APC - ICT Policy Handbook

User costs – “free” to the user but who pays?

The internet appears ‘free’ to the user. If you are using a dial-up account, once you have paid for a telephone connection to an ISP and the subscription costs of the account, browsing the internet or transferring emails is essentially free. But the internet has never been free in the monetary sense. In the beginning, the US government paid for the infrastructure, first via its military research programme (AR PA), then through the universities (NFSNet, etc). Now, internet is big business, and someone has to pay for everything. The ordinary user pays the ISP to be connected. The ISP may charge a monthly flat rate, or according to the amount of data that the customer receives/sends, or the number of hours connected. The ISP in turn has to pay the telecommunications company or another ISP for its connection to the internet, and perhaps for the rental of computers or other services, apart from its normal running costs. The customer also has to pay for the telephone line, with a monthly rental and a fee for each (normally local) connection. Many telephone companies charge by the minute for local calls so the longer the user is connected the more expensive the connection. In fact, even when some ISPs offer ‘free’ connections to the internet, they usually offer the connection for free, but by arrangement with the telephone company they earn money from the telephone charges for the call.

When accessing a web page or sending an email, the packets are routed through the various interconnected networks of the internet to reach their destination, over links which are paid for by the ISPs. Take the case of browsing a website in Fiji from the UK. The user establishes a connection to the internet by dialling up to an ISP’s nearest point of presence (POP). From this point on, the user does not pay anything else. The ISP provides a leased line from that POP to its central node, and then to other ISPs where it exchanges traffic with that network, which in turn does so with another network, and so on. The packet is routed through the network of networks until it reaches its destination. The ISP pays the telecommunications company or a larger ISP for its connection, and they in turn pay a larger carrier. The larger companies enter a variety of commercial agreements among themselves, to share the huge number of cables, routers, and computers that make up the global internet infrastructure. In this way, the distance to the website or POP server is not important to the end user, because it does not influence the cost of the communication. We pay for unlimited access to any point on the global network.

The Economic Toolkit for African Policymakers¹, published by the World Bank in 1998, splits the cost of internet access for the user into its component parts. It found that for 30 hours of internet access in Africa, 15% of the cost is accounted for by telephone access, 42% by equipment (37% on computer and 5% on modem), and 43% by ISP subscription. The main upfront cost is computer equipment, which is subject to import tariffs and other taxes in some countries. Equipment costs can be reduced massively by public internet access centres, which lower entry barriers be removing these costs. This is especially so in developing countries where there is a limited number of computers, which naturally therefore limits the number of potential internet users.

Telephone access costs

The cost of dialing up to the ISPs POP using a telephone line depends on two things: first, the amount of time spent online and second, the tariffs which are charged by the telephone operator. The amount of time which is a user needs to spent online depends heavily on the bandwidth of the line, which is measured in the number of kilobits per second, as this determines how long it takes to download a file. Telephone lines are designed to carry voice conversations at 64 Kbps, and modems work at either 56 Kbps or 64 Kbps. In practice however, speeds are usually much lower than this because of congestion, attenuation of signals, or poor quality analogue lines and switches. Faster lines using ISDN (integrated services digital network) or DSL (digital subscriber line) technologies can be obtained where telephone companies have installed them, or wireless technologies.

The availability of different Internet access lines

The availability of different Internet access lines is one point for policy interventions. In the UK for example, action groups have been established in many rural areas to lobby British Telecom to install ADSL in their local exchanges. This pressure can work. For BT it encourages would-be users to register their interest in having ADSL, and as this develops the critical mass of users required to justify the cost of upgrading the exchange can have the effect of encouraging the company to install the equipment.

For heavy internet users, it frequently becomes more cost efficient to lease a dedicated line from the telephone company and pay a flat price regardless of how much time is spent on the line. Leased lines can be bought at 64 Kbps, 128 Kbps, 256 Kbps, 512 Kbps, 1024 Kbps, 2048 Kbps or above. Organisations or businesses with many computers will connect them together onto an internal local area network (LAN), and then share one ‘leased line’. At the point where the costs of dialing-up exceed monthly outlay on a leased line (for example if there are 10 computers which are used for internet access), it becomes more cost effective to lease a line. Where ‘broadband’ has been deployed (DSL), this is also charged at a flat fee in the same way that leased lines are.

The tariffs charged by the telephone company for lines and leased lines are the key cost factor for internet access, therefore another area of concern for policy makers. These tariffs are set by the telephone company, but in many countries are subject to approval by the regulator. A key lobbying point is for the telephone company to use local rates for dial-up tariffs.

