dialog Consult-VATM, 2013.

It remains to be seen to what degree the deployment of LTE networks, which provide higher data rates, will continue motivating users to replace their fixed broadband access lines with mobile broadband access.

8.5.1 Operators

The fixed broadband market is characterized by an extensive use of xDSL lines, dominated by the incumbent operator Deutsche Telekom. Alternative operators have been relying on the regulated unbundled lines of Deutsche Telekom for competition. Infrastructure-based competition is provided by cable operators who own their own access networks. By using DOCSIS 3.0 cable operators are able to provide a high-speed broadband service and they have been able to realize the highest growth rate in fixed broadband lines in recent years.

In the wireless broadband market four operators own physical wireless access infrastructure and the competition between them is quite strong.

8.5.2 Main features of policy developments

Cooperation between the federal, regional and local levels of government has proven to be the central element of the policy measures aimed at broadband development in Germany. Coordinated efforts were first made to boost the demand side in the early years of broadband deployment, while the supply side was expected to be driven more or less by market forces alone, once the competition had been introduced and some general anti-competitive regulatory safeguards related to the market power of the incumbent had been adopted. However, over recent years governmental efforts have become increasingly concentrated on support of the supply side and range from incentive measures to promote and facilitate investment (risk and infrastructure sharing, consideration of investment for price-setting measures, etc.) to the provision of financial support in various forms. Market failures in rural areas have required an even more direct public involvement in the form of infrastructure and services provision by Stadtwerke (communal enterprises) and laying of empty ducts by municipalities for their subsequent rental to the broadband providers. At the same time, demand-promoting measures have retreated to the background, so that currently German policy can be called a bit one-sided.

These developments indicate a possible (re)turn of German telecommunications policy towards industrialism, at least with regards to fixed broadband development.

Throughout most of the period considered, Germany has been a close follower of the policies developed by the European Commission in Brussels. However, realizing the growing importance of broadband for Germany’s competitiveness and its status as a European and global economic power has resulted in attempts to anticipate – and even determine to a degree – the EU’s information society policies and broadband policy in particular. For instance, Germany was hard on the Commission’s heels auctioning the first Digital Dividend and nowadays it is preparing most thoroughly for the second round.

Against this background, one should consider the fast growing importance of the radio spectrum policy as part of the broadband developments during recent years. So far, the policy measures in this field are very different from the measures for the promotion of fixed broadband: the federal government is mostly concerned with the provision of the scarce resource of radio frequencies.

8.5.3 Expectations based on theory

Taking the theory of the ladder of investment as a way to analyse the level of competition achieved by the different operators in the fixed network market, the following points can be observed:

1. Alternative operators have not had the motivation to deploy their own fixed broadband access infrastructure. The conditions to use the copper lines of the incumbent operator were quite clear and attractive, and hence they have been using the local loop of Deutsche Telekom to provide broadband services. This has brought competition to the services market.

2. Introduction of new broadband technologies (vectoring), extending the life of copper wire is likely to cement this situation and further postpone the emergence of infrastructure-based competition from alternative operators.

3. Cable operators are using the access networks that were initially deployed to provide TV services. By employing DOCSIS 3.0, they are able to provide a high-speed broadband access service. Cable networks are not regulated: that is, they are not subject to any access or price regulation. Increasing infrastructure-based competition has been the result.

4. There is limited deployment of FttH/B networks. Neither the incumbent operator nor the alternative operators have identified an attractive business case to make the necessary investment in large-scale roll-out of fibre-based networks.

8.5.4 Lessons learned

The following lessons can be derived from the case analysis made:

1. – Our analysis shows that even though German telecommunications policy and regulation are technologically neutral, they impact the deployment of xDSL-, cable- and fibre-based networks in different ways. While policy measures seem to respond beneficially to the needs of xDSL and cable providers, fibre providers – both the incumbent and alternative operators – lack motivation and resources to roll out infrastructure.

2. – The policy and legal framework has failed to provide sufficient measures for alternative operators to enter into infrastructure-based competition with the incumbent. However, the latest (2012) amendments to the German Telecommunications Law are intended to address this issue; their full effect is yet to be seen.

3. – Setting high data rate targets as part of the broadband strategy – i.e., pushing beyond the capabilities of xDSL networks – could be instrumental in promotion of fibre and cable development. However, the recent development in vectoring and bonding technologies will extend the life of the copper lines. It may move the alternative operators down the ‘ladder of investment’, as unbundling becomes more complicated and potentially less attractive, driving alternative operators to procure bitstream-like access instead. With this development, effective infrastructure-based competition becomes more unlikely in the near future.

4. – With about half of all households being covered by cable networks, infrastructure-based competition nevertheless is likely to grow as users demand increasingly higher data rates.

5. – The market strength of the incumbent operator remains significant and, thus, justifies continued asymmetrical regulation.

6. – Mobile broadband networks play an increasingly important role due to the higher data rates they can provide, but at the moment they are predominantly used to ensure basic broadband coverage (1 Mbit/s) and to close the gaps in fixed broadband provision. Fixed access networks continue to provide much higher data rates. Moreover, mobile networks are becoming increasingly dependent on fibre-based fixed backhaul networks.

9.3.1 The market before the review

The UK first introduced competition in the telecoms market in 1984 with the licensing of Mercury Communications as the sole competitor to BT in domestic markets. In return Mercury entered into commitments to develop a domestic trunk network, but relied on BT for local access, except in some business districts. It used the Cable & Wireless international network for overseas calls. In 1985, Oftel determined interconnection prices between the two firms that were initially favourable to Mercury, as a method of supporting entry.

In 1991, following the UK government’s review of the duopoly, this period of ʻmanaged competitionʼʼ drew to an end as other firms were permitted to enter the market. The end of the duopoly allowed cable companies, until then permitted to offer telephony only in partnership with Mercury or BT, to enter the market providing infrastructure-based competition to BT, which was Oftel’s preferred form of competition.

A further policy designed to support infrastructure competition was differential access prices for service-based and infrastructure-based competitors. The former had to buy inputs from BT at discounted retail prices, whilst the latter could buy at lower wholesale rates that were the same as BT charged internally¹⁵.

By the time of the review, the UK telecoms market was more competitive than most other markets at the retail level. BT had a market share of around 25% in the retail broadband market, 82% of exchange lines and 60% of fixed traffic¹⁶. BT’s strength in the calls market came from ʻother callsʼ (including calls to free dial-up internet access), which accounted for 47% of all minutes and in which BT had a 70% market share. By contrast, BT had only a 34% share of international call minutes, although such calls were only a small part of the market (just 2.3%).

