INTUG  
RMIT - Royal Melbourne Institute of Technology

The rise of broadband telecommunications

Ewan Sutherland


Good day ladies and gentlemen


My thanks to Professor Bill Martin for the invitation and for the opportunity to speak here today.

It is always good to be reminded that I spent much of life in and around universities, studying and teaching in British universities.

The definition of a university teacher is someone who talks in someone else's sleep. I will do my best to keep you awake.

I want to try to paint a global picture of policy developments in the deployment of broadband, to ask where we are and where we are going, then to try to identify reasons why this might be important and what we might do about it.


What is broadband?

The definition of broadband is what you want to make of it. Some people include ISDN, but that is only 2 x 64 kbps channels and I think that is a bit low. Some ADSL offerings are around 256k or 512k, downstream, which is still to me "fast narrowband". Once you get to 2 or 4 megabits per second, then you are getting serious. I think a crucial cut off point is when you can and cannot watch decent quality video. Some services for residential users go up to  8 or 12 or even 20M bps.

The A in ADSL stands for Asymmetric, that is you can receive more than you can send. For many purposes this is adequate. For other purposes, mainly business or content provision, you need more symmetric capacities.

Of course, it is always on and flat-rate. Those are crucial characteristics. That means there should not be some arbitrary download limit of a few gigabytes per month.

We have to recognise that we are talking about a subject where the basic terms are poorly agreed and where, almost inevitably, technological change drives us on.


The global policy debate

The question of broadband Internet access became a matter of prominence in policy circles when it was discovered that South Korea had shot ahead of Canada, to take the number one position in the OECD. That is measured as broadband lines per 100 hundred population.

This came as something of a blow to the Canadians who had worked very hard to remain ahead of the USA. Given their geographical problems, it was and remains a considerable achievement. Their "outback" is every bit as challenging as that of Australia. Providing high bandwidth capacity to places where aerials freeze up and the aurora borealis interfere with satellite transmissions is not simple. 

We had some amusing OECD meetings where governments were producing data generated the week or the day before in order  to prove that they were much better than the "old" numbers of only a few week previous. It was slightly sad and mildly embarrassing to watch your former colonial power, the United Kingdom, claim that the upswing had begun, the problems had been overcome, that a great leap forward had been taken and that a thousand flowers were blooming. The UK, at the commencement of 2003, still had only about one million lines to share amongst its 57 million population.


Table 1    Top economies by broadband penetration (June 2002)

Country
 density
 rank
Korea (Republic) 19.3
 1
Hong Kong, China 12.1
 2
Canada 9.7 3
Taiwan, China 7.2
 4
Belgium 6.3 5
Sweden 5.4 6
Iceland 5.2 7
USA 5.0 8
Denmark 4.5 9
Singapore 4.4 10
Austria 4.3 11
Japan 3.9 12



Australia 1.6 20
New Zealand 1.0
 25

Source: ITU Asia-Pacific Telecommunications Indicators, Geneva, 2002.

 The ITU comments:
It is interesting to contrast the advanced Asian economies with Australia and New Zealand, the other developed economies in the region but which are lagging in broadband deployment. The situation in the latter two countries is more akin to developments in North America and Western Europe. Incumbent operators have slowed deployment by dragging their feet on unbundling. Competitors have been reluctant to invest in their own facilities. Prices are high (and pricing structures obtuse) with operators keen to avoid cannibalizing their ISDN and leased line revenues. Similar to North America, DSL is a harder proposition to sell in Australia and New Zealand; the former has flat rate pricing for local calls while the latter does not charge for local calls. Finally, the Australian and New Zealand governments have more of a laissez-faire attitude than Asian ones. Their attitude is that broadband should be left to the market rather than influenced by government policies.
Source: ITU Asia-Pacific Telecommunications Indicators, 2002, p 32
The "obtuse" pricing structure being an offer of 256kbps with a 256M download limit and a charge thereafter for each additional megabute.

It is really metered broadband in Australia and New Zealand that seems to be the barrier. Flat rate dial-up is available in Canada but has not stopped take-up of broadband. It is also available in South Korea, though it is barely used. This strongly suggests that it is broadband pricing and not dial-up pricing that determines the speed of take-up of Broadband. Even in the UK, demand greatly increased following the reduction of broadband prices and again flat rates for dial-up had been available and were well established.

