Leased lines currently provide critical building blocks for the Internet. A recent survey undertaken by Matrix Information and Directory Services (MIDS) found that by far the majority of organisations access the Internet via 56 kbit/s and 64 kbit/s connection speeds with about 18 per cent each, or 36 per cent of all connections (Table 8 ) .i The authors or the MIDS study noted that once an organisation has reached the point of registering a domain and making itself visible on the Internet they were likely to have dedicated capacity.ii This appears to indicate that access to the Internet via 56 kbit/s or 64 kbit/s leased lines is currently the most common way for organisations to provide services and content over the Internet. Some organisations are accessing Internet via ISDN at 64 kbit/s but they are still in the minority.iii
This evidence suggests the pricing of access to the Internet via leased lines at 56 kbit/s and64 kbit/s is currently critical to the development of services available over the Internet. In some cases thisis because bandwidth at higher speeds is not available, or not available at affordable prices. As new applications are developed leased lines with higher access speeds can be expected to become more in demand, as evident from the popularity of 1.5 Mbit/s lines in North America and to a lessor extent 10 Mbit/s leased lines.
A major problem is that Internet access at speeds higher than 2 Mbit/s in Europe appear not to be generally available at attractive prices. In fact the different rates of access via 1.5 Mbit/s lines marketed in North America (12 per cent of respondents) and 2 Mbit/s lines in the rest of the OECD area (4 per cent of respondents), recorded in the survey shown in Table 8, is remarkable. It suggests that capacity at higher speeds is far more available in the competitive markets of Canada and the US. A recent study by DANTE investigating the non-technical barriers to developing high speed backbone pan European networks concluded:
“The major concern of European PTOs with regard to providing high speed leased lines is their selling of this bandwidth for carrying voice circuits. Up to bandwidths of 2 Mbit/s the cost per voice channel is sufficiently unattractive for reselling; 34 Mbit/s lines can carry approximately 1000 voice circuits, which makes reselling attractive if the costs are considerably less than 17 times the cost of a 2 Mbit/s circuit. Since international voice telephony is a highly profitable business, the PTOs try to keep possible competitors out of their country either by not selling the bandwidth required or not at reasonable rates. Their monopoly status enables them to do so.”iv
Similarly Wim Vik from EUnet, a leading European IAP, has stated:
“We pay ten times or more than ten times as much for the same transport in Europe vis-à-vis the United States. We do not have commercially available 34 Mbit/s links within Europe. It is starting to be provided in Europe for the academic networks, but if we are talking about multimedia and other applications which we want in Europe, in order to be able to compete with our American friends, we are at a very large disadvantage vis-à-vis our American competitors.”v
If the pricing of capacity at the 2 Mbit/s level and above is dictated by concerns about the development of competitors, including IAP providers, this forces up downstream prices for users. Much of the initial demand for high capacity links to the Internet backbone networks has come from the rapidly growing IAP industry in Canada and the US, and its ready availability enables IAPs to pass on lower prices to their customers accessing the Internet at lower speeds. These latter users make up the majority of organisations on the Internet. The MIDS study concluded “...most organisations on the Internet have small numbers of computers and people. A few organisations have very large numbers of computers and people, but they are rare, just as Fortune 500 companies are rare. The Internet consists mostly of small organisations, not large ones.”vi Accordingly the following sections contain an examination of pricing via 56/64 Kbit/s leased lines and ‘dial-up’ services over the PSTN.
Table 1 . Internet Access Connection Speeds, 1995
The reason Helix can offer a less expensive service is because it purchases capacity from are seller of BC- Tel’s network in British Columbia. If best practice pricing is defined as the lowest available prices then the suppliers in Canada (US$7 000) and Finland (US$6 000) appear to have set the current benchmarks. This finding is confirmed from within the Internet access industry by EUnet’s Wim Vik who says Finland has the least expensive prices in Europe because of the competitive provision of infrastructure. viii In fact very little separates the prices of these countries when it is considered the Canadian charge includes a router at the customer site. By way of contrast with the low prices found in Finland and Canada, ten OECD countries had a leased line basket price per annum higher than US$10000, five a basket price higher that US$20 000 and a further three with charges higher than US$30 000. Overall the average price for leased line access to the Internet in countries with telecommunication infrastructure competition was 44 per cent less expensive than countries with monopoly provision of infrastructure.
The approach employed in the comparison in Table 9 , consistent with the OECD’s price comparison methodology in other areas, was to use the lowest price available to a customer for leased line access to the Internet. In other words if suppliers gave customers the choice of renting or purchasing equipment, as an unbundled charge, the price of the option excluding equipment (router or other) was selected. Some IAPs bundle the cost of renting a router in their overall price and do not list a price for customers wanting to purchase this equipment separately. The IAP supplier chosen to represent the US, Netcom, offers connection for $1 200 without a router and $3 400 with a router included. As the connection charge is spread over three years in the basket, this charge does not have a large impact on the total basket for those operators that bundle the purchase of equipment with their overall connection prices compared to those IAPs that unbundle these charges. In general the monthly rental charge is much more influential in determining the cost of an Internet access basket.
It is interesting to note that in eight of the ten least expensive countries, suppliers quote prices that include all PTO charges. By way of contrast in only one of the 15 more expensive countries—Austria—does a supplier include PTO charges. The most likely explanation for this trend is that the IAPs bundling PTO prices do not view them as an obstacle to customers purchasing access or that they can provide their own connection facilities. In fact five of the eight OECD countries that allow telecommunication infrastructure competition have amongst the lowest basket of Internet access prices for leased lines at 56 kbit/s and 64 kbit/s. Finland and Canada have the lowest prices, while the UK and Sweden are ranked sixth and seventh in Member countries. The US is ranked eighth and although less expensive prices were found for some IAPs based in particular cities it was decided to use Netcom because it is a national supplier. In October 1995 Netcom had 232 000 subscribers and more than 200 points of presence in the US. The pricing structures of the other countries that permit infrastructure competition, Australia, Japan and New Zealand, appear to be more expensive but this can depend on usage patterns. The geographical location of Australia, Japan and New Zealand may contribute to higher international leased line charges, for backbone Internet capacity, that are passed on to IAP customers but it is difficult to determine how much impact this factor has on the overall price. It should be noted that the underlying prices for international leased lines between these countries and major Internet gateways would-be bundled in IAP charges rather than the domestic leased line charges shown over 2 km in Table 9. In other words some part of the higher cost of the IAPs shown in Japan and New Zealand would reflect the underlying charges of PTOs such as KDD or Telecom New Zealand. In the case of Telstra it would be an internal cost to the company and presumably reflect tariffs charged to ordinary customers of international leased lines. In November 1995, Telstra increased the international capacity of its network from 6 Mbit/sto 10 Mbit/s.ix
The IAPs in countries with monopolies over telecommunication infrastructure competition which bundle PTO charges in the price to users, all have PTO charges for the provision of national leased lines at 64 bit/s that are lower than the OECD average. In particular, Iceland, Switzerland and Ireland have among the least expensive prices for 64 kbit/s leased lines over short distances and this contributes to the irrelatively low basket price. To construct the 56/64 kbit/s basket shown in Table 9, where IAPs did not bundle PTO charges, standard leased line prices at two kilometres were applied to represent this cost.
In a competitive market, such as Australia, it is necessary to ask whether customers do actually pay these prices. Telstra’s prices were selected to represent Australia because the company is a national supplier. Since there are two major infrastructure suppliers (Telstra and Optus) and numerous resellers, Australian IAPs do not necessarily have to pay Telstra’s charges. This is the main difference between a competitive and a monopoly market in terms of Internet pricing. In a competitive market the IAPs can simply opt for an alternative Internet backbone provider. So for example, MCI and IBM have Internet connections/subscriptions in the US and use their international communication links to connect to their Australian facilities. The extremely important feature of a liberal market is that the incumbent can not control the pace at which service is rolled out in an anti-competitive manner. Where a PTO has monopoly power over providing the underlying infrastructure for the Internet backbone network, all IAPs would have to pay that PTO’s prices and they would set a baseline for IAP charges. In terms of market development this is crucial because if a monopoly PTO was not in the position to offer service, or did not want to offer more than a limited service, they could strongly influence market growth.
Some may ask why a monopoly PTO would want to do this given that an expansion of the Internet market increases the size of the overall communication market. There could be many reasons why a monopoly PTO would want to retard growth. For example, in August 1995, when the OECD surveyed Internet provision in Member countries it was not possible to obtain ‘dial-up’ Internet access as a service from many PTOs. Since that time a growing number of PTOs have announced they will launch a service or have commenced by offering leased line access with a view to entering the ‘dial-up’ business. In the interim, PTOs in monopoly markets have the option, if they so choose, of holding back the growth of their future IAP competitors by charging high prices for capacity. This may be as simple as not introducing a reduction in prices that would otherwise have been made or delaying the installation of a new order for a leased line. This is not likely to happen in a market with infrastructure competition and vigorous national and international resale.
Much of the discussion of Internet charging practices has focused on the advantages and disadvantages of
flat-rate and usage based pricing. As indicated in Table 9 , in most OECD
Member countries it is possible to obtain a 56 kbit/s or 64 kbit/s leased line at a fixed rate per month with
unlimited usage. For the purpose of this comparison flat rate pricing options for 56 or 64 kbit/s access to the
Internet via leased lines were selected. If a flat rate pricing option was not available from the suppliers
contacted the lowest cost option is shown. Accordingly the baskets shown for Australia, New Zealand,
Greece and Portugal are not wholly comparable with the other OECD countries.