9.3.2 The review

As early as 2002 some players in the telecoms market began to lobby for a strategic review of telecoms regulation and the break-up of BT¹⁷. Those who argued for a review believed that the market was being held back by the vertically integrated nature of BT and that separation would increase dynamic competition in retail markets. They argued that investment by competitors was being held back because they expected BT to use its integrated structure to sabotage any product development by rivals, thus preventing investors from earning a reasonable return on any investment. It was not necessary for BT actually to have harmed its rivals for them to change their behaviour; all that was required was an expectation of discrimination to deter investment. Although there is little hard evidence to support the claim that investment was being held back, the very low uptake of LLU can be seen as backing up the claim. In July 2004, there were just 13,000 LLU lines out of a total of 4.4 million broadband access lines.

Ofcom launched the TSR in April 2004 with a first phase consultation document¹⁸ in which it set out the purpose of the TSR as to:

Ofcom set out its analysis of the sector looking both at the level of competition at the time and towards the future with an analysis of technology trends. Stakeholders were asked ʻfive fundamental questionsʼ and sixteen more detailed questions. The five fundamental questions were:

In the Phase 1 consultation document, Ofcom found a mixed picture of benefits to UK consumers. It found, for example, that whilst there was plenty of competition in the fixed voice market, most of this was based on service provision and that BT still provided most access infrastructure, despite Oftel’s policy of infrastructure competition. Similarly, BT dominated the broadband access market at wholesale level, although the retail market was competitive¹⁹. Ofcom’s overall conclusion, therefore, was that the 20 years of competition prior to the TSR resulted in only partial benefits for UK consumers in both residential and business markets.

Responses were received from over eighty interested parties, including fixed and mobile operators, consumer representative organizations, independent experts and individuals with no affiliation. Although the responses were wide ranging, a central theme to emerge was the problem of discrimination and the ineffectiveness of the legal/regulatory regime to prevent such behaviour. Some operators went further and argued that both Article 10AD and the UK law allowed dominant operators to discriminate by not providing equivalent products in equivalent circumstances.

In its response to the TSR, Cable & Wireless, then the UK’s second-largest fixed line operator and which had absorbed Mercury Communications in 1997, stated:

Energis, another competitor to BT which has subsequently been acquired by Cable & Wireless, discussed the problem of ʻundue discriminationʼ. It stated:

Energis was established in 1992 by the national electricity transportation network, National Grid. It used National Grid’s network to develop a fibre-optic trunk network which formed the basis of its offering to business customers. Energis, through its subsidiary PlanetOnline, worked with the electronics retailer Dixons to create Freeserve, an Internet service provider (ISP) providing free dial-up Internet access based on an 0800 number. At its peak in 2000, Energis had a stock market valuation of £10 billion, however in July 2002 it was placed into receivership and was acquired by Cable & Wireless in 2005 for a little under £600 million.

The essence of these responses was that preventing discrimination was not enough when the obligation of ʻno undue discriminationʼ allowed justifiably different treatment by the dominant firm of its own downstream business and that of its competitors and allowed different treatment when there was no material effect on competition.

Cable & Wireless’s example of LLU provides a good example. LLU allows competitive operators to rent the copper local loop that runs between the local exchange and the customer premises. The LLU customer needs to install its own equipment in the local exchange to allow broadband signals to be sent over the local loop. It can then sell that service to consumers. BT, like other incumbent operators in their own countries, does not use LLU to provide broadband access themselves. So BT’s retail division was buying a different product from its competitors²³.

There was a similar concern with narrowband or traditional voice access products. The externally supplied product is known as wholesale line rental (WLR) and allows a downstream competitor to rent from BT a local exchange line conditioned for voice services. BT itself did not use WLR to provide voice services at the time of the TSR.

Referring back to the legal definitions of discrimination, BT could well argue that differences between internal and external cost and terms were justified and that therefore they were not discriminating under the definition of ʻundue discriminationʼ.

However, such a defence was unnecessary as no discrimination cases were successfully brought against BT: indeed Ofcom did not find explicit evidence of discrimination during its review. What became clear, however, was that competing communications providers (CPs) lacked confidence in a system that allowed BT to duly discriminate as evidenced by the paragraphs from the Cable & Wireless and Energis responses quoted above. The expectation of different treatment was enough to change the behaviour of downstream competitors.

On 18 November 2004, Ofcom issued its Phase 2 consultation document.²⁴ This reviewed the comments received from the first phase and put forward specific proposals for future regulation of the electronic communications market.

Central to Ofcom’s analysis in Phase 2 was the concept of ʻenduring economic bottlenecksʼ (para. 1.17) which it described as those areas of the network where ʻeffective, infrastructure based competition is unlikely to emerge in the medium termʼ. In possibly the most damning paragraph in the Phase 2 consultation, Ofcom said that competing operators who rely on BT for access ʻhave experienced twenty years of:

Ofcom concluded that the ʻno undue discriminationʼ remedy by itself had proved inadequate to address the competition problems caused by economic bottlenecks and that a stronger remedy was needed. It partially laid the blame at the door of its predecessor, Oftel.

In the last sentence of this quote, Ofcom discusses what has been referred to as ʻcumulative materialityʼ. This is the idea that it is possible for there to be many minor differences between an internal and an external wholesale product which, when each difference is taken alone, appear unimportant but which when they have a cumulative impact can result in a significant disadvantage for the external customer.

9.3.3 Equivalence of input

Ofcom’s principal proposal arising from the TSR was to strengthen the non-discrimination remedy by requiring what it termed ʻreal equality of accessʼ which would prevent BT having justifiable reasons for providing different services internally and externally. This would require ʻequivalenceʼ at the product level and clear behavioural changes by BT.

At the product level, Ofcom stated that equality of access implies BT’s wholesale customers should have access to:

1. the same or a similar set of regulated wholesale products as BT’s own retail activities;

2. at the same prices as BT’s own retail activities; and

3. using the same or similar transactional processes as BT’s own retail activities. (para. 1.36)

Ofcom termed these characteristics of equality of access ʻequivalence of inputʼ. One purpose of the proposal was to strengthen the incentives for BT to provide fit-for-purpose wholesale products without intrusive regulation.