Some people claim the South Koreans have poured billions into broadband, that it is all paid for by the state. Well, you have to recall that in 1997 South Korea went through a very severe economic crisis. Therefore, one is left wondering where such vast sums of money would come from. Moreover, South Korea is a member of the IMF, WTO and OECD, all of which report regularly on the South Korea eocnomies and governmemt policies. Everything is a matter of public record.

Of course if a medium-income economy such as South Korea could afford to subsidise broadband, and I am not saying it did, then it ought to be far easier for Australia, a much richer country.

France is presently engaged in the fascinating spectacle of trying to put 0.6% of this year's GDP into France Telecom and it is not a pretty sight.

It is likely, that the South Korean government will make a substantial profit from its very modest investment in broadband. Far more important than any investment was a commitment and a vision of where it wanted to be. Currently, that is 20Mbps to each home by 2008.

The South Koreans set out with the same objective as Europe and North America, but managed to deliver. What is different about the Korean strategy?

In Japan and South Korea, the day when demand for broadband access will levels off is in sight, Korea early next year, Japan a couple of years later. Then the operators will have to concentrate on getting market share in China and other emerging markets - perhaps even Australia - plus the development of services. They will also push existing customers to increase their use of bandwidth and to take new services.

The story gets a bit messy, because people were so dazzled by the Korean performance that they sought ways to dismiss it or to discount it. They saw it as unattainable, so they tried to denigrate it.

Now, South Korea does have a very unusual housing pattern. It is vastly more densely populated than Canada or Australia, even than the UK, it is more like Hong Kong or Singapore. This is both in terms of the percentage living in urban areas and especially those living in apartments. So, it was possible to argue, that South Korea was unique.

It was hardly plausible to suggest that Canada, in order to regain the number one position in broadband, rehouse its population in two concentrations of skyscrapers.

A related reason for the willingness to embrace the "uniqueness" argument was the previous failure of Singapore in its bid to be a broadband city-state. Korea appeared to be something special.

However, the sudden appearance of the Koreans at the head of the league table had been too diverting. Gradually, it has emerged that South Korea was far from unique. Japan has, if anything, grown more rapidly. Since the middle of 2001, Japan has seen exponential growth of ADSL and has now reached some six million ADSL lines, plus 2 million cable lines, with new subscriptions at speeds of 12 Megabits per second.

Looking further south, Taiwan, Hong Kong and Singapore have come strongly into the picture.

Most impressively, China has been moving at speed. By the end of last year it had attained 3M broadband subscriber lines and will quickly become the largest market in raw numbers. It can copy the South Korean example of wiring up its very many high rise blocks. It will also have to address the "outback" question, but it can first install tens of millions of relatively easy urban lines. Nonetheless, there are problems with the poor quality of many lines.

So it can be argued that there is at least two models that work in North-East Asia, admittedly one of those linked to fairly dense urbanisation. South Korea does top this up with the use of satellite for some rural area. In Japan cable and ADSL cover fairly conventional suburban areas.

Broadband has become an Asian phenomenon much more than it is American or European.

By this time next year, the largest markets for Internet access (narrowband and broadband) and for mobile telephony will all be China. In crude numbers, they will have overtaken the rest of the world. In terms of spending per capita they will still have a long way to go. In terms of growth, it will match any country.


Cable television

INTUG has its office in Belgium. On the whole Belgium is doing quite well for broadband. It is one of the European leaders, which is something of a surprise, since the incumbent operator is not noted for being a leader. It is more of a dinosaur, not the ones they make Hollywood movies about. This is not a smallish, vicious and carnivorous dinosaur, rather it is one of the enormous, slow ones with a tiny brain largely disconnected from external events. Its excellent performance in broadband is the result of luck, more than anything else. It is hard to find anyone who can legitimately be given the credit for this.

Belgium is a relatively small country, fairly densely populated and has had almost total cable television coverage for many years.