Technically, some may view Telstra’s charge at 100 per cent utilisation of a 64 kbit/s leased line Internet
access as a maximum flat rate charge (Table 10 ). This charge, and others for
high rates of utilisation, are presumably set at a punitive level to encourage IAPs to shift to higher capacity
leased lines, once they cross certain volume thresholds, and for end users wanting high speed access to
utilise ISDN services. Indeed the theoretical price a customer would pay for 100 per cent utilisation of a 64
kbit/s leased line is US$ 80 000 per annum. Instead what the data for Australia, and the other countries with
usage based charging show in Table 9 , is the minimum price that can be paid by
customers if they only use their 64 kbit/s access up to the level where additional usage charges would be
incurred.
The main impact of Telstra’s charging practices, in line with their underlying philosophy of using volume
based charging to discourage saturated international links (see below), is that IAPs orient their services
toward selling ISDN access to business customers rather than flat rate unlimited access via leased lines. A
random survey of the home pages of several IAPs in Australia—Acay Network Computing, Access One,
AUSNet, Hutchinson Telecommunications, and Netspace, showed that their prices generally had a volume
component for ISDN access to Internet. Only Access One offered 64 kbit/s ISDN access to the Internet
without charges increasing beyond a certain utilisation level (US$1 838 for installation and US$827 per
month).
In November 1995, more typical of Australian IAP prices, was Netspace, which was offering 64 kbit/s ISDN
connections to the Internet for US$1 838, and a monthly charge of US$484 for 30 per cent utilisation levels
(US$827 with a 60 per cent utilisation). For semi-permanent ISDN 64 kbit/s links, Hutchinson
Telecommunications charged US$1 838 for a connection and US$514 per month. Naturally, these charges for
ISDN access are much lower than the ones shown for Telstra in Table 10 . IAPs
are able to do this because they purchase capacity from Telstra, or another supplier, and resell this capacity
among a number of customers. For example, an IAP might purchase a 128 kbit/s connection for US$3 680 per
month. If this link was resold to two business users (both with 64 kbit/s ISDN access) at US$500 per month
and 200 dial-up users at US$30 per month the IAP would receive US$7 000.
If the IAP access prices via ISDN connections were used to represent Australia, it would improve that
country’s position relative to other countries in Table 9 , and significantly lower
the average price for countries with competitive markets. In one sense the prices of the Australian IAPs are
closer to those available in Canada and Finland than a initial examination of Telstra prices would suggest,
although customers would also need to pay Telstra charges for ISDN transmission between their facilities
and those of an IAP. On the other hand if the leased line prices available in Canada and Finland were
matched by Telstra or other Australian backbone Internet providers, IAPs in that country could offer lower
prices. How much the Australian pricing has to do with geography, rationing use, encouraging balanced
traffic flows or other factors are the key questions. It is also true that in August 1995, Telstra had only
recently taken over the provision of AARNET, an Australian Internet access provider for the academic
community, and the inherited tariff structure may not reflect future pricing plans. In particular low bandwidth
(64 kbit/s) seems relatively expensive compared to Canada and Finland while higher bandwidth (2Mbit/s)
relatively inexpensive relative to many parts of Europe.
If using baseline prices had placed Australia (Telstra), New Zealand, Greece and Portugal at an advantage,
in comparative terms, it may have been necessary to develop an assumption based on an average usage
pattern to compare costs. This was not necessary because the baseline costs for permanent leased line
access in Australia, Greece and Portugal are relatively expensive even before usage charges are considered.
On the other hand, for comparisons between these four countries it should be recognised that New Zealand
is very much advantaged relative to Australia, Greece and Portugal because it has far less usage included in
the fixed rental. Telstra’s charging structure is different from most other IAPs because it measures the
average port utilisation of a leased line customer rather than expressing usage by the amount of data
transferred. If a user in New Zealand, using ICONZ as their IAP, made similar use of their line to a Telstra
customer, their charges would also be similar (in practice this is difficult to determine because ICONZ
charges different rates for local, national and international traffic).
There are IAPs in several countries shown in Table 9 with flat rate pricing that also offer usage based
charging, such as Canada, Italy, and the Netherlands. Given certain usage patterns these options could be
less expensive for customers. For example one option available in the Netherlands would be less expensive
than the option shown in Table 9 if a customer used the leased line at a level
equivalent to that included in Telstra’s charges. In the case of Turkey the maximum charge that can be
incurred by a customer is shown in Table 9 , although a user could pay less
based on levels of usage much lower than included in Telstra’s charge.
With the commercialisation of the Internet the choice of pricing structure has become a major issue. Should
IAPs charge a flat rate, one based on utilisation, or a combination of the two. Generally PTOs prefer
bandwidth or time based charges for most types of telecommunication services because it encourages users
to ration network use. It is also said to allow operators to better predict patterns of use which enables more
effective network planning. On the other hand many users prefer unmeasured pricing because they know in
advance how much they will pay and, in the case of traditional leased lines, flat rate shave represented a
significant discount on standard PSTN charges beyond a certain point for the same volume of service.
If IAPs choose to charge for usage it raises the question of which aspects of usage will be charged and
how this will be measured. Telstra’s pricing is based on bandwidth and volume charges(measured by
average port utilisation). Users only pay for utilisation for ‘incoming traffic’ rather than ‘outgoing traffic’.
According to Telstra, “The new tariff is designed to encourage Australian content provision to a global
market, by not charging customers for sent traffic. The tariff has been structured to charge for received only
traffic, to be fully distant independent (including both Australian and International traffic) and truly make
Australia a competitive source of information content.” x If the Internet
becomes a significant market place for electronic commerce such a pricing policy may have a number of
advantages for the export of information services. On the other hand it might be pointed out that a flat rate
tariff for leased line access would achieve the same effect. Yet, over the longer term, if some type of
international settlements scheme was ever implemented for the Internet, Telstra’s pricing scheme may
influence user behaviour by encouraging Australian users to ration consumption of incoming international
traffic. Under a flat tariff structure this would not be a consideration for users.
Aside from the method of charging the main issues for Internet access are the level of price and the quality
of service. Just because there is a certain price structure for network use does not mean it is expensive or
inexpensive. Australia, Japan and New Zealand all have relatively expensive IAP charges for dedicated
leased line access with three different price structures. What impact geography has on the relatively higher
prices due to the cost of international backbone networks would require further analysis. On the other hand
flat rate pricing is no guarantee of low charges as evident in countries such as Austria, Germany, Italy and
Spain. Here the problem is not geography but the relatively high prices charged for international leased lines
that do not need to bridge continents to access major Internet gateways. At the same time these issues can
not be considered apart from the quality of service.
Depending on developments in areas such as network capacity, compression technologies and new applications (such as bandwidth demanding video services) such questions may come increasingly to the fore. This raises some interesting issues for policy makers. If regulation restricts the provision of telecommunication infrastructure to monopoly or duopoly providers, it is necessary to ask the whether it is acceptable for PTOs to ration service via price if other potential infrastructure providers are not allowed to meet a need in the market. In other words, if PTOs can not provide infrastructure, or will not provide infrastructure because of the potential for competition with their traditional sources of revenue, is regulation creating an artificial scarcity of connectivity? If more than 14 Internet backbone networks have been able to develop in the US, in a commercial environment, and relatively inexpensive flat rate tariffs are available to business users for high speed leased lines, it raises the question of why it is difficult to get equivalent high speed links in Europe at the same price.
As with other telecommunication services, price comparisons need to be complemented by quality of
service analysis. A feature of Telstra’s Internet access pricing agreement with customers is that the
company will guarantee certain performance levels. According to Telstra its Internet Network will be
engineered to provide a network service availability of more than 99.75 per cent, with a 2 hour response time,
24 hour repair time service for customers. Telstra is also providing customers with monthly reports on traffic
statistics, network availability, faults and restoration. In respect to backbone networks:
For Internet traffic within Australia Telstra will provide a domestic Internet service that is designed to
carry at least 98 per cent of network traffic between any two Telstra Internet Points of Presence with a
transit time not exceeding 300 milliseconds. This includes transit and queue times at any intermediate
points in the network. This measurement will be determined by random sampling conducted by Telstra
over each week of the domestic network over Telstra’s points of presence.
For Internet traffic, between the international gateway in Australia and the international gateway in the
United States, Telstra will provide an International Network Service designed to carry at least 95 per
cent of network traffic between Australia and the United States with a transit time not exceeding 300
milliseconds. This measurement will be determined by random sampling conducted by Telstra over each
week of the international network between the Telstra Australian Internet gateway and the international
Internet gateway located in the United States. xii
Table 3 Telstra Internet Access Pricing, August 1995 (US$PPP)
The charging practices and prices for ‘dial-up’ access to the Internet vary enormously throughout the OECD area (Table 11 ). In general the experience of OECD is that capacity, particularly high speed leased lines, is less expensive in those Member countries that allow telecommunication infrastructure competition. xvi Trying to compare the underlying cost of capacity for IAPs to access backbone networks would be extremely complex. For example it would need to take into account the cost of leasing capacity from the IAP facilities to those of the PTOs and, in some cases, the charges applied by the PTOs to IAPs based on the volume of traffic. For larger IAPs, that directly lease international capacity, the cost of international leased lines (which include the prices of two PTOs), would also need to be considered. In those cases where the PTOs are also the IAPs, it would be necessary to determine whether they paid, for the purposes of internal accounting, the prices listed for use of their own infrastructure.