Ofcom also stated that it was important that there is equivalence throughout the product development process and product life cycle. It implied that BT’s wholesale customers have the same ability as BT’s retail activities to introduce changes or have problems addressed.

The final stage of the TR was the issuing by Ofcom of a ʻStatementʼ including a set of undertakings by BT in lieu of a reference under the Enterprise Act 2002²⁵. Paragraph 2 (Definitions and Interpretation) of Annex A sets out what is meant by Equivalence of Input

Since the signing of the original Undertakings a number of amendments have been introduced and brought together in a consolidated version. In this consolidated version ʻsameʼ is helpfully defined as meaning ʻexactly the sameʼ.

Whereas the non-discrimination requirement left room for some degree of ambiguity, the definition of EOI makes it clear that BT must provide exactly the same products internally and externally under the same conditions, etc.

The list of products to which EOI was applied is set out in paragraph 3.1 of the Undertakings. These are:

These products existed at the time of the TSR and were offered by BT in the wholesale market. Thus it can be argued that they had to be reverse engineered to be offered under EOI terms.

However, the Undertakings also commit BT to providing certain (at the time) future services on an EOI basis (para. 3.1). These are listed as:

1. a) Wholesale Extension Service Access Product;

2. b) Wholesale Extension Service Backhaul Product;

3. c) Wholesale End-to-End Ethernet Service;

4. d) IP based Bitstream Network Access products that are the successors to IPStream or DataStream; and

5. e) A successor product to Wholesale Line Rental if:

   1. i) such a product is provided using BT’s NGN, based on Multi-Service Access Node (MSAN) access; and

   2. ii) BT is determined by Ofcom to have SMP in a Network Access market or markets which includes that product.

Looking further to what was in 2005 the future, the Undertakings place certain obligations on BT regarding the provision of next-generation networks (NGN). Section 11 of the Undertakings makes it clear that BT will provide network access to its NGN on an EOI basis.

The Undertakings seek to ensure that BT designs-inequivalence of inputs into future products. Thus whilst there may have been a cost associated in ensuring that ʻoldʼ products were made fit for EOI, new products should not incur the same costs.

Ofcom also introduced the concept of ʻequivalence of outcomeʼ which was a weaker form of equivalence, more akin to non-discrimination, and was applied to products that at the time were expected to become redundant as they were overtaken by new services such as those listed above.

The difference between the ex post remedy of non-discrimination applied ex ante and the design of the specific ex ante remedy is central to an understanding of the Undertakings and their impact on the UK telecommunications market.

EOI was and remains a radical change from the ʻno undue discriminationʼ requirement placed on BT in markets where it has SMP. Under the non-discrimination approach, BT did not have to design-in to existing products and processes the equality of treatment of internal and external customers. Each could use a different product and process and differences between the two could be justified, allowing BT to charge different prices or to impose other non-price terms.

BT, or indeed any other incumbent firm, could legitimately argue that its network was built for use by a single integrated firm and was not designed for access by other networks. It was designed to carry calls from the calling party to the receiving party (end-to-end calling) and not to pick up calls or deliver calls to other (national) networks. Therefore, BT could argue that it faced lower costs to deliver a call end-to-end on its own network than to carry calls to or from an interconnected network. Likewise it could argue that it could provide different order-processing systems internally and externally.

Therefore, under the definition of non-discrimination adopted by Ofcom, its treatment of internal and external customers differently was objectively justifiable and therefore not unduly discriminatory. Nevertheless, industry participants and Ofcom determined that the competition policy principle of non-discrimination was not sufficient to stimulate effective and sustainable competition downstream of the economic bottleneck and so a stronger, specifically ex ante remedy was required to overcome the incentive to discriminate.

That remedy, equivalence of input (EOI), requires equal treatment to be designed into products. BT’s commitment in the Undertakings is to provide the ʻsameʼ product, timescales and information with the same degree of reliability. BT is also expected to respond to requests for new services from wholesale customers using the same process: i.e., it should not distinguish between a request from BT Retail and external customers.

There has been no legal testing of equivalence but it would seem unlikely from the unequivocal wording of the Undertakings that BT could claim external customers were not in an ʻanalogous situationʼ to their internal customers.

Equivalence can therefore be regarded as a specifically ex ante approach to redress incentives for discrimination, whereas the ʻno undue discriminationʼ obligation was an ex post remedy applied ex ante.

9.3.4 Functional separation

In the Undertakings, BT also agreed to a change its organization form and incentives for managers. The new organization form became known as ʻfunctional separationʼ though the term itself is not used in the Undertakings. BT made a significant number of commitments, the three most important and relevant of which were to establish:

1. a separate Access Services business unit with a separate brand name: One of the first deliverables from BT was the establishment of Openreach, a new business unit separated from the rest of BT with responsibility for providing the majority of the input equivalent wholesale products. Although not explicit within the Undertakings, it was a perceived aim of BT, through establishing Openreach, to develop a different culture which treated all of its customers in an equivalent manner. The introduction of Openreach ensured that there was a ʻcleanʼ interface with all the operators competing in the downstream markets and greater transparency for monitoring compliance with the Undertakings;

2. a Code of Practice for employees: it was obviously essential that the detailed set of commitments made in the Undertakings was understood clearly by the employees affected and so a simple code of practice was needed, backed up by training and support services for employees;

3. an Equality of Access Board (EAB): this body provided an independent means to monitor the implementation and administration of the Undertakings, to ensure that BT remains compliant with its commitments. Although it is a body internal to BT, its independence comes from the fact that three of its five directors are required to be independent of BT.

The purpose of these organisational changes was to remove the incentive to discriminate and so to facilitate the implementation of equivalence of input. Openreach managers have a set of incentives that are not connected to the overall performance of BT, but only to the performance of Openreach. In theory at least, this should encourage managers only to consider their own division, rather than the effect of their decisions on the profitability of other divisions or the business overall.

9.3.5 Parallel actions

The signing and implementation of the Undertakings were not the only actions taken by Ofcom to attempt to stimulate the take up of LLU and therefore change the competitive dynamics in the UK market. Two other actions were important: Ofcom’s review of the wholesale local access market (WLAM) and the setting up of the Office of Telecoms Adjudication (OTA).