Belgium divides into the northern Flemish speaking part and the southern French speaking part, with Brussels a sort of plurilingual bubble in the middle. The strange difference has been that the cable television companies in the north launched into cable modem service, while those in the south did not. There does not seem to be a rational explanation for this.

Looking around the OECD countries there is a similar problem of trying to understand why cable modem service can be a strong player in some markets, yet poor or absent in others.

In the European Union there was a Commission Directive issued in the mid-1990s under  Article 90 (then Article 86) of the Treaty of Rome. This had the effect of forcing incumbent telecommunications operators into selling off their cable television interests. It was to achieve competition between the PSTN and cable networks and to break the stranglehold of the PTTs.

In Germany, it was resisted by Deutsche Telekom so successfully, that its cable interests are only now being sold. Consequently, cable modem service is still not a factor in Germany for broadband. Deutsche Telekom managed to delay that option so effectively, that it may be too late.

In the Republic of Ireland, the cable interests were sold off. However, the two companies concerned have been or still are bankrupt and the quality of the networks was very poor. Thus in a country which lives on its fame as a high-tech location, employers cannot give their employees broadband access at home. It is doing considerable harm to the image and the prospects of future economic growth.

In some countries there is no cable television network operators to play a part. This is true in Italy and in some parts of the UK. In others, the significant content, especially key sporting events, is held by the satellite operators, making cable television is a poor option for operators and customers alike.

In the developing world there is a different problem. Some countries have only two channels; one of government ministers being shown inspecting factories and military bases or visiting schools, while the other shows people reading religious texts. Only in USA can you make money from the latter. Without true media diversity there is very little money to be made and so no cable television, not even triple play.

The term "triple play" comes from the USA where it means the option to delivering on a single cable into a residence:
  1. voice telephony
  2. broadband Internet access
  3. television
Across the three, an operator ought to be able to recover its investment and make a profit.

The case here in Australia has been rather messy. It involves Telstra, Optus and the press magnates, Murdoch and Packer. In effect, the result has been a hard core duopoly in which neither party has condescended to compete. This has left the cable networks out of the broadband game. It has all the artifice of one of those World Wrestling contests.

It has also left government with few tools to push them to make them compete.

Nonetheless, the Government could and should separate the cable interests of Telstra. It might sell this to the highest bidder, conceivably even Optus or Murdoch/Packer. Alternatively, it might de-merge the cable interests as the Australian Cities Cable Company, with separately quoted shares and a new board of management. The government shares in that could then be sold in a public offering, without the same controversy as those of Telstra.

In all countries there is a strong case for getting cable operator started in competition for broadband access. It is the easiest infrastructure to be used to compete with the PSTN.


Unbundling the local loop

An enormous part of the debate on broadband has been devoted to the issue of unbundling the local loop. It has proved to one of the most frustrating policies to be implemented in years. Both in the USA as UNE-P and in Europe as LLU the progress has been like the trench warfare in the Great War. It has been fiercely resisted by the incumbent operators, who see it as a threat to their profitability, if not their existence.

Germany was an early mover on local loop unbundling. These were used mainly for ISDN and some voice telephony. This is now declining as Deutsche Telekom runs a win-back campaign to get customers onto its ADSL offering.

One of the more remarkable achievements of Deutsche Telekom was not to have just unbundled local loop, but to have nineteen (19) flavours of unbundled local loop. It always struck me as a good examination question to ask candidates to explain the differences between the 19.

At the beginning of January 2001 the European Union saw a new piece of legislation take effect, a Regulation on the unbundling of access to the local loop. The same legislation in all fifteen countries at the same time. The results are wildly variable. Looking at the numbers, it is hard to believe they operate under a unified legal framework.