Undertaking comparisons throughout the OECD area would be further complicated by geography. In other words it might be expected that the relatively high prices for international leased lines would be magnified in those countries with few near neighbours compared to those with many. While this would reveal the actual costs to users it could highlight geography rather than efficiency. Finally a comparison of listed prices would not take into account the discounts that tend to be more widely available in competitive markets, or the availability and quality of provision of high speed capacity.
Despite the difficulties in undertaking accurate comparisons of the underlying infrastructure prices paid by IAPs it is possible to compare their prices to end users for ‘dial-up’ services with the underlying market structure. Table 12 shows a marked correlation between lower IAP prices for ‘dial-up’ services and the liberalisation of underlying telecommunication infrastructure provision. Six of the eight countries with the lowest cost for a basket of IAP ‘dial-up’ charges allow telecommunication infrastructure competition—UK, NZ, Finland, the US, Canada and Sweden. If Australia’s November 1995 prices were considered it would join this group. On average IAP prices for ‘dial-up’ services are nearly three times more expensive in countries without underlying telecommunication infrastructure competition than those with competitive markets.
Geography could be a factor in the two countries with competitive markets that were relatively expensive— Australia and Japan—in August 1995, where because they have few near neighbours there may be higher international cost components. That being said, if this is a factor, it might be suggested that the charges reflect price structures for leased lines derived from PSTN rates for international telephony, that now bear little relation to the cost of providing international capacity. More importantly prices are now spiralling downward in both countries. In Japan, prices were brought down dramatically between August and November 1995. In fact the price of Fujitsu’s InfoWeb service in November 1995 was roughly half the price for Tokyo Internet recorded in August 1995.
In the case of Australia, competition among IAPs is rapidly reducing the price of access. By November 1995,
Acay Network Computing, the company shown to represent Australian pricing had introduced an offer of
US$29.41 per month which included 40 hours of use (with additional hours priced at US$1.47). Clearly this
price is much less expensive for 20 or 30 hours of service than the August 1995 Acay price shown in Table 11 . This change was not due to a reduction in Telstra charges which
remained at the same level as Table 10 , and was presumably in response to
demand for consumers for a flat rate pricing structure and an increasingly competitive IAP market.xvii An
important point, which is often overlooked in theoretical discussions of Internet pricing, is that in a
competitive market operators will respond to the needs of the market in terms of pricing structures and
making new facilities available.
Between 1994 and 1995 the number of IAPs is reported to have doubled in Australia from around 50 to 100.
One of these IAPs, Netspace Online systems, launched a service option in October 1995 priced at US$10.40
for 20 hours of use per month and US$16.54 per month for 40 hours of use for non-commercial users. As at
November 1995 these were probably the lowest prices for Internet access in the OECD area. Because
Australia has a liberal telecommunication market the relatively high underlying Telstra listed prices seem to be having little impact on
IAP prices in Australia. In a policy environment allowing infrastructure competition and national and
international resale, IAPs can purchase capacity from alternative suppliers. In the US, the market had
become so competitive by October 1995, that Pipeline USA, a PSINet company was offering unlimited
Internet access for flat rate of US$19.95, including all software, no connection charge and a seven day free
trial. xviii
It is also true that in some countries with monopoly markets for underlying infrastructure, but competitive IAP markets, the prices are being reduced. For example, in Spain the initial IAPs charged very high prices without really differentiating the market between business and residential users. Inpractice, of course, this meant the service was available only to business or individuals prepared to pay relatively high prices. The EUnet price shown in Table 11 to represent Spain was selected because the EUnet network prices were readily available for most European countries in August 1995, and provided a European wide perspective on pricing. However in early August 1995 the only ‘dial-up’ tariff package available to EUnet users in Spain— ’MyEUnetDial’—was relatively expensive compared to most packages aimed specifically at the mass market.
At the beginning of September 1995, EUnet (Spain) revamped the pricing of its ‘dial-up’ option and aimed it more specifically at business users. Prices were actually increased for connection and the monthly charge doubled (with the time included increased from 3 to 6 hours) with additional hours beingpriced at one rate (instead of peak and off peak). For the general public EUnet (Spain) then recommended Aba forum, another IAP, and its prices are much more in line with the experience of most of the rest of the OECD (although around three times as expensive as Netspace in Australia). As with Australia and Japan it is tempting to include the latest information, and it is shown for these countries, to demonstrate thedynamic nature of change. However, like the Australian and Japanese data, the Aba forum prices are not included in any of the OECD averages because they are for November 1995.
The other OECD country with few near neighbours, New Zealand, has relatively low charges but ‘dial-up customers’ passing average use of a residential line for telephony rather than accessing on-line services. For example telephone users in Japan spend on average only 10 minutes and 36 seconds per day on the telephone.xix
Because this is a national figure that includes business traffic the average for residential users may be lessagain. In the case of on-line services it would be expected that a user would make less calls of a much longer duration than for telephony. Moreover, instead of being spread over different distances calls would all be made to a point of access within a local calling area. While the latter assumption may not hold truein all cases, especially for some inhabitants of rural and remote areas, it would be the case for most usersin a major city.
The cost of accessing on-line services outside local calling areas would be prohibitively expensive for calls longer than the average telephony usage based on standard long distance charges. Accessing IAP and ISPs by 800 numbers, while less expensive than long distance rates, is also relatively expensive compared to local calls. For example, Concentric Network, a large US IAP, charges a subscription of US$10 to access the Internet via an 800 number which includes 2 hours on-line time per month. xx After the initial 2 hour period customers pay US$5 for every additional hour. America Online charges US customers a surcharge of US$6 per hour to access its network via an 800 number. xxi
This raises many questions for policy makers in terms of universal service and regionaldevelopment. If providing the widest possible access to information infrastructure meant that rural usersshould pay the same rates as those in urban areas, this would not be possible within the structure of traditional telecommunication charges when there was not a local Internet access point of presence. Indeed traditional telecommunication charging practices might discourage business from locating outside urban centres, employees from opting for tele-work, or rural communities and residences from benefiting from services available to users in cities at more affordable prices.
In times past universal service issues such as these have been used by many PTOs to justify monopolies in telecommunication infrastructure provision. Today the most innovative solutions are beingdeveloped in competitive markets. For example Demon Internet, the largest provider of Internet access inthe UK, has introduced a service which enables any user in the UK, irrespective of geography, to access itsservice for the price of a local call. xxii For this service, customers call via BT’s local access network to Demon Internet Numbers in every local charge group area via ‘virtual points of presence’. These numbers are easily identifiable since the last three digits are ‘666’ in each zone. The traffic is then carried by long-distance service providers, (Energis or Mercury) to Demon Internet’s operations centre which houses its modem stacks. Clearly such a pricing structure has major advantages for customers in rural areas in terms of universal service, and rural users in monopoly markets are going to justifiably ask why they can not obtain similar benefits.
The charging practices for local calls for residential users are set out in Table 13 . The basket of calls shown inTable 14 includes 20, 30 and 40 local calls of one hour duration per month at standard rates. In other words a user is assumed to be on-line for between five to ten hours per week. The Jupiter/Yahoo survey found that the average on-line user spent 80 hours per month and that the average session lasts slightly loner than one hour (75 minutes). xxiii This may suggest that a comparison of longer periods than the one chosen by the OECD should be made. On the other hand it is also necessary to remember that this is an average that takes into account the use made by employees during office hours, including access via dedicated connections, and that based on PSTN and IAP pricing the experience throughout the OECD area is likely to be very mixed.
Moreover it has been found that surveys of Internet use, that are made over the Internet, tend to overestimate average on-line time. The CommerceNet/Nielsen Internet Demographics survey of the US and Canada, found that the average Internet use for individuals was just under 22 hours per month. xxiv The CommerceNet/Nielson study found that average use of on-line information services was just under 10 hours per month (e.g. CompuServe, America Online). The evidence of this survey suggests that baskets of Internet and on-line information service use between 20 and 40 hours provide a reasonable reference to the prices paid by average users. In future more data will be available on Internet usage patterns because the advertising industry wants to build a picture of on-line use to increase the effectiveness of their expenditure. In a similar manner to technologies that measure actual viewing patterns of television, the first ‘PC-meter’ software is now being implemented in the US, among participating consumers, to build up pictures of on-line use.xxv
Many residential users would, of course, take advantage of off-peak rates to reduce their PSTN on-line time and consequently the cost, but a more detailed knowledge of usage patterns would be necessary to construct a basket of proportional use between times with standard and off peak rates. Table 15 shows the same basket as Table 14 with PSTN charges at off-peak rates. The off-peak rate selected was the one available in OECD countries at 20:00 hours. For some countries the time of day does not matter because there is a flat rate or an unmeasured rate (Australia, Canada, Greece, Mexico, New Zealand, and the US (Florida)). On the other hand for Japan, Spain, Switzerland and Turkey the time is significant because major PTOs in these countries do not offer an off- peak rate at 20:00 hours. xxvi
The other important information which needs research and analysis for comparisons of PSTN and Internet access is the impact household demographics have on on-line time. This would be true not only for the number of people but which members of a household accessed services at what time of day. Analysis of this type although still scarce can produce surprising results. For example, CompuServe conducted a survey of women’s use of on-line services in which 5 000 subscribers participated. xxvii Nearly 25 per cent of CompuServe’s members world wide are women, although of course, many more would access services through a subscription in another name (e.g. a spouse). CompuServe reports that the survey found that, on average, of the leisure hours women subscribers have per day they spend more than four hours of that time on a computer. The seemingly high rate of computer use may indicate that these women spend more than the industry wide average of less than 30 minutes per day accessing the databases of information service providers. It is also indicative that on-line surveys produce results of the most enthusiastic early adopters of computer technology and services.