In May 2004, Ofcom began its market review of the WLAM in line with its obligations under the CRF. The review defined the market on technologically neutral grounds such that both copper local loops and the cable access networks fell in the same market definition. Ofcom found BT to have SMP in the relevant market and so imposed a number of ex anteremedies designed to ensure access to local loops by third parties on fair and reasonable terms.

For the purposes of this case study, the most important obligation placed on BT was that of cost orientation. Condition FA3 of the formal Notification required that BT provide network access for a price that was ʻreasonably derived from the costs of provision based on a forward looking long run incremental cost (LRIC) approach and allowing an appropriate mark up for the recovery of common costs including an appropriate return on capital employed.ʼ

To establish the appropriate cost oriented price, Ofcom also investigated BT’s weighted average cost of capital (WACC) calculating a separate WACC for BT’s low-risk local access network business and the rest of the company, and the value of BT’s copper network. Both these consultations provided Ofcom with evidence that allowed it to reduce the price of both fully- and partly- unbundled local loops.

Table 9.2 shows the development of the price of fully- and partially-unbundled loops from 2002–2009²⁶. The data show a sharp decline in the monthly rental average cost of the first year between 2003 and 2005 with some fluctuation of charges since then but at prices between one-half and two-thirds of the 2005 price.

Table 9.2 Cost of LLU, UK, 2002–2009

Data sources: European Commission, Implementation Reports 2003–2009.

The substantial reduction in the price of unbundled local loops would clearly have a significant impact on the economics of unbundling for competitive operators and may well have been at least as important as the functional separation of BT in encouraging investment by competitors.

The second parallel development was the establishment of the Office of Telecom Adjudicator (OTA) in 2005, now superseded by OTA2. The principal purpose of the OTA was to ensure that the processes for LLU and other ʻcurrent generationʼ access products (for example, wholesale line rental) were industrialised and ʻfit for purposeʼ. It was technically independent of both Ofcom and the industry although its accommodation was provided by Ofcom and its members were representatives of BT and its competitors.

The original OTA was established to ʻfacilitate swift implementation of the processes necessary to enable competitors to gain access to BT’s local loop on an equivalent basis to that enjoyed by BT’s own businesses. The Telecommunications Adjudicator will also be able to bring all parties together to find a prompt mediated resolution of working-level implementation disputes.ʼ

OTA2 has a slightly different objective: ʻOTA2 will facilitate the swift implementation of processes where necessary to enable a wider range of Communications Providers and End Users to benefit from clear and focused improvements, in particular where multi-lateral engagement is necessary.ʼ

The OTA2 website sets out a six point ʻvisionʼ:

1. the OTA2 will champion end user issues;

2. Communications Providers will benefit from a competitive telecommunications infrastructure based on Openreach products that has no operational barriers to success;

3. there will be implementation (as quickly as is reasonably possible) of new product functionality, features and services relating to In-scope Products that will be seamlessly introduced;

4. migrations between broadband and narrowband products of both BT and other Communications Providers will be seamless, timely and with minimal interruption to service for end users;

5. dips in operational quality performance of In-scope Products provided by Openreach will be unusual and will be proactively managed by Openreach to ensure the least impact on Communications Providers and end users; and

6. participation by Communications Providers in the OTA2 Scheme will be widespread and representative.

Although OTA2 mentions end users in its objectives, the scheme participants are all drawn from the supply side of the market with no user representatives. The main scheme participants are the larger communications providers: BT, Openreach, BSkyB, Cable & Wireless, Everything Everywhere, Exponential-e, Global Crossing, O2, Scottish and Southern, TalkTalk Group and Virgin Media. Smaller communications providers are represented through the Federation of Communications Services.

Equivalence of inputs and functional separation, reduced LLU prices and the OTA can be seen as a three-pronged strategy to encourage the development of LLU as the principal wholesale product for broadband access. It may not be possible to separate the effectiveness of any one part of the strategy and of course it may be that the three together were critical to ensure increased adoption of LLU. In the next section of this case study we examine the UK broadband market today, at both the retail and wholesale levels.

9.4.1 The retail market

Broadband penetration has increased substantially since July 2005 and now stands at 22.1 million lines or 83.7% of households²⁷. This level of household penetration places the UK ninth in the OECD countries and 6^th in the EU, behind Sweden, the Netherlands, Denmark, Finland and Luxembourg. South Korea, Iceland and Norway also have higher levels of household penetration than the UK.

Figure 9.2 below extends Figure 9.1 to cover the period 2002–2013. A comparison of the rate of growth of broadband lines in the UK before and after the Undertakings shows no significant difference: diffusion of broadband continues to show the classic ʻSʼ shaped growth. There appears to be a slowdown in the growth of broadband in the wake of the 2008 financial crisis, with growth resuming in late 2010.

Figure 9.2 Broadband Penetration, UK, 2002–2013

Data sources: ECTA; Ofcom.

The retail broadband market has also seen substantial consolidation amongst suppliers. The four largest suppliers (BT, Sky, TalkTalk and Virgin Media) now have a combined market share in excess of 90%. At the time of the Undertakings, neither Sky nor TalkTalk was a significant player in the retail market, so both have entered the market on the back of an improved climate for LLU.

Sky’s market growth has been largely organic, benefiting from its strong position in the Pay-TV market to cross sell broadband to its existing customer base. TalkTalk Group, by contrast, has been acquisitive, acquiring the UK customer bases of AOL and Tiscali.

Table 9.3 shows various acquisitions that have taken place in the market over the period 2006–2009.

Table 9.3 Selected ISP mergers, UK, 2006–2013

Data source: Ofcom Communications Market Report 2013.

Ofcom explains the spate of mergers and acquisitions in the sector by the increased need for scale as a result of the increase in the uptake of LLU. For each unbundled exchange, the unit costs fall with each additional subscriber, meaning that scale is important for an ISP operator to be profitable, compared with using wholesale bitstream products from BT.

Whilst consumers’ choice of supplier may have diminished since 2006, consumers have benefited from an increase in the average connection speeds available.

Figure 9.3 shows the average actual broadband data rates for the period 2009–2013, which has grown steadily from 4 Mbit/s in 2009 to over 14 Mbit/s in 2013. Between 2012 and 2013 there was an exceptionally large increase in actual access data rates as more customers signed up to ʻsuperfastʼ broadband products based on fibre to the cabinet or DOCSIS3.

Figure 9.3 Average broadband connection data rates, UK, 2009–2013

Data source: Ofcom.