Table 2    ECTA DSL Scorecard (as at end December 2002)


 DSL cable
 other
 Total

incumbent wholesale

 LLU



incumbent
retail
 OLO
retail
 OLO
retail
 cable
modem
 other
totals
Austria 143,000 31,000 5,500 360,000
539,500
Belgium 423,500 78,400 3,641 280,000
785,541
Denmark 184,600 0 47,652 122,000 5,000 359,252
Finland 175,000 5,000 39,000 54,000 500 273,500
France 987,000 413,000 6,769 248,519
1,655,288
Germany 2,800,000 0 210,000 120,000
3,130,000
Greece 0 0 0 0
0
Ireland 2,645 9 633 4,000 3,000 10,287
Italy 475,000 175,000 50,400 0
700,400
Luxembourg 4,300 0 130 70
4,500
Netherlands 316,450 0 43,152 700,000
1,059,602
Portugal 43,657 9,046 54 207,000
259,757
Spain 734,087 223,117 3,099 230,000
1,190,303
Sweden 314,000 104,000 5,063 150,000 200,000 773,063
UK 300,000 290,000 2,000 736,000 6,500 1,334,500







Totals 6,903,239 1,328,572 417,093 3,211,589 215,000 12,075,493





The European Commission has admitted that the Regulation had failed to take into account the capacity of the incumbent operators to engage in "sportif" tactics. That means playing the man, not the ball.

The bloody-minded defensive tactics have been deeply depressing. There have been attempts to overcharge for colocation space and equipment. There have been the almost inevitable claims about the need for security clearance for staff from competitors, questions of access to the toilets and so on. It was approached by the operators as an opportunity to make life difficult for competitors in the hope they would die or go elsewhere. 

The contrast with Japan is illustrative. There have been no law suits, no protracted disputes, just deployment and the development of what appears to be a competitive market strucuture. The market shares look very good, without undue dominance by NTT. Many companies are taking the wholesale shared access offer on the local loop, together with a fibre backbone network provided by NTT at cost orientated prices and can offer 8 or 12 M bps for about USD 20 per month.


Table 3  The cost of broadband in Japan


 Downstream
bps
 Upstream
bps
 Monthly charge
NTT
 12M
 1M
 ¥3,580
8M
 1M
 ¥3,480
1.5M
 512k
 ¥3,180
Yahoo! BB
 12M
 1M
 ¥2,480
8M
 1M
 ¥2,280
Source: operator publicity material 1 March 2003.


One aspect of this does give cause for concern, though it maybe no more than chance. In Japan a full local loop is available for about USD 16 which is broadly comparable with other countries, perhaps a little higher. There is an offer of shared access for about Yen 173 per month. Now this is very cheap. At first sight, it is so low as to suggest anti-competitive pricing, in order to make uneconomic the construction of competing infrastructure.

The justification offered, is that Japanese customers have paid a substantial sum for the original connection and with "shared access" will continue to pay NTT a substantial monthly subscription for the local loop. Thus the costs to NTT will be very limited. Nonetheless, one day there will have to be some rebalancing. 

South Korea was periodically listed as not having unbundled the local loop, so it passed a law. There appear to be no unbundled local loops in South Korea and no intention of their being any. However, they cannot now be tarred with that particular brush. They have a policy of open access, to which I will return.

There are bad examples of local loop unbundling and some not so bad, but few good examples, notably Japan and Hong Kong.

Nonetheless, local loop unbundling has achieved a growth of broadband access in many countries. It was the big stick that caused operators to discover that a business case did exist for broadband. Suddenly, we saw Belgacom and TDC rolling out ADSL in Belgium and Denmark. The result is a modest supply of broadband but from wholesale markets that are dominated by the incumbent operators in both cable and PSTN.

While the retail market may look better, the problems in the wholesale market may ultimately cause problems there.

The worry or concern is that we do not know if that market structure will allow for the move forward to higher speeds or whether it will freeze.


Structural separation

Part of the policy debate is about finding the next big stick with which to beat the incumbent operators. This is necessary because of their inertia, their protection of their existing revenues and lack of imagination in moving on. Many incumbent operators are really law firms that happen to run networks.

Consequently, we must look at structural separation. It is nothing if not big, though how effective a stick it might be, is another matter.

Here we would be going beyond mere unbundling the local loop into the separation of the "loop company" from the operator. It was debated recently at the OECD and in other places. Such debates are not very convincing, they seem too abstract, too driven by economic theories or they seem based on dubious parallels.

Doubtless there are practitioners of economics, the dismal science, in the room, so that reminding them of the definition of an economist as being someone who has never met a real person, but had one described to them, will be unwelcome

In October 2001, the OECD published a report from its Competition Committee. At its simplest, this advocated the structural separation of everything. This has been followed up in what for the OECD has been a heated exchange on the applicability of structural separation to telecommunications.