The features women respondents to the CompuServe survey ranked most highly in what they wanted from an on-line services were in areas such as children’s education and health. Ranked number one of major concerns for respondents was the importance of children being comfortable with computers and other types of technology. While these results provide only one perspective on what users may be looking for from the world of on-line services and the Internet, they nevertheless suggest caution in simply assuming that analysis of off-peak rates is sufficient. While users will undoubtedly take advantage of off-peak rates for work and leisure related activities the needs people have for information infrastructure pricing are likely to be very diverse.
Accordingly there may be a number of reasons why ‘standard rates’, and off-peak rates no later than 20:00 hours, are an appropriate starting point for analysis of pricing in terms of information infrastructure. For example, with the increasing interest in information infrastructure applications for:
tele-work, an examination of rates during ‘office hours’ is apposite (and evening rates for employees taking work home);
health applications, analysis of the tariffs for the hours pertinent to medical consulting need to be considered;
social policies aimed at improving the self reliance and home care for the elderly or disabled who may need to access services 24 hours per day;
delivery of public services may in some cases need to be effected during ‘office hours’;
education, where both the class room hours and the ‘home work hours’ after school predominantly fall in standard rate times (and to a lessor extent evening off-peak rate times);
managing households involves tasks that continue around the clock and may not all be able to be accomplished in off-peak times;
small business development because for business users without dedicated capacity most transactions will occur during ‘office hours’ or early evenings;
The results of analysis of PSTN charges for a basket of on-line services in Table 14 show an extraordinary range of prices for the same service. The tremendous differences in the cost of accessing the Internet, or other on-line services, via the PSTN are the result of pricing structures that, while not producing large differences in a basket of PSTN charges for average telephony use (mixture of calls over different distances, times of day and comparatively short duration), have a huge impact on the cost of on-line service access. Inevitably those countries with unmeasured local calls for residential users (Canada, New Zealand and some parts of the US) and un-timed calls (Australia, Mexico and Greece) have the lowest charges for accessing on-line service via the PSTN. These countries are followed by Finland, Iceland and Sweden who traditionally record low basket prices in OECD telecommunication comparisons. It is also note worthy that Finland has unmeasured calls during off-peak hours. Some of the more expensive countries, for standard PSTN charges, include Ireland, Switzerland and the UK because of its relatively high cost of local calls at standard rates.
Price rebalancing in these countries, and others with timed local calls, has contributed to the relatively high costs of their ‘on-line basket’ (see section below on rebalancing) and this issue is going to be at the forefront of discussion about consumer, schools, patients, tele-workers and small business access to information infrastructure. For example, users in Switzerland and Canada, countries with virtually identical levels of telephone penetration, would differ in the amounts they paid to the PTO by ten times for 20 hours per month of on-line access via the PSTN. At the same time, because the Canadian charges are fixed, the difference would increase to 20 times the price for 40 hours of on-line time per month. To put this into perspective the normal difference between the most and least expensive countries for an OECD basket of residential services is between two to three times the cost.
The dramatic difference in the results of alternative pricing structures for local calls is revealed by the domestic comparison of US charges. Table 14 shows the prices that would be paid in four States— California, Colorado, Florida and Texas where local calls are available at unmeasured rates compared to the prices of timed local calls for a user in New York City. The cost of a flat rate service in Florida is more than 10 times less expensive than a measured rate with 40 hours of local calls in New York.
To date, it has been the inherited local call pricing structure rather than competition that is the primary factor in determining the PSTN access cost of an on-line basket. Indeed, the country where local competition is most advanced—the UK—has relatively high charges for the PSTN component of an on-line basket at standard rates (Table 14). On the other hand the lower charges available for backbone capacity, where competition has had a far longer time to make an impact, in the UK have made it possible for the IAP component to be the least expensive in the OECD area (Table 12 ). It is also true that local call prices have been falling in the UK due to price regulation and local competition from cable telephony providers. As a result of these factors the UK has a price well below the OECD average for a total Internet access basket shown in Table 16 and the least expensive at off peak rates in Table 17 .
Several other interesting developments have emerged in the UK market where cable communication companies are providing competition in the provision of local telecommunication services. xxviii First is the innovation they have brought to the pricing of local services ranging from volume discounts, inexpensive connections, discounted service packages (which include cable television and telephony), and relatively inexpensive or free calls between their customers during off peak times.
By way of example Table 15 shows the pricing of Videotron, a UK cable communication company. Videotron’s off-peak rates are shown where customers (with and without CATV) can call other Videotron subscribers, which include a number of IAPs, at no charge. While the innovation of free cable to cable customer calls was introduced to attract telephony customers to the networks of cable companies, the option is proving popular with users wanting to access on-line services. In fact 35 per cent of Videotron’s telecommunication customers own a PC.xxix For Videotron customers uncharged local calls apply to voice and data telephone calls made in analogue form from a residential Videotron telephone customer to any other local Videotron telephone customer, between 7pm to 7am Monday to Friday, all day weekends and UK public holidays. As a number of IAPs in the Videotron area are also its customers, other Videotron customers can call these numbers free of charge during off-peak hours.
Some PTOs and IAPs have ambivalent feelings toward such options because users might tie up facilities for long periods with calls that do not generate revenue. In one trial of a flat rate tariff in Hull set-up by Demon Internet and Yorkshire Electricity, users paid US$0.09 per call for a unlimited connection time. xxx This reportedly caused problems for Demon Internet because a number of users established near permanent links which prevented other users from logging on. This is perhaps why some IAPs in a number of countries where uncharged or unmeasured local calls exist, initially preferred to charge for usage (Australia, New Zealand, Greece). While in practice there is no incentive with unmeasured calls (Canada, US) to remain logged on, because users do not incur a new charge for additional calls, new communication capabilities may act change this situation. As more communication facilities are incorporated into PCs, including faxes, answering machines, the Internet telephone and audio/video capabilities users may wish to establish near permanent links if there is flat rate or unmeasured pricing.
A competitive local market has also brought forth benefits for BT customers. For BT customers, even where cable has not yet passed their homes, the incumbent PTO has improved its performance. First is that a competitive market for local service, and price regulation via a price cap, has acted to bring down the cost of local calls in the UK. In early 1994 BT brought its morning peak rate down to the level of its afternoon rate to create one standard day time rate. Moreover competition has encouraged BT to develop an innovative range of tariff options for residential customers. Accordingly a user subscribing to BT’s premier line plan could benefit from a 15 per cent discount on call charges. xxxi This could be combined with a “Friends and Family” option that would provide an additional 5 per cent discount on five numbers. What this means is that users subscribing to these plans could pay 20 per cent less for calls to an IAP point of presence (Table 14 ).
In the future local competition is likely to be the key policy instrument used by governments in the OECD area to expand affordable access to information infrastructure. For the present it can be noted that those countries with low PSTN charges (where the primary influence has been the inherited pricing structure) and those countries with low IAP ‘dial-up’ prices (where a contributing factor appears to be infrastructure competition driving down the cost of the underlying backbone capacity) are best placed to expand access to many new ‘information infrastructure’ applications. In the case of the UK where the inherited pricing structure at standard rates appeared to be inappropriate for the widespread development of access to the Internet, competition is being applied as tool to encourage PTOs, cable communication companies and IAPs to come up with innovative solutions.
One of the more promising developments, in terms of technology, are modems for cable television lines which may allow ordinary subscribers to take advantage of the broadband capabilities of cable communication networks. xxxii A number of cable modem trials are underway in the US and PSI and Cablevision are reported to be offering a commercial service in Cambridge, MA. xxxiii In the US, Netscape has announced an alliance with TCI Technology Ventures to develop cable based Internet services for residential users. The two companies believe that the key to the new service will be the speed with which information is transported and estimate that transmission speeds more than 500 times faster than current Internet offerings will be available. Service is planned to be commenced in selected parts of the US during 1996. xxxiv In the UK, the Cable Communications Association, an industry group representing cable communication companies, has predicted that cable modems will be available for widespread use in early 1997.
The challenges before those OECD countries without competition in telecommunication infrastructure appear to be formidable. For the total basket of 30 hours per month of ‘dial-up’ Internet access (i.e. PSTN plus IAP charges) seven of the eight countries with infrastructure competition are below the OECD average while 12 of the 17 countries without infrastructure competition are above the same benchmark. The major barrier to a widespread expansion of ‘dial-up’ access to on-line services would appear to be the impact of PSTN rate structures based on telephony. For example, in 1995 the cost per annum of an average basket of residential telecommunication services in the OECD area was US$387(Table 18). The average price for ‘dial- up’ access to the Internet, including PTO and IAP charges, was US$1 290 (@ 240 hours per year) or US$1 655 (@ 360 hours per year) at standard rates. On average, on-line users would pay three times their normal telephone bill for five hours per week of on-line time and five times their normal telephone bill for ten hours on-line time per week at standard rates. The situation is ameliorated somewhat by off peak charges but, on average, the baskets still represent a large increase in payment by the users compared to the cost of traditional telecommunication usage patterns.