This increase in average connection data rates has largely been driven by the upgrade of DSL lines from ADSL to ADSL2⁺ and by the upgrade of the Virgin Media cable network to DOCSIS3.

Table 9.4 below shows a selection of super-fast broadband implementations and trials as identified by Ofcom in July 2009.

Table 9.4 Selected super-fast broadband implementations and trials, UK, July 2009

Data source: Ofcom Communications Market Report 2009.

Since July 2009, there have been a number of further announcements, in particular by BT and Virgin Media.

In October 2010, Virgin Media announced plans to upgrade to 100 Mbit/s from December 2010 with the entire network being upgraded by mid-2012²⁹. In March 2011, Virgin Media announced that its 100 Mbit/s service had passed one million homes and was on track to meet the 2012 deadline. In November 2013, Virgin Media announced plans to increase the access data rate to over 150 Mbit/s to all of the 12.5 million homes its network passes in 2014³⁰.

BT has responded to Virgin Media’s various upgrades, and arguably to political pressure, by launching Fibre to the Home (FttH) and Fibre to the Cabinet (FttC) products, which it retails under the brand ʻInfinityʼ. This fibre based service provides access data rates of up to 40 Mbit/s. Infinity was first trialled in exchanges in London, Cheshire and Glasgow and in January 2014 was available in around 1,900 local exchange areas covering over 75% of UK homes³¹.

In January 2013, according to the EC Broadband Scorecard, 14.5% of broadband lines provided data rates of 30 Mbit/s up to 99 Mbit/s, while 0.9% provided data rates of 100 Mbit/s or above.

9.4.2 Broadband services

With the growth in the average access data rate has come a change in how the Internet is used. One of the most important applications that demands high bandwidth is ʻcatch-up TVʼ, through services such as BBC i-player, and the equivalent from the other TV channels, which has enjoyed particularly strong growth. In March 2013, there were 272 million requests for TV and radio programmes on BBC i-Player, up from 78 million in 2009³². Equivalent data on ITV Player were not available.

BT has also entered the TV market offering IPTV over its FttC/H network. In 2013 BT secured the rights to show Champions League football matches live with a bid of over £800 million putting it squarely in competition with Sky’s and Virgin Media’s TV offerings. In fact, over 2013 there has been a noticeable shift in the locus of competition in broadband markets from Internet access to TV, with products such as ʻSky plusʼ and Virgin Media’s TiVo becoming increasingly popular.

9.4.3 The wholesale market

Whilst there have been many changes in the retail market, it is perhaps the wholesale market that has seen most change since the Undertakings were signed.

In 2006, 2010 and 2013, Ofcom conducted market reviews of the wholesale broadband access (WBA) market . This market lies at the intermediate level between the wholesale local access market (WLAM) and the retail broadband market (RBM). Whereas BT and Virgin Media self-supply the whole broadband value chain, most other ISPs operate only in the WBAM or retail market. LLU operators enter at the WBA level, buying unbundled local loops as their input, while retail ISPs enter at the retail level, buy bitstream access either from BT or an LLU provider that operates in the WBA market, as illustrated in Figure 9.4. It should though be noted that most LLU operators also buy bitstream access from BT in areas where it is uneconomic for them to place their own equipment in exchanges and purchase unbundled loops.

Figure 9.4 Broadband market structure, UK, 2013

Legend: WNIA: Wholesale Network Infrastructure Access (Market 4 in the EC 2007 Recommendation); WBA: Wholesale Broadband Access (Market 5); RBA: Retail Broadband Access (not included in the 2007 Recommendation); VM: Virgin Media (the cable operator).

When Ofcom’s predecessor, Oftel, first reviewed the WBAM in 2003, it was only able to identify one national market³³ and Oftel found BT to have SMP in that market. By the time Ofcom reviewed the market again in 2006 and 2010, it found there to be three geographic markets based on the number of LLU operators plus Virgin Media present in exchange area. According to the 2010 market definition the three markets are defined as:

1. Market 1: exchanges where only BT is present (11.7% of premises);

2. Market 2: exchanges where two Principal Operators (POs) are present or forecast and exchanges where three POs are present or forecast but where BTs share is greater than or equal to 50 per cent (10.0% of premises); and

3. Market 3: exchanges where four or more POs are present or forecast and exchanges where three POs are present or forecast but where BTs share is less than 50 per cent (77.6% of premises).

In the final statement Ofcom found BT to have SMP in Markets one and two, but no firm was found to have SMP in Market 3. This means that in exchange areas covering more than three quarters of all premises, BT is no longer dominant. Although the market definitions and size of the markets was somewhat different in the 2006 WBAM review, the finding of SMP was the same.

By 2013, the market had developed yet further and Ofcom found only two markets:

1. Market A: exchange areas where there are no more than two Principal Operators (POs) present or forecast to be present, which accounts for 9.6% of UK premises.

2. Market B: exchange areas where there are three or more POs present or forecast to be present, which accounts for 89.7% of UK premises³⁴.

This finding is a major change from the 2003 market review. Ofcom found that effectively competitive markets were operating below the retail level and thus the retail market in some 90% of the UK was not dependent on regulation at the wholesale level, albeit it is dependent on regulation at the LLU level.

This change in the market structure, and thus the finding of SMP, has come about because ISPs have substituted LLU for bitstream or wholesale access. TalkTalk Group now has over 2,700 exchanges enabled, covering some 24 million homes within 5 km from the exchange, 97% of all households. Sky has enabled over 2,300 exchanges covering 23.3 million homes within 5 km³⁵.

Figure 9.5 shows the number of wholesale copper access lines, i.e., those not offered by BT, by type: resale, bitstream and LLU. Resale lines peaked in January 2007 as LLU began to take an increasing share of the market. Until January 2006 almost no customers accessed broadband via LLU, but by January 2008 LLU was the most popular single method of wholesale copper access and by July 2010 two-thirds of wholesale copper access lines were via LLU. In July 2010 some 29% of all retail access lines were via LLU, whilst bitstream and resale combined reached around 22%.

Figure 9.5 Wholesale copper access methods, UK, 2002–2009

Data source: Based on ECTA Broadband Scorecard.

Over the same period, cable’s share of the market has declined from 59% to 20%. In part, this decline is due to BT extending its broadband availability to around 99% of homes whilst cable still only reaches about 55% of homes: BT’s larger physical reach ensures that twisted copper wire has been able to gain a larger market share.