The United Kingdom has been an interesting test-bed for structural separation. There are good and terrible examples. In gas and electricity distribution, there has been some success in generating competition, without any obvious adverse effects. However, the same move in the railways, the creation of Railtrack, was a disaster. It was so bad as to require the government to re-acquire the company. It may be some years before the UK rail sector is sorted out.

At one time there were merchant bankers looking for business. However, they seem to have gone off to other projects.

A slightly distrubing development is the thought that the Loop Company might be a vehicle to carry debt. Thus, the French and Germany governments could transfer a lot of debt from their incumbent operators. This would be very unwelcome.

It would be hard to achieve structural separation cleanly. A legal instrument would be tough to construct that allowed some freedom to the Loop Company, but constrained it from competing with operators.

The position seems to be that, in the light of the OECD debate, both major Australian political parties have concluded that this is a bad idea or at least not the quick fix to the broadband policy dilemma. Thus it is off the agenda. There are few countries where it seems to be likely to be implemented in the near future. Either the political clout of the incumbent operators or the prospect of a painfull implementation have caused it to be held in reserve.


Other techologies

The story of the Wireless Local Loop (WLL) has been disappointing. It failed in the UK as Ionica and in many subsequent attempts in a varoety of countries. Yet, there is no clear explanation as to why.

One possible cause is that the high variable costs, for site acquisition and preparation, make it uneconomic. This may be compounded by problems of obtaining consent for equipment on shared buildings.

In a small number of countries there have been attempts at Ethernet To The Home (ETTH). Notably, in Sweden and to a lesser extent in Italy. These remain experimental, though nothing should be done to preclude such options.

In the Italian case, the mandatory sharing of ducts appears to offer a valuable policy tool.

In some countries there have been trial on the deployment of fibre direct to the home. However, the economic and business case seems to be very problematic. The benefits necessary to justify the installation of a new cable simply cannot be found.

In Japan, the deployment of fibre optic cables has been a long-standing government policy. From January 2002, the number of subcribers rose from 12,000 to some 230,000 at the end of January 2003. This compares with two million cable subscribers and six million ADSL subscribers. It is about three weeks growth of the ADSL subscriber base.

Wireless LAN has been ceased on my many people as a solution, often to problems which have many other and easier answers.

WLAN as an add-on to ADSL provides an excellent option for residential subcribers or for small retail outlets. It also has a role in community applications, including rural areas.


Broadband economics

An obvious question is why a government might want its citizens to have broadband access, other than to look good in the OECD statistics.

One answer lies in teleworking. Here there are considerable environmental benefits from reducing traveling and commuting. There also positive employment effects, by bringing into the workforce people on the margins.

In terms of Foreign Direct Investment (FDI) you can hardly claim to be trying to attract high-tech industry if the best home offer for their workers is ISDN. You need to have always-on, flat rate broadband at high speeds and low rates.

Another consideration is the development of the next generation of applications. In the first instance this will be for fixed networks, but clearly with a view to rolling then out to mobile networks as they become available. Korea, for example, has hopes to taking a share in the games market, at the expense of Japan and the USA.

Some countries, want to exports broadband equipment, such as routers and modems. Others want to sell the next generation of domestic appliances which are Internet-enabled.


Voice over the Internet

Ironically, one of the killer applications of broadband is voice. So much so that manufacturers have developed specific devices for this. Devices you plug into ADSL modems to access the PSTN at some very low prices.

Yahoo BB! offers US$ 0.02 per minute for calls to the Japanese PSTN. To keep life simple, the offer the same rate to the USA. This has now been matched by NTT.

The operators face the imminent threat of reduced revenues for telephony. This will not be compensated for by increased volume of calls, there is a limite to the number of calls we can make. Consequently, with falling prices, we will see a decline in total revenues from the voice telephony. It is merely a matter of time, when it cuts and how quickly it cuts over.

We also face the possibility of an alternative to mobile voice telephony as devices are developed allowing Voice over WiFi and other non-GSM and non-CDMA networks.