While users will undoubtedly be prepared to pay higher charges for services they value it is also true that a major examination of charging practices will need to be undertaken by many PTOs if they are to foster efficient use of information infrastructure. Markets can be used by policy makers to assist in this process and eight of the 13 least expensive countries are applying competition policy to encourage structural adjustment. This is not to argue that these eight countries still don’t face large challenges. For example only Canada, New Zealand and the US have a one to one ratio between a relatively inexpensive basket of residential PSTN services and on-line services (@ 240 hours per year) at standard rates. On the other hand the US performance varies where there are not unmeasured rates available for local calls(i.e. New York City).
Sweden and Finland, although having relatively inexpensive on-line baskets by comparison with other OECD countries, are relatively expensive at standard rates compared with their residential telephone baskets—which it must be said are amongst the lowest in the OECD area. Nevertheless both countries have amongst the highest growth rates for access to the Internet and rates are reduced considerably during off- peak hours. For Australia and Japan the major barrier would appear to be the high cost PTOs charges to IAPs for capacity rather than the cost of local access via the PSTN, which is relatively inexpensive in both countries. By way of contrast the UK has relatively expensive local calls for an on-line user but inexpensive IAP charges made possible by competition in the provision of backbone capacity. The approach of the UK, since the end of the duopoly, has been to apply competition to the local market with a view to gaining similar benefits to those achieved in the pricing of backbone infrastructure. In summary the group of eight countries with competitive markets are either starting from a point where they have relatively inexpensive backbone capacity or local PSTN service for Internet access.
On the other hand far greater challenges are facing those countries that have not yet moved to liberalise their telecommunication markets. Of the monopoly countries, perhaps only Iceland will start from a relatively advantageous position. The other country with a monopoly market which is enjoying high growth rates in access to the Internet, Norway, has relatively expensive charges compared to most countries with competitive markets. The Netherlands also has relatively low prices for those countries with monopoly markets. Greece and Mexico have an advantage with their existing price of local access but the lack of competition in the provision of backbone infrastructure appears to result in higher IAP prices.
Most other countries with monopoly markets have far more severe problems because they have relatively high local PSTN access charges for a basket of on-line services and high prices for backbone capacity that IAPs pass on to customers. Unless these countries implement policies to address this situation the economic and social benefits governments envisage for information infrastructure will be very slow to eventuate. In monopoly markets there does not appear to be any innovation or structural adjustment to address new growth areas. Instead, the tariff rebalancing (refer later section) that is occurring is making the local charges in these countries even more expensive relative to those with competitive markets. It may or may not be true that higher tariffs for local service better reflect cost causation but the problem with monopoly provision of local service is that there is little incentive to reduce these costs. Overall this considerably raises the cost of universal service and makes many of the aims governments have for information infrastructure prohibitively expensive in these countries.
Table 4 . ‘Dial-Up’ Internet Access Prices in the OECD Area, August 1995
Table 5 . Ranking Internet Access Provider Charges (‘Dial-Up’), August 1995, US$ PPP
Table 6 . Local Charging Practices for Residential Users in the OECD Area
Table 7 . Internet Access by PSTN (Standard PSTN Rates), January 1995, US$ PPP
Table 8 . Internet Access by PSTN (Off-Peak PSTN Rates), January 1995, US$ PPP
Table 9 . Total ‘Dial-up’ Basket (Standard PSTN Charges), US$ PPP
Table 10 . Total ‘Dial-up’ Basket (Off-Peak PSTN Charges), US$ PPP
Table 11 . Internet Access Pricing and Average Residential Telephone Bills (Standard PSTN Rates), US$ PPP
Table 12 . Internet Access Pricing and Average Residential Telephone Bills (Off-Peak PSTN Rates), US$ PPP
Box 2:
One interesting example of ‘Internet on-line banking’, in respect to access pricing, has taken place in the town of Apollo (population 2 300), near Pittsburgh, P.A., in the US. Apollo Trust Co, a community bank based in Apollo, is offering its customers free access to the Internet for one hour per day. Since the program started more than 900 people, of which 550 are from the bank’s 8000 customers, have joined the service. xxxvii Around 100 Apollo customers are using the service for home banking. The Apollo Trust Company provided the service to non customers for the price of US$100 per annum and it is free for customers. The bank has donated free access to the town of Apollo’s library, churches and two highschools and provides a bulletin board with a listing of community events. Prior to the bank offering its service, residents of the Apollo community could only access on-line services via the PSTN at long distance rates. This made access to on-line services very expensive for users in an area which has suffered substantial job losses in recent years due to restructuring in the steel industry.
The primary motive for the bank providing this service is to promote ‘home banking’ by Apollo Trust customers. Customers can check account balances, transfer money between accounts, review bankstatements and receive information about Apollo Trust and its services. As the local telephone companyhas an unmeasured rate option for local telephone calls, customers can make use of their full hour per daywithout incurring additional PSTN charges. The bank has a 56 kbit/s leased line to an Internet backbonenetwork for the service but they are currently looking to upgrade this to a T1 line. To date it is reportedthat the costs of the service have been met by the additional loans made to Internet customers. Initiallyapproved as a one year trial the Bank now plans to continue the service. In 1996 the charging structurewill be revamped, although the free hour per day will remain a permanent feature for bank customers. The primary reason for the price restructuring is to allow users to stay on-line for more than one hour per dayfor a fee.
Information Service providers such as America Online (AOL) are also moving to introduce bankingservices for their subscribers.xxxviii AOL and Intuit, a provider of financial software have announced thatAOL subscribers will be able to carry banking services from the first half of 1996. The electronic bankingservice is planned to be incorporated into a normal AOL account such that subscribers can access theirfinancial accounts in participating institutions and undertake tasks such as checking statements, balances, transferring funds and paying bills. AOL has stated that they will not have an additional surcharge on topof the normal monthly fee for the service and that the pricing of access will be at the discretion of each participating financial institution.
In Europe two new ventures were established in 1995, AOL-Bertelsmann Online and Europe Online, but they were yet to commence service at the time of writing. xl As with IAP facilities, customers of ISPs generally access their networks via the PSTN. In the past the main difference between the two was that IAPs provided access to the Internet whereas ISPs provided access only to proprietary databases. In 1995 this distinction started to break down with a number of ISPs introducing Internet access and most others planning to follow in 1996.
Since several ISPs have very large and rapidly growing subscriber bases, compared to most IAPs, their decision to provide paths from their networks to the Internet will dramatically increase access. For example, by September 1995, AOL and CompuServe each had more than 3.5 million subscribers.xli In the 3rd quarter in 1995, AOL added 800 000 additional subscribers (65 000 new subscribers per week toward the end of this period) while CompuServe added 320 000.xlii At the end of October 1995, AOL was already a leading provider of Internet access from the home and in early November 1995 passed the four million subscriber mark. According to AOL the company “...processes more than 13 million hits on theWeb, delivers e-mail to four million people—half of which is Internet mail, and AOL members post more than 70 000 messages to Newsgroups.”xliii Microsoft’s MSN service added 525 000 subscribers in the first three months of service.xliv
In the past ISP pricing was based on providing customers with access to proprietary information rather than plain access to the Internet. Accordingly the price of Internet access from ISPs is much higherthan IAPs because it includes access to proprietary databases. Table 21 demonstrates the difference between accessing the Internet via several ISPs in the UK and an average of four IAPs. The basic CompuServe rate is three times more expensive for 20 hours per month and five times more expensive for 30 hours per month than the IAP average.
Recognising the need to adapt pricing to compete with IAPs, and facing the launch of MSN, anumber of ISPs have been restructuring pricing. In February 1995, CompuServe reduced the price for connection charges 50 per cent to US$4.80 per hour while raising its monthly membership fee by US$1. xlv
In September 1995, CompuServe further reduced its on-line charge to US$2.95 per hour while retaining the same monthly rate of US$9.95 (Table 22). The option aimed at being competitive with IAPs is the ‘Super Value’ package which includes 20 hours per month of on-line time for US$24.95. This pricing option includes access to the Internet and CompuServe’s proprietary databases.
CompuServe’s US pricing is similar to the UK pricing except that customers pay a higher rate for additional hours under the ‘Super Value’ option. While CompuServe charges are generally uniform the reason CompuServe gives for charging more to customers in some parts of the world than the US is the high cost of leased lines from monopoly PTOs. xlvi The same principle applies to all IAPs and ISPs paying relatively high charges to provide European connectivity.