A telecommunications monopoly in a country means that users have no choice of a telephone company to use in order to dial up to their POP, and must accept the tariffs charged. In a competitive environment, users can choose between a number of local exchange carriers, are free to choose the lowest tariff and in so doing lower the cost of internet access. A monopoly over international traffic means that ISPs are not able to establish their own independent links to foreign ISPs and to the internet backbone, and must route all traffic and bandwidth through the telephone company, which can charge whatever price it likes for the international bandwidth.

ISP Subscription costs – what the user pays

All the other costs of the internet are hidden from the user. The ISP bears the rest of the upstream costs of the internet, but as a business it must pass these costs back

Tariffs can prove more important than the speed of connection for determining access prices. The figure of 15% of the total costs of access, which is spent on telephone calls in the African Toolkit model, is based on local rates. Where users have to make a long-distance phone call to the nearest POP, these costs skyrocket and account on average for 72% of the costs (the rest evenly split between the cost of ISP subscription and the equipment). In the 1990s, before ISPs were established in many countries, to get access to the internet users had to make an international call to dial into the POP of a foreign ISP – which was astronomically expensive.

to the end user through the account subscription fees. The ISP subscription fees are usually the highest component of the cost of internet access. Ultimately, unless internet access is subsidised by an organisation (such as a university or a company), the user pays all the costs. ISP subscription fees vary widely, and a number of models have been employed (for example the ‘free ISP’ model).
Using these two variables – telephone charges and the ISP subscription – the cost of internet access varies widely across the world. Most noticeably, the internet access costs are very much lower in OECD countries than in developing countries. In Asia for example the cost of 30 hours internet access per month in New Zealand is US$11.74 (US$11.74 ISP charges and free local call charges), which is less than a tenth of the cost in Kiribati US$142.8 (US$140 ISP charges plus US$2.8 telephone charges).

ISP Costs – what the provider pays

ISPs, which provide the ‘on-ramp’ onto the internet, are run either as businesses or in the education or government sector. Either way, the ISP bears all the upstream costs of the internet. In the case of commercial ISPs, these costs are passed back to the customer through subscription charges, plus profit. In the public sector, they are run on a not-for-profit basis and may to some extent be subsidised.

Internal network infrastructure: Because of the historic growth of the Internet in US and OECD countries, leased lines are concentrated in the developed core of the global economy. ‘At the end of 2001, there were around 375,000 permanent leased line connections to the Internet around the world,’ according to Netcraft’s leased line survey quoted by OECD^³. ‘Some 89% of leased line connections to the Internet are in OECD countries. The largest number of leased line connections is in the United States with one third of all connections in the world. Japan makes up 12% followed by the United Kingdom (7.1%), Germany (6.6%) and Canada (3.3%’)

.External network infrastructure: The ISP in turn must establish a leased line from a telecommunications company in order to connect to the internet. Again, the cost of leased lines varies greatly between countries, depending on the cost of providing the line and the degree of competition which exists. For this reason, it is economically attractive for ISPs in each tier to connect with each other through formal or informal peering arrangements. Because communication is cheaper locally, and national long-distance is cheaper than international traffic, Tier 3 ISPs would avoid, if possible, sending traffic to a neighbouring ISP through an international link.

Even within OECD countries, ‘local leased line prices remain of concern where there is currently insufficient competition. For users, in those markets, this means that incumbents can continue to charge prices that are not disciplined by competition.^⁴.’

In developed countries such as the USA, upstream service providers are located in the same country as Tier 2 and Tier 3 ISPs, and the telecommunication costs for the latter are lower – they only have to pay local or at most regional leased line rates (plus the transit charge) to gain access to the internet backbone. Furthermore, the generally higher level of competition in developed countries has resulted in lower prices for leased lines.

In many developing countries, where the internet has been introduced relatively late, ISPs are small and there may only be a handful of them. ISPs must therefore establish an international private leased line to exchange traffic with a foreign Tier 1 internet backbone provider (IBP) in order to gain connectivity to the internet backbone, so that users would be able to access a website hosted on another network on the other side of the world. These Tier 1 IBPs are usually based in the USA or Europe. Like local leased lines, this international private leased line circuit (IPLC) can be provided by the national telephone company or another data communications operator, or can be gained independently by the ISP establishing its own satellite connection where they are licensed to do so. A local 64 Kbps leased line in the USA cost US$80 per month and twice that (US$190) in Kenya, whilst an IPLC from Kenya was charged at US$1,687 (some 21 times as expensive).

Cost of internet access in developed countries (eg Germany)

Sources: Broadband Access for Business, OECD (2002);
Towards a Knowledge Based Economy, UNECE (2002).

Cost of internet access in developing countries (eg Ghana)