The "creative destruction" caused by the application of new technology continues.

Mobile broadband

In one sense it is tempting just to dismiss mobile telecommunications, over the failure to develop usable networks and plausible models for 2.5G and 3G.

The mobile operators, insofar as they are anything, are financial engineers. They knew once how to make the financial markets happy and supply them with ever rising share prices.

If they understand mobile voice telephony it was based on market abuses in call termination markets and international mobile roaming.
Their offers in GPRS are appalling. You require a university degree in computer science to configure the devices. Then you have to play very high rates.

Vodafone rates are £1 per megabyte in the UK and £10 per megabyte in the handful of adjoining countries where they offer the service.

Compared to the prices for Wireless LAN, this is a "no brainer". You go for WLAN and wait until the mobile network operators move the decimal point in their prices.

So that whatever the potential of 2.5G and 3G we have to discount it to zero.

Rural Australia

The technologies and economics of broadband mean that provision in sparsely populated rural areas will be more expensive than in the CBDs of the five large cities. This fact may be uncomfortable for those who live in the outback, but for the present it is the economic reality.

The crucial question is the willingness of politicians to ensure the provision of broadband and to pay for it.

It could be added to the universal service obligation and Telstra could be forced to deliver it. However, that not be cheap. To make matters worse, it would further distort competition. On the basis of Telstra's showing to date, the form book says it would not deliver.

It is hard to imagine some entrepreneur turning up and offering to create The Murray-Darling Cable Company or the Priscilla Queen of the Desert Broadband Company. Laying fibres to Ayers Rock would require a special sort of insanity.

There is a role for government, state as well as Commonwealth, in the aggregation of demand. In taking the needs of all government offices and functions in a given area and inviting businesses to bid for it by competitive tender. It is important to ensure such tenders are on a scale to allow in new entrants and do not merely serve to further secure the position of Telstra. It is absolutely essential to ensure that markets are not distorted.

With a secure public contract, a private supplier can justify infrastructure and can build out its network. It can then make a commercial offer to businesses and individuals.


Conclusion

As a set of technologies, broadband Internet access is certainly and justifiably a major concern of policy makers. The key technologies are, for the present, ADSL and cable modems. Soon there will be more advanced xDSLs, such as Video DSL. There is also Wireless Local Loop (WLL) and Ethernet to the home (ETTH) or business (ETTB). In some cases there may by Wireless LANs, though in more confined areas. The nature of the mix varies from country to country and from CBD to outback.

In the medium term there is the possibility of mobile broadband, but that relies on the strange and increasingly depressing saga of 3G. The chances of that being sorted out in the near future are minimal, it will not happen this year and possibly not next year. We can safely discount 3G for the present.

Already there are 30 million customers worldwide for fixed broadband Internet access, and that number will double by the end of next year, then double again. The bandwidth available to those customers is also growing. A 2003 offer should be around 9Mbps and next year higher still. South Korea is already running trials of VDSL at 50Mbps. You need that to make your plasma screen television work

The geographic centre will be somewhere around the Yellow River. Indeed, the centre of the world of telecommunications slowly moves to North-East Asia.

It is abundantly clear that there is no single solution, there is no recipe to follow. Each successful country has pursued its own route.

Where we find commonality, it is in the failure which is characterised by refusal to compete, by bloody-minded if creative obstructionism, described as 3D: deny, delay, degrade. It is the incumbent laying seige to its competitors until they starve to death.

There are new applications and new services, providing additional value for customers being built on those networks. That work is being done first where there are large markets, in North-East Asia. Australia is, like Europe, being left out. We cannot say left behind, since we do not know if it will be able to catch up.

We are also see value being destroyed in the declining price of voice telephony. Customer spending moves to new applications and operators have to recognise that. They simply cannot defend those revenues, other than my market abuses and even then only in the short term.

Investments in broadband technologies are generating new economic growth. The new application are generating even more important growth on top of that. This is an important

Thank you very much for your attention. A copy of my presentation will be available on www.intug.net in the next few hours.


copyright © INTUG, 2003.
 http://www.intug.net/talks/es_2003_03_rmit.html

Last updated 3 March 2003.