It could be said that CompuServe’s new prices are very competitive with IAPs, when access to proprietary databases is taken into account, although less so at higher levels of usage. AOL took adifferent approach in June 1995 when it purchased Global Network Navigator (GNN) so that its customerscould subscribe to AOL’s standard service with access to the Internet, or simply access to the Internet via GNN. xlvii In October 1995, AOL announced the prices for accessing the Internet via GNN, without access to AOL’s proprietary databases. The price for customers was US$14.95 for 20 hours of use at any time ofthe day and a further US$1.95 for each additional hour. The first month, including unlimited usage, isfree. AOLÆs standard charges for its proprietary service at this time were US$9.95 per month, includingfive hours of on-line time, with additional hours priced at US$2.95. An additional complication incomparing prices is that ISPs periodically offer customers free on-line time for undertaking specific tasks(such as filling out questionnaires) or by paying their accounts in a particular way. For example, inNovember 1995, Prodigy was offering 10 free hours on-line to customers paying with a particular creditcard rather than paper cheques.xlviii
In practice both CompuServe and AOL are pricing 20 hours of on-line access to the Internet atUS$14.95 with additional hours at US$1.95. This compares to NetcomÆs price for US users of US$19.95for 40 hours with additional peak hours (9 am to midnight) being US$2.00 in October 1995. In terms ofprice the best option simply depends on a userÆs needs. Someone who only wanted 20 hours per weekwould be better taking, for example, the GNN option. On the other hand for someone requiring on-linetime in excess of 22 hours it would be less expensive to sign up with a company such as Netcom. At thesame time ISPs and IAPs are trying to differentiate their service in other areas. The major advantageNetcom has to offer ædial-upÆ customers is access through more than 200 points of presence at speeds of28.8 kbit/s. This speed is not yet offered universally by ISPs although networks are being upgraded apace. In 1995 only IAPs offer Internet access at speeds higher than 28.8 kbit/s. AOL plans to offer ISDN accessto its services in 1996. xlix
Another way IAPs distinguish themselves is to offer browsers that have been developed in-house. Netcom offers a product called Netcruiser which is client/server interface to the Netcom networkand, in turn, the Internet. ISPs such as CompuServe and AOL are acquiring specialist Internet companiesand tools so they can incorporate Internet products with their services. In the 12 months to October 1995, AOL purchased eight companies designed to help strengthen their positioning in the Internet.l In March1995 CompuServe bought SPRY Inc, a maker of an Internet navigational software, for US$100 million.li
This acquisition enabled CompuServe to offer Spry’s ‘Mosaic’ software through the CompuServe network. In other developments related to Internet navigation during 1995, AOL acquired the WebCrawler, an Internet search tool, while Yahoo Corporation’s Internet Guide was incorporated into CompuServe’s ‘Mosaic in a BoxÆ. In time certain search tools or upgraded versions of these tools, nowfree to use, may only be available to customers subscribing to a particular ISP or IAP. Another modelwould be for these tools to be advertiser supported, with IAPs and ISPs selling space in the same waybroadcasters sell time over their networks.
The other major area in which ISPs and IAPs will increasingly distinguish themselves is in the area of content provision. While IAPs, in the main, initially provided plain access to the Internet withsome links from their home page to search tools they are increasingly providing content. The entry ofcompanies like AOL will probably increase this trend as they provide propriety information for theirInternet only customers. This is where it becomes more difficult to make straight comparisons betweencompanies such as GNN and Netcom. For example GNN provides subscribers with access to diverseinformation produced by, among others, celebrity travel writers, restaurant chefs, winery owners, and post modern authors. GNN will also be providing access to news and information from groups such as Reuters and ‘The Sports Network’. In other words even the AOL’s offer of Internet access via GNN without an opening to its proprietary databases, does not mean that value added content is not being built into the service.
The mixture of carriage and content to attract customers to Internet, and other on-line services, would appear to be the fundamental reason for a series of alliances and mergers. The investment made byMCI in News Corporation, prompting the merger of their on-line services, is a prime example. NewsCorporation purchased Delphi in 1993, while MCI has a long association with the Internet and says it isthe world’s largest carrier of Internet traffic with more than 40 per cent. lii Following the News Corporation/ MCI alliance, was the creation of an alliance between AT& T and CNN Interactive to develop a range of multimedia and on-line services.liii Indeed the Internet focus of the above companies is a prime example of the convergence of services with not only PTOs and media companies forming on-line joint ventures but IT companies, such as Microsoft through MSN and Apple Corporation via ‘eWorld’, establishing on-line services. In November 1995 Apple Corporation also acquired a 5.1 per cent stake in AOL.liv
The importance of content in the new communication environment rightly receives much of thefocus in the process of convergence which is bringing companies from previously diverse industriestogether. It is also important to realise that ISP and IAP services are significant, and increasing, platformsfor communication. Jack Davies the President of AOL has stated,
“The reality with America Online is more than 50 per cent of our usage has nothing to do with content. It is about chat, it is about e-mail, it’s about posting messages on message boards. It’s people communicating with other people. So anytime you hear somebody say content is king .... that is we believe a substantial myth and that the development of this medium is very much about community. It is about bringing content and community together in an interesting fashion and making it interactive and participatory rather than one way.”lv
As Internet based technologies evolve this role can be expected to increase. For example, inNovember 1995, a new service was announced by Mobilemedia and AOL which enables customers ofAOL to send text messages to people with MobileMedia pagers.lvi The new service, which has theappearance of an e-mail service to the AOL customer, is available at no additional charge to the normalmonthly fee. Like the æelectronic banking serviceÆ available to AOL subscribers, these services areevidence of an emerging trend for ISPs to provide value added communication services in a diverse rangeof areas. ISPs will try to add value to communities of interest that were once closed forums because theyrecognise that their customers want to be able to access information outside their proprietary worlds. Inother words, by integrating technologies like HTML, ISPs can build in communication links from interestgroups to related information for their subscribers.lvii
The other question that is raised by the entry of ISPs into the Internet access market is the impact their pricing may have on the development of IAP pricing throughout the OECD area. Compuserve’s new ’Super Value’ tariff structure, based on UK and US prices, could have a direct impact on IAP pricesbecause it is less expensive than IAP prices in many OECD countries. CompuServe’s new rates came intoeffect after the August snapshot of prices shown in Table 11 was taken. The new rates aimed at Internetusers, introduced in September 1995, were less expensive than the IAP prices for 18 countries for 20 hours of on-line time per month shown in Table 12 . For 30 hours per month CompuServe’s rates are less expensive than the IAPs shown in Table 12 for 15 countries. CompuServe rates in the US are even morecompetitive because additional hours are less expensive. As they restructure their prices to compete with CompuServe other ISPs can be expected to set prices at or below these levels. Clearly for countries in which CompuServe has a local point of presence, allowing customers to call at local rates, their service will be very competitive. While AOL is less well established outside the US than CompuServe its pricesto US users, for 20 hours of on-line time, are less expensive than all the representative IAPs shown inTable 12 except for the UK for August 1995.
Most ISPs have not yet restructured their tariffs along the lines of CompuServe and AOL for thenew environment but changes are expected as they introduce access from proprietary networks to theInternet. CompuServe provides on-line services to customers in more than 140 countries but unless there isa local point of presence allowing direct connections via local PSTN charges users need to paycommunication surcharges. In October 1995, CompuServe had a local access points in all the Capitals ofOECD countries and a large number of provincial centres, although penetration of points of presencevaries from one country to another. In Germany and Japan, CompuServe has established local numbers inmore than 60 cities while for several OECD countries only the Capitals have local points of presence. InSpain, for example the company has local points of presence only in Madrid and Barcelona, in Portugalonly in Lisbon and Porto, and in Turkey, Ankara and Istanbul. Outside these areas, if there is not anational number CompuServe customers can reach without communication surcharges, the additionalcharges users pay mean the service would be significantly more expensive than an IAP with a local pointof presence. For example users in New Zealand outside Auckland, Christchurch and Wellington wouldneed to pay US$10 per hour prime time and US$6 per hour non prime time to access CompuServe. Nevertheless for the vast majority of users, Internet access is increasingly becoming global business with IAP/ISPs competing around the world irrespective of national borders.
Table 13 . Selected On-line Information Service Providers
Table 14 . Internet Access via On-line Information Service Providers in the UK, US$PPP
Table 15. Compuserve Pricing with Internet in the US
In general users in competitive markets have benefited more from the rebalancing process thanthose in markets without competition. For example while the total cost of a basket of business servicescontinues to fall in competitive markets they have recently risen, on average, in those markets withoutcompetition (Figure 5 ). The main difference is that while fixed charges have risen in both monopoly andcompetitive markets between 1990 and 1995, the rise in competitive markets has been more than offset by falls in usage charges. On the other hand, in countries without infrastructure competition there has been a slight trend in the opposite direction with a recent rise in average usage charges.
There has also been a process of rebalancing underway within usage charges. In general the price of national calls over longer distances, and international calls, has been falling while the price of local calls has been rising. Figure 6 shows the price trend for the cost of calls over different distances incompetitive and non-competitive markets. The price reductions in markets with telecommunicationinfrastructure competition over longer distances are the most impressive. This is the market segment inwhich competition has been most fierce to date. Yet the average price of local calls has also been falling, due to price regulation and the introduction of local competition, in countries such as the UK. By way ofcontrast the trend in non-competitive markets between 1990 and 1995, has been for continuing rises inlocal call prices. Moreover the modest falls in long distance prices in non-competitive markets wereeroded by increases between 1994 and 1995. To the extent local call charges are already a barrier to usersaccessing information infrastructure rises in local call prices in monopoly markets will exacerbate theproblem.
In respect to rebalancing, PTOs say the new charges better reflect their costs of providingtraditional telecommunication services. Nevertheless the rebalancing process has tended to becontroversial because of the difficulty in precisely assigning cost causation to different parts of thenetwork. Much of the costs involved in building and maintaining telecommunication networks are jointand common costs. Accordingly it can sometimes be an arbitrary decision as to which service hasgenerated what share of the total cost. Moreover PTOs in a monopoly environment have generally notcollected the information necessary to accurately assign costs. When the information is available separatestudies can, and often do, come up with quite different results depending on their methodology. One example has been the endeavours in a number of countries to determine the cost of universal service where different methodologies have produced widely differing outcomes.