What we have seen in the UK therefore is a dramatic change in the nature of competition from one based on a mix of cable and the reselling of BT’s bitstream products, to one where LLU has a significant and growing market share with unbundled loops largely replacing bitstream.

This change in the wholesale market is important as LLU allows ISPs considerably more control over their own products. Whilst care should always be taken over simple correlations, especially over time, Figure 9.6 shows a strong correlation between LLU as a share of all competitive copper access lines and the weighted average access data rate enjoyed by users. However, it should be noted that the greatest percentage increase in access data rates happen when LLU had only a very low market share (less than 10%).

9.6 Relationship between LLU and average access data rate, UK, 2002–2008

Data source: Ofcom.

10.2.1 Telecom 1.0: the history of the Italian telecommunications sector (1853–1995)

In 1853, the Italian government established a government-granted monopoly on long-distance transmissions (i.e., telegraphy service) sanctioning a strong public presence in the telecom sector, with the main consequence that all relationships between the state, the licensee and the customers had to be regulated by laws. A law of 1892 enforced public presence by stating that, on licence expiry, all installations should return to public ownership, without any compensation fee. The main effect of such a normative prescription was to reduce the propensity to invest and, as a consequence, to restrain the increase of users, especially in the smaller cities. However, because of the deployment of the first long-distance networks, from 1904 to 1907 the overall number of customers increased, though remaining far below that of similar European countries.

In 1923, a Royal Decree (R.D. n.339) initially reserved to the state the right of installation and operation of telephone networks for both private and public use. However, in 1925 existing concessions were consolidated and awarded to five private companies (Stipel, Telve, Timo, Teti and Set) for regional telephony and R.D. 884/1925 entitled the state agency ASST (Azienda di Stato per i Servizi Telefonici) to coordinate and interconnect them for interregional and international calls. Moreover, other private companies, Italcable, financed by Italian migrants to Argentina, and Italo Radio, were entitled to provide intercontinental telegraph and telephone services from 1921 to 1941, the year that they merged.

The basic rationale of such a governance model for the Italian telecommunications system stemmed from the necessity to avoid a high concentration of services in the hands of a single (private and/or foreign) licensee, while reaching the economic policy objective of using private funds to finance investments, a necessity given the heavy public debts accumulated during World War I.

The radical shift towards the full public control of the telecommunications sector began in the late 1920s and was scaled up in 1933 with the foundation of IRI (Istituto per la Ricostruzione Industriale; i.e., the Institute for Industrial Reconstruction, a state-owned conglomerate aimed at counterbalancing the effects of the global depression on the Italian economy) and STET (Societa’ Torinese per l’Esercizio Telefonico), its sector holding for telephony. STET acquired control of Stipel, Telve, and Timo – i.e., the telecommunications assets of SIP (Società Idroelettrica Piemontese) – through the emission of state guaranteed bonds, which could be converted into shares.

STET became public in 1936 and by the end of the conversion process more than the 40% of the equity was in private hands. In 1958 STET also acquired the control of the remaining private licensees (SET and TETI). In 1964, following the nationalization of electric utilities in ENEL, through the enforcement of D.P.R. n.1594/1964, the five regional licensees merged with SIP. Note that the regional electric utilities were also controlled by IRI. Hence the nationalization indemnities obtained by SIP could be used to finance the diffusion and modernization of the telephone services. The merged company was renamed Società Italiana per l’Esercizio Telefonico (SIP) and became the nationwide licensee responsible for almost all telecommunications within and among the 210 newly-established districts. It was controlled by STET but also retained the private shareholders of the previous SIP, who were attracted by the profitability of the telephone business.

In the same year, STET also acquired the control of Italcable and Telespazio (responsible for satellite communications). As a result, STET was entitled to technically and managerially address all the licensees in the telecommunications sector, while the state agency ASST, still in charge of national long-distance and international services, lost much of its strategic role.

Moreover, especially in the 1980s, STET pursued a strategy of vertical integration and expansion in the adjacent industries of the emerging ICT ecosystem: IT services, telecom manufacturing, infrastructure building, research and development, data services, publishing and advertising. The attempt to enter the cable-TV business failed, because of both technical and regulatory problems, and the venture was transformed into a satellite Pay-TV platform.

The redesign of the governance structure of the whole sector spurred a new growth period in terms of users and installations. From 1964 and for almost ten years, a steady growth in the rate of user uptake was recorded, also as a consequence of substantial investments to update the network, the introduction of new tariff policies for both local and long-distance calls, the abrogation of additional fees for residents outside the cities and, last but not least, the low risks associated with a monopolized structure of the sector. In those times, major attention was also given to cost control, reduction of the financial leverage and the use of depreciation for self-financing.

During and after the oil crisis of the 1970s, the sector suffered due to the tariff freeze imposed by the government to try to keep inflation under control.

In contrast, during the 1980s SIP enjoyed a close relation with the government, which provided funds channelled through IRI to increase equity, allowed the increase of tariffs and the decrease of licence fees; and allowed the flow of subsidies from international and long distance traffic to the operation of local networks. SIP also received financial support from the Italian bank CDP (Cassa Depositi e Prestiti, which finances public infrastructures using postal savings) and the EIB (European Investment Bank). This enabled the company to sustain heavy capital expenditures in network capacity expansion and modernization, and helped to reduce the gap in infrastructure quality between northern and southern Italy.

The first half of the 1990s saw further important steps in the growth of SIP, such as the activation of the first large-scale mobile network and of the first digital access network – ISDN.

In 1994, the consolidation of the mixed public-private monopoly of telecommunications services controlled by IRI was completed: STET, SIP, Italcable, Telespazio, Iritel (formerly ASST) and SIRM (maritime services) were all merged to form TI.

10.2.2 From monopoly to competition: the early days (1995–1998) and beyond

The public-private telecom monopoly has fostered the development of a telecommunications system with users’ interests protected in terms of availability of services and the level of tariffs (within the usual pre-liberalization context of tariff unbalance and publicly-financed universal service obligations). High fixed costs for network deployment and economies of scale justified the monopoly of the access network at a micro-economic level. However, the creation of a European telecom regulatory framework and the ambitious ideas exposed in the Green Paper (1987), substantiated in 1990 by the ʻtwin Directive setʼ (Competition in telecommunications and Open Network Provision), eventually led to the liberalization of the Italian market, with sizeable effects on competition. At the same time, the privatization of TI became a core issue at both national and European levels, as it contributed to pay off IRI’s debts in other industries and prepare for Italy’s adoption of the euro, from 1 Jan 1999 starting with non-physical payments.