It is also true that the largest price falls in the rebalancing process have been in the mostcompetitive market segments (long distance and international calls) and the largest rises have taken placein those market segments where competition is prohibited by regulation or is immature (local access andshort distance calls). This is one of the main reasons telecommunication regulators have monitored tariffrebalancing to safeguard against PTOs unfairly shifting the burden of cost, or not passing the benefits ofgreater efficiency, to those market segments where there is not effective competition.
To date tariff rebalancing would have been academic if all users made use of telecommunication networks in the same way. In reality, of course, telecommunication users do not make equal use of thenetwork. The reason that the cost of telecommunication has fallen faster for business users, thanresidential users, is their different use of various services. In general business users make greater use ofmore expensive long distance calls than residential users and consequently they have benefited more fromthe price reductions to long distance calls. In the new telecommunication environment a new division ofusers may occur, not so much between business and residential users, but users of on-line services (such asthe Internet) and others continuing to only utilise traditional telecommunication services.
In future policy makers will need to be aware of the impact tariff restructuring may have on thesetwo groups of users. Recent price restructuring introduced by France Telecom provides a good example ofthe impact tariff restructuring may have on different types of users.lviii Prior to the change FranceTelecom’s customers were charged for calls within their local area in units of six minutes. In 1994, thesize of most local calling areas was enlarged but the unit of call measurement was cut to three minutes. Inother words subscribers could dial, at local rates, to a larger calling area but for calls beyond three minutesthe charge increased (i.e. the price of a six minute call doubled to a location within the previous callingarea but was much less expensive in the extended area). According to France Telecom almost threequarters of its local calls were unaffected by the change because they last less than three minutes. At thesame time France Telecom brought down the cost of calls over longer distances by expanding the time inwhich units are measured from 17 second to 19 seconds.
For users of traditional telecommunication services, based on ‘telephony patterns’ of use, FranceTelecom calculated the overall restructuring represented a 2.4 per cent reduction in prices to customers.lix
In addition under the new scheme customers increased, by an average of seven times, the number ofpeople they could call at local rates. On the other hand for ‘on-line’ users the changes represented a mixed blessing. While on-line users had an extended local calling area that might take in additional points ofpresence from IAPs, or other on-line service providers, the change doubled the cost of 30 hours of PSTN usage per month at standard rates from US$28 to US$56. Over the course of a year, while potentiallyhaving little impact on a telephony user, this would increase the bill for PSTN timed charges for an on-line user from US$339 to US$678.
The fundamental principle forwarded by France Telecom for these changes to the tariff structure was that because of technological change, distance counts less and less in the cost of providing telephony while the length of calls is becoming a key determining cost. lx If this principle is correct France Telecom’s new tariff structure better reflects underlying costs. Indeed, under the former system users of local calls atless than three minutes would appear to have been subsidising those making calls of longer duration. Onthe other hand the new price structure adds considerably to the cost of an on- line basket of services by making calls of a longer duration more expensive and may make such services less affordable. Yet any new policy aimed at reducing the cost of such a basket, along the lines of flat or unmeasured rate pricing would run counter to the principle behind the initial rebalancing. Such is the dilemma for PTOs in the new communication environment.
It is also true that France Telecom’s changes, announced in 1993, may reflect the thinking ofPTOs in the pre- competitive world of local telecommunication. For example, with increasing local competition in sight, NTT has introduced a flat rate tariff for off-peak hours which the company says is not based on distance or duration. In its “1995 Annual Report” NTT noted that affordable rate structures, that reflect diversified needs, will be increasingly important in the new telecommunication environment. Accordingly from August 1995, NTT introduced “Telehodai” a PSTN based service which allows users toplace an unlimited number of calls to specified numbers during late night and early morning hours for aflat monthly rate.lxi NTT has also provided a similar ISDN based service. NTT’s new pricing direction is contrary to the principle espoused by France Telecom but is clearly a response to an increasing need ofsome users to access information infrastructure at more affordable prices and for the company to have apricing structure which enables new services to be developed.
While the issue of tariff rebalancing has been relatively controversial in the past this can be expected to increase in the future because of its impact on the use of information infrastructure for thosecountries heading in the France Telecom rather than the NTT direction. In the telecommunication tradepress the issue of raising local call charges is just beginning to receive attention in respect to on-lineservices. For example, an article entitled “On-line Customers Hit by Telekom Price Hike” in EuropeanTelecommunications noted that while Deutsche Telekom was bringing down the cost of long distance callsits proposed tariff schedule will increase the cost of local calls by as much as 156 per cent. lxii The articlequoted a user stating the belief that the new tariff structure would slow the development of informationinfrastructure. Barry Berkov, a CompuServe executive has stated:
“Even if we have a CompuServe node in your local city and even if it is a local call, your localtelecom costs are likely to be more than the cost of accessing CompuServe and payingCompuServe’s charges itself. I think that’s a situation that definitely needs to change if we aregoing to see the explosion that we would like to see in the use of on-line services.” lxiii
How the development of on-line and Internet services over the PSTN will develop is difficult toforesee. If consumption of ISP services is any guide it suggests consumers do not like to pay surcharges, and if they view communication carriage as an additional cost in the manner of a surcharge, they may bereluctant to pay. It has been suggested that CompuServe backed away from its initial pricing structurewith surcharges because AOL was able to surpass its subscriber base in a relatively short time withoutsurcharges. In other words PTOs that want to sell value added information and entertainment basedservices to their customers might have to re-evaluate the underlying pricing of carriage to take this factorinto account. AOL’s Jack Davies says their experience has been that:
“What consumers tell us is pricing has to be simple and affordable. Our services are charged at US$9.95 per month, five hours free - no surcharges. What consumers tell us is that they don’t like surcharges, they will not pay for services with surcharges .... they are going to want itpackaged all together in one simple affordable price.”lxiv
However this is in North American markets where unmeasured local rates are widely available. For manyother OECD countries, where unmeasured rates, flat rates or relatively inexpensive local calls areunavailable even this level of on-line use may be unaffordable. In Japan, on average, people spend morethan 4 hours per day on the following activities—television, radio, videos, CDs, newspapers andmagazines.lxvi Hence, as NTT recognises, if it would like to provide the carriage for telecommunicationand information service access in all its forms over the PSTN, new charging practices will be essential.
This raises an interesting question in respect to the pricing of different types of multi-mediaservices. Currently users generally receive television and radio in real time whereas other forms of mediacontent in their traditional guises are received in pre-packaged form such as CDs, rented video tapes, andnewspapers. For those users that continue to need to browse or receive content on-line in real time, localcharges when expensive will continue to be an obstacle to use. On the other hand existing chargingpractices might be less of a hurdle if content could be downloaded in a relatively short time. In otherwords a PTO offering a video on demand service might ask the customer to pay for the contentdownloaded (e.g. a movie) with existing local call charges if it could be quickly downloaded for replay without on-line charges. This would be a step away from usual PTOs charging practices for high speed transmission by volume. From the customer’s perspective the main question would be how long it would take to download a product?
At present cable communication companies, because their inherited systems have coaxial cable providing the link to residential premises, are ahead in terms of the speed with which signals can bedelivered over what PTOs can provide over the PSTN using copper wires. For example, according to arecent report on interactive television by Frost and Sullivan, cable communication companies haveupgraded the number of US households capable of receiving interactive (two-way) capabilities from 1.5 million in 1993 to 10 million in 1995. lxvii By way of contrast Frost and Sullivan say US PTOs will equip 1.6 million homes with ‘video dial-tone’ services by the end of 1996 and that this will reach 5.5 million by 2001.
The problem at present is that because regulatory environments have in the past not allowed cable communication companies to offer telecommunication services, in most OECD countries, there has been no market for the development of cable modems and set-top boxes. Where regulatory restrictions have been lifted cable communication companies are racing to bring cable modems to the market. TCI and its partners say that their ‘@home’ service will enable “...a high-volume cable connection that pushes information at 10 megabits a second through cable wiring - directly to your computer. A movie that currently takes 27 minutes to download will take only around 9 seconds to display.”lxviii
It is planed for ‘@home’ to be launched in Sunnyvale California in early 1996 and TCI and itspartners aim to have one million customers in the US by the end of that year. It is hoped that widespreaddemand will bring down the current cost of a cable modem from US$500 to US$300 or US$200. lxix
Hewlett-Packard has announced its cable modem product “HP QuickBurst” is scheduled for launch in mid- 1996.lxx The company says ‘QuickBurst’ will offer the same access to networks that telephone modems do only at speeds ‘thousands of times faster’ but has not yet announced pricing.
Where regulatory regimes allow PTOs to operate cable systems they are also experimenting with cable modems (e.g. Bell South with Intel modems in a Georgia trial). This raises the question of whether an alternative infrastructure will be developed by PTOs in parallel with the PSTN for final delivery into small business and homes. Under this scenario the transport of PSTN traffic for the final connection to thehome may be gradually transferred to the new broadband links (hybrid fibre/co-axial cable) rather than the progressive introduction of new services over the PSTN. This might allow PTOs to side-step the difficult issue of how to integrate the pricing of new and existing services because the last link would be delivered on separate infrastructure. This network model would allow PTOs to charge for different services in thesame way cable communication companies charge different rates for cable television and telephony in theUK.
By introducing cable television in 1995 Australia has started much later than many other OECDcountries. In practice this has meant Telstra and Optus, the two carriers licensed to provide PSTN network infrastructure, have been free to select any technology to deliver new services. Both Telstra and Optus are actively using coaxial cable as part of their broadband network expansion. In the relatively short timework has been underway Telstra has passed 600 000 homes installing more than 7 000 kilometres ofcoaxial cable.lxxi In late 1995 Telstra was passing up to 5 000 homes per day with coaxial cable. Use ofcoaxial cable by both Telstra and Optus suggests they believe that the last part of delivery of some new services by existing copper wires, is still some way off as a technology of first choice.
The view that hybrid fibre/coaxial cable networks will be the most prevalent network architecture, in the short to medium term, is also espoused by Frost and Sullivan. lxxii Still, either becauseregulation prohibits PTOs offering cable television service or because they wish to more efficiently utiliseexisting infrastructure, many PTOs are conducting trials of new services, such as video on demand, overexisting infrastructure. If fixed link transmission of video services, that has in the past required coaxialcable, can be squeezed down the existing copper wires that still provide the last part of thetelecommunication connections to small businesses and virtually all homes, it might significantly reducethe future cable investment requirements of PTOs. In October 1995, hopes in this direction were boostedby AT& T Microelectronics announcing the development of a set of integrated circuits able to transmitswitched digital video over ‘ordinary phone wire’ and disclosing that a number of other companies, suchas France Telecom and Southwestern Bell, planned future deployments using the technology. lxxiii Thetechnology enables the transmission of up to 51.84 MBit/s from 100 to 300 metres over ordinaryunshielded twisted pair wiring. Moreover the encoded video is transmitted at a different frequency to the voice band used for telephony allowing simultaneous voice and interactive video to travel over the samecopper wiring.
The major benefit of such a technology is that it would enable much of the existing PSTN, the part providing the final connection to premises, to be used for new services. Fibre optic cable would be used for transmission up until the final 100 to 300 metres. However wide deployment of technologies capable of transmission of 51 Mbit/s and the fibre optic cable necessary to transport signals over longer distances to all parts of public networks are still in the future. In November 1995, Bell Atlantic announced a commitment to use AT& T Paradyne’s GlobeSpan technology. Bell Atlantic and AT& T Paradyne say this technology will enable services, such as video-on-demand and Internet access, at transmission rates of up to 6 Mbit/s over existing copper wires. At this speed the GlobeSpan technology would provide speed improvements that are more than 200 times faster than those supported by a 28 Kbit/s ‘dial-up’ modem.
In future what will become doubly difficult for PTOs with measured rates, and who say costs aredriven by duration, is to price new services differently over the same copper wire infrastructure. If aregulated or monopoly PTO, with measured pricing, lowered prices for particular telephone numbers ornew services, (e.g. video on demand), by introducing flat rates or unmeasured rates for customers, theywill be in the position of having to explain why such pricing options are not extended to users oftraditional telephony services. Since the new services may be causing additional investment in networks, that would not be required for traditional services, it will become increasingly difficult to explain priceincreases for telephone service at a time when discounts are being offered for new services. At the sametime whatever pricing schemes PTOs introduce for the PSTN they will have to be available to all servicesuppliers on an equal access basis or PTOs would have an anti-competitive advantage.
Both companies posted increased revenues and profits based on this development. Indeed in November1995 several Ameritech subsidiaries were offering ‘second line sales’ to attract customers with discounted installation.
If the increase in demand for second residential lines is due to users wanting additional access toon-line services it is a promising area of new growth for PTOs. In this context it will be interesting tofollow whether demand for second residential lines occurs faster in those countries with uncharged localcall rates, or low rates, which enable users to stay on line for longer amounts of time and thereby generatedemand for additional household lines. On the other hand raising fixed charges may discourage a majornew area of growth for PTOs and may also act to slow demand for information infrastructure services.
Given the new pressures on rebalancing how might policy makers proceed. Fortunately severalvery robust policies are emerging to deal with these problems particularly at the local level. First andforemost competition has been a major tool in assisting with the rebalancing process because it hasencouraged operators to become more efficient in meeting customer demand. Tariff rebalancing should not be a process of shifting prices to cover inefficient provision of service. A reduction in costs brought about by greater efficiency can reduce the necessity to raise prices, particularly in those cases where thereis a financial loss to the PTO due to a commitment to universal service. Indeed the cost of meeting universal service obligations should fall as PTOs become more efficient. Price rebalancing in competitive markets between 1990 and 1995 has been undertaken in a way that held average prices to residential users constant while passing benefits to business users who make greater use of long distance and international telecommunication. Prices should, of course, be cost orientated but PTOs need to ensure that those costs are as close as possible to best practice level.
When rebalancing has occurred in monopoly markets, it raises the question of whether the process is simply shifting the burden of cost between different services and categories of user. Those countries that delay the introduction of competition to ‘assist in the rebalancing process’ may actually be providing incentives for PTOs to rebalance in directions that will ultimately prove unworkable in the new telecommunication environment. While the real key to meeting the challenge posed by new network patterns of use for rebalancing prices is to increase PTO efficiency, it is also true that competition can assist in this process by encouraging technological and service pricing innovation.
In the case of access to the PSTN, opening local markets encourages technological innovation in both the conventional PSTN and alternative infrastructure. Developments in the UK are a prime exampleof this trend with companies striving to improve access via the PSTN, cable communication and wirelessnetworks. Moreover as local competition develops in Australia, Canada, Finland, New Zealand, Japan, Sweden and the US, the potential market for local competition opens up from 27 milliontelecommunication mainlines in the UK to more than 270 million mainlines in those OECD countries permitting infrastructure competition by 1995. Whereas cable communication operators and manufacturers previously had no incentive to develop cable modems, a very large new market is beingopened up by liberalisation. More significantly for the first time there is real market pressure on PTOs, and in turn their equipment suppliers, to lower the price of access to networks and reduce the pressure for rebalancing in that direction.
Even if the rebalancing process was halted in many countries new ways of pricing networks andservices are going to be needed. This is the other major contribution a competitive market can make. Indeed the first signs of pricing innovation for on-line access to information infrastructure are alreadyevident in competitive markets. The cable communication company Videotron’s offer of uncharged local calls in off-peak periods, between its customers and IAPs, is a leading example of such a tariff innovation. Recognising there is a growing market for the installation of second residential lines Videotron also offersdiscounted connection prices. Similarly Diamond Cable, a UK cable communications company offers a 50 per cent reduction on the rental of second lines. It would be expected that innovative pricing by cable communication companies will be further boosted when cable modems become widely available.
Competition for the provision of service at the local level need not only come from cable communication companies. Fixed wireless competition for the provision of first and second lines is expected to emerge first in the UK during 1996. After several years of testing, Ionica has announced that it will commence its fixed wireless service in some areas of the UK in March 1996. lxxvi Nation- widecoverage using the Northern Telecom supplied equipment is expected within five years. The Ionica system uses digital radio signals to connect homes and businesses to local exchanges. An interesting feature of this technology is that it can support one or two ‘lines’ to a local exchange. In other words the customer can elect to subscribe with one or two lines and change their service configuration at any time without a visit by Ionica staff. The expected benefit of this service is that the reduced cost of infrastructure provision will allow lower prices to consumers.
Introducing wireless competition is also being examined in other parts of the OECD. In MexicoTelmex has a monopoly over long distance and international PSTN services until the 1st January 1997. However in theory the local telecommunication market in Mexico has been open for several years. Inpractice, local competition has not developed because any new market entrant could only offer local service and would have to interconnect their network to Telmex for long distance and international services. Telmex has had a monopoly over long distance and international calls. From 1997 Telmex is required to interconnect its network and any market entrant is permitted to offer a full range of PSTN services. To this end Iusacell, a mobile cellular company, has been conducting a trial with wireless technology in order to provide an alternative to those areas without service, a long waiting list or with a poor service. Iusacell’s tariffs for the trial are higher than Telmex’s although less expensive than local service in many other OECD countries (Table 23 ).
In a country with a relatively low telephone penetration rate opening markets to new suppliers will have major benefits in terms of expanding service. The FCC’s Susan Ness has stated, “We believe that competition can, in fact, provide faster and better opportunities to build out an infrastructure particularly where there is a low tele-density. The competition improves incentives for innovative network solutions...”lxxvii In OECD countries with higher telephone penetration rates, such as Finland, local competition is being used to provide customers with choice and boost the incentives PTOs have forinnovation.
For their part PTOs will not sit still in the face of competition and this will impact on rebalancing. Indeed PTOs have shown, when they are faced with competitive markets, a remarkable ability to offer innovative pricing. Developments in mobile telecommunication pricing in liberal markets provide one example where low user schemes have made networks increasingly affordable for a wider range of users.lxxviii Faced with increasing demand from users with a choice of supplier PTOs will adapt pricing to the needs of their customers. They will also strive to bring new technologies to the market, often ones which have been developed by cable or wireless competitors, and improve the efficiency of their own service. In short, the new telecommunication environment may produce surprising directions in tariff rebalancing in response to competition and the need to price networks differently in the recognition that not doing so would limit the growth of revenues from new services. Restricting competition to allow monopolies to rebalance in directions that may, or may not, be appropriate could significantly delay many of the potential advantages of information infrastructure.
Figure 4 Time series of business and residential baskets
Figure 5 Comparison between different markets(Business basket)
Figure 6 The impact of competition of usage charge rebalancing
Table 16 Telmex and Iusacell Telecommunication Charges, US$