A long-term reform process was initiated with State Law n.58/1992 transforming the state agency ASST into IRI-owned Iritel in order to separate ownership and management from regulation. The reform of the governance system was more precisely defined in 1995 and 1997 by the establishment of the regulatory authorities for public utility services (State Law 481/1995) and the actual setting up of the Autorità per le Garanzie nelle Comunicazioni(AGCOM, the telecommunications and media regulator, Law 249/1997), which began its operations in March 1998 (more on the creation of AGCOM in Section 10.2.4). The new corporate model required major organizational changes such as adopting a divisional structure to make the different lines of business more independent and their individual performances more visible. Moreover, key legislative initiatives targeted the abolition of exclusive and special rights previously assigned to the incumbent. Crucially, D.P.R. 22/12/1994 introduced non-exclusivity for the incumbent in the provision of mobile communication based on GSM technology⁷. Moreover, State Law n.103/1995 – finally transposing EC Directive 90/388 – allowed any European service provider to offer (existing or potential) data services and closed user groups to business customers. Overall, the full transposition of European Directives was quite slow, mostly due to lobbying and delaying actions from established industry groups.

Among the most important steps, was the enactment of D.L. n.546/1996, converted into State Law n.650/1996, which allowed the government to more easily implement the EU directives; State Law 55/1997 and D.P.R. 318/1997 provided the core elements of access and interconnection as well as universal service.

Full liberalization was completed with the (emotional, one could say) award of the first authorization for public voice telephony service in February 1998 to Infostrada (today owned by Wind), quickly followed by a large platoon of network and service operators. The most notable were Fastweb, Wind, BT-Albacom and Tiscali. Vodafone entered the fixed broadband business at a later stage, although today it has a substantial customer base, almost as large as Fastweb. Wind, currently the most important competitor to TI in terms of fixed broadband market share, was born in 1997 as a spin-off of the national electricity incumbent (ENEL). France Telecom and Deutsche Telekom were part of the initial set of key shareholders. However, they soon exited the venture and sold their shares to ENEL. In 2005, ENEL then sold Wind to Wind Telecom S.p.A. (formerly Weather Investments) owned by the Egyptian group that also owned Orascom Telecom. In 2011, Wind became part of the Vimpelcom group, following a merger between the Russian company, with the participation of Telenor, and the Egyptian Orascom Telecom. Wind started off as an integrated fixed and mobile operator and incorporated Infostrada, but was able to exploit this integration only to a certain extent. Today, Wind is a major mobile operator.

10.2.3 The early days of Internet access in Italy

Internet access ‘in Italy’ (i.e., by means of a permanent physical node) started as early as 1986 in Pisa. Italy was tied for third European country to get on the Net, following the UK and Norway and at the same time as the Netherlands.⁸ Generalized access to the Internet was essentially launched as a scientific project – at a cost equivalent to about EUR 2.5 billion – in 1988 to unify the various scientific institutions and their large computers in Italy by means of a single network, GARR (Gruppo Armonizzazione Reti della Ricerca – from the ministerial committee which gave its name to the network).

The first commercial ISPs for residential users appeared several years later, in 1992–1993, building on their experience with bulletin board systems (BBS). In 1994 I.net became the first ISP for business customers, joined a little later by IT.NET.

I.NET and IT.NET used links provided by Unisource and Pipex, respectively. However, the carrier network, the landing points of the marine cables and all the core connections that made narrowband traffic possible were almost entirely provided by TI; without TI little would have happened in this regard.

However, as noted in other countries, the incumbent was essentially taken offguard with respect to the ‘early adoption of the Internet’ when considered as a mass-market for businesses (which TI only entered in 1996) and residential customers (with TI’s first offer in 1998, under the brand name of TOL – Telecom Italia Online).

With regard to the residential market, the uptake of Internet access (with very slow and expensive connections, especially if long distance calls were required) was a bottom-up affair, more from the edge than from the core. Especially between 1992–1996, Internet access was mainly propagated by former BBS (bulletin board system) managers turned into service providers. Names like Agorà, MC Link, Video On Line and I.net, amongst others, are early examples of more professional services, all active since 1992–1993. Agorà, based in Rome, was particularly significant as it was born out of necessity, for the Radical party (a small political party, which became an important civil liberties movement in the 1970s) to connect with its many affiliates throughout Europe using the Internet as a means to avoid costly phone calls and faxes. Video On Line (started in 1993) was the first mass-oriented provider and it built on its connection with the research center CRS4, at that time directed by the Nobel prize laureate Carlo Rubbia, also a former director of CERN (where the Web was born). See Figure 10.1 for a screenshot of its home page. Video On Line is (self) credited with hosting Europe’s first online version of a daily newspaper (L’Unione Sarda) in March 1994.

Figure 10.1 Video Online home page, Italy, 1996

Source: Video Online, 1996.

Tiscali, which was established January 1998, was considered an heir to Video On Line, as they both originated from Sardinia and as it fostered the mass-market adoption of access to the Internet, being the ‘ideal type’ of a bold, fast competitor, without ever making a profit in more than 10 years of its operations.

Tiscali’s key strategic move consisted in pioneering the ‘reverse’ (indirect) interconnection mechanism. From March 1999 onward, Tiscali offered a subscription-free Internet service whereby the customers only had to pay for the time they were online. Tiscali then derived its income from the ‘indirect interconnection’ option, whereby the interconnected operator (TI in this case) pays to a carrier (Tiscali) an interconnection charge for bringing traffic to its network. In other words, TI offered an interconnection service to the third-party operator Tiscali. At the time, giving up paid subscriptions was a bold move, which allowed Tiscali to gain an important share of the market. This made Tiscali’s IPO in October 1999 a real success: at the apex of the dotcom bubble, Tiscali exceeded Fiat’s market valuation. Tiscali used these financial resources to expand its ISP operations in several advanced European countries. But in subsequent years economic and financial problems forced it to sell all its ventures outside Italy.

Another extremely successful IPO was at the origin of E-Biscom in March 2000. This venture was the result of an agreement with AEM (the energy utility of Milan) to create an infrastructure company (Metroweb) and a telecoms operator (Fastweb). E. Biscom had participations of 33% and 70%, respectively. E.Biscom was incorporated into Fastweb in 2004.

10.2.4 AGCOM, the first integrated regulator in Europe

AGCOM (Autorità per le Garanzie nelle Comunicazioni) was created as a single, or ‘integrated/convergent’ regulator in July 1997 with the task of applying European legislation in telecoms and media. However, operations only started in March 1998, with the formal appointment of the first board. It was essentially modeled after the experience of the Canadian regulator (CRTC) and the Federal Communications Commission (FCC) in the US. For the first time in Europe, a regulator would be responsible for jointly regulating media and telecommunications. Others, like OFCOM in the UK, followed suit only several years later. AGCOM superseded the Media Authority⁹ and inherited most of its competences from the Ministry of Communications (now part of MISE, the Ministry for economic development), although some job-sharing with the latter continues to this day, especially with regard to the issuance of authorizations, spectrum management and broadband policies. Overall e-policy making, including today’s ʻDigital Agenda for Italyʼ initiative, is managed by the Prime Minister’s office, together with MISE. In 1998, creating an integrated regulator was ambitious. However, while a number of telecom/media subjects are discussed jointly by AGCOM’s board, the management of single regulatory issues remains largely divided between the regulators’ offices and treated without too much emphasis on ‘convergence’.

On the other hand, being ‘convergent’ has often put AGCOM at the forefront of regulatory and market developments, sometimes resulting in innovative thinking. Nevertheless, critics have pointed out that, especially with regard to audiovisual matters, the governance of the regulator exposed the regulator to political influence, with a President appointed by the Prime Minister and a board of four Commissioners nominated by the Parliament.

In the end, if regulatory performance should be judged by metrics, it is uncontestable that Italian consumers benefit from low retail prices (including mobile pricing, notwithstanding termination rates which for a number of years remained within the higher bracket of EU members), and ease of choice (one mobile number out of four in Italy is ported). Today, TI has roughly 14 million direct access lines (63% of the access market) while a further 7.2 million lines of its network are managed by altnets. In addition, 0.5 million lines (NGA/FWA) are owned by alternative network operators.¹⁰

10.2.5 The privatization of TI

A few key points of the privatization of TI are relevant for our discussion on the Italian broadband market.

The privatization model chosen for TI is the subject of several, mostly critical, analyses. They all underline that the original model (a public company with a large set of shareholders) failed to be properly implemented (or, was the problem inherent in the model of choice?) and that, eventually, controlling shareholders emerged and charged the company with high financial leverage. The debt still hampers investment, and has radically downshifted the position of the company in the international markets. It is difficult to disentangle the complementary effects of privatization, liberalization, pro-competitive regulation, financial bubble and global difficulties of telecom operators in engaging with the Internet world. However, there is no doubt that under these circumstances the privatized company was unable to pursue the same bold growth strategy and to secure to the shareholders the same handsome financial performance as the previous mixed-capital venture.

Essentially, the private TI went through four distinct changes of controlling groups. Firstly, in 1997, the Treasury granted the control of the company to a small nucleus of investors including Ifil (the Agnelli family), Mediobanca and other recently privatized banks, bank foundations¹¹ and building companies, with the private placement of less than 10% of voting rights. Moreover the Treasury placed most of its other shares through a public offering: about 1.5 million small investors acquired 32% of the equity. The weakness of these initial core investors led to what has been labeled as the largest hostile takeover in European history in 1999: Olivetti bought 51% of the equity through a public offering. The investment, worth more than EUR 30 billion, was financed with debt but also by selling to Mannesmann the control of Infostrada (subsequently acquired by Wind), and that of Omnitel, the second Italian mobile operator, eventually acquired by Vodafone.

The operation, as much as the original privatization, was marked by heavy-handed intervention from political parties leading to a leveraged buy-out which made no economic sense and allowed a few key investors to take control of TI through a long chain of Chinese boxes that ended in Bell, a Luxembourg-based holding company.

In 2001, the third change of the controlling group did not require a public offering. Olimpia, an ad-hoc vehicle created by the Pirelli and Benetton groups, banks and some of the Bell shareholders, acquired 23% of Olivetti’s equity from Bell, paying twice the market value; EUR 7 billion in cash was sufficient to acquire the control of TI, a company which at the time had a market capitalization of EUR 55 billion.

A lot of divestments, consolidations (between Olivetti and TI) and the buyback of the floating capital of TIM were executed only to help the controlling group to pay its debts. But eventually Olimpia had to give up and a new vehicle, Telco, took control of TI with 23% of equity in 2007. Telefónica of Spain owns more than 40% of Telco and has negotiated a further increase of its holdings, de-facto becoming the controlling shareholder of TI. These leveraged buy-outs (LBOs) resulted in a huge debt that peaked at EUR 40 billion in 2005.¹² After years of painstakingly repaying, the debt is ‘down’ to EUR 28.2 billion as of 31 December 2012.¹³ However, for many years investments have been severely constrained in order to assist Bell, Olimpia and now Telco to maintain their control.

10.3.1 Broadband development: the importance of television in shaping competition and regulation in Italy

No serious discussion on the evolution of broadband in Italy can take place, in our view, if the role of television is not carefully considered. As we pointed out in the introduction, the ʻoriginal sinʼ of Italian broadband is the lack of competitive pressure stemming from the absence of CATV. Television competition also affected access conditions following regulation introduced by the Competition Authority.

The early liberalization of the free-to-air (FTA) television market, also allowing foreign broadcasters to transmit from Italian territory, was coupled with legislation (DPR 29 Marzo 1973, n. 156) that briskly halted CATV in its infancy. Part of that legislation was later on declared unconstitutional and amended in 1975. However, by then FTA TV was already established as the (cheaper) means of choice for local broadcasters, essentially making the deployment of a nationwide CATV uneconomical.

As around 1995 it became conceivable to use a CATV infrastructure to compete with emerging pay-TV satellite services and as a broadband vehicle, TI chose to invest heavily in the so called Socrate Plan, aimed at deploying a fibre+coaxial cable network in all major Italian cities¹⁴. This move was criticized as a market pre-emption by the dominant player in infrastructure, as it took place on the verge of full liberalization from 1 January 1998. However, the passive infrastructure was only partially deployed as TI’s broadband investments were shifted towards ADSL. The pay-TV company of the TI group (Stream, created in 1997) migrated to satellite and was eventually sold to Newscorp in 2003.