The purpose of the Toolkit is to enable decision makers in African countries to assess the economic impacts to be expected from the new cable and satellite projects. The intention is that the Toolkit will be used by participants at the Conference and subsequently to analyse the consequences for their own country of an individual project. Its use for this purpose will be demonstrated in workshop sessions at the Conference.

In this introduction to the Toolkit two case studies are presented to illustrate the possible effect of new telecommunications infrastructure. The first case study is for a coastal African country with a large population; the second is for a landlocked country with a relatively small, dispersed population. It is assumed that the coastal country has made significant progress in sector reform, attracting a strategic partner for the national operator and licensing a second fixed network operator. It also has two cellular radio service providers. In the landlocked country, the sector remains state-owned and monopolised.

These two country cases have been selected to illustrate significant differences in ease and costs of access to cable and satellite services. They are intended to be representative, rather than exemplary. In each case study, the impact of a GMPCS system and of access to a new submarine cable scheme are assessed.

To facilitate comparison between the two countries, as many factors as possible are held constant between them. So, for example, tariffs for telephone services charged by the national operators in the two companies are the same, even though it might be expected that sector reform would lead to a more rapid adjustment of tariffs towards costs in the coastal country.

The centre piece of the Toolkit is a model of the telecommunications sector, showing quantitatively the interaction of economic variables such as tariffs, traffic levels, revenues, costs and profitability. The impact of a cable or satellite project is expressed primarily through its effect on demand and on the costs of provision of service. The model has a modular format, enabling each element to be adapted to fit the relevant project for each country. The degree of detail in the model is strictly limited to the purpose for which it has been produced – to provide a guide for strategic decisions.

A separate explanatory guide to the model will be provided at the Conference.

The model has been prepared in US dollars, being the most convenient currency in which to compute the revenues and costs of international projects.

The advent of the GMPCS service provides a major extension of coverage of mobile services, effectively covering all parts of Africa. This is sufficiently valuable to ensure take-up of the service, but the number of users remains small (by 1998, 5,000 in the coastal country and 1,000 in the landlocked country). GMPCS is expected to achieve a share of the mobile market in both countries of about 2%. The direct economic benefit for the coastal country is about $4 million a year, and in the landlocked country less than $1 million.

Although neither country acts as a regional service provider, GMPCS is commercially significant for the leading cellular mobile operator, who is able to market satellite connections jointly with its own service. GMPCS therefore boosts the total number of cellular subscribers, as well as generating more calls.

For the national operator, GMPCS produces only a small gain in revenues, arising mainly from interconnection charges for calls to and from users in the country. The universal coverage offered by GMPCS is used to improve network access in remote areas, primarily by payphones. But the size of the satellite payphone network is limited by the relatively high call charges that have to be charged (or, alternatively, if calls by satellite payphones are priced the same as normal payphones, by the scale of continuing subsidy required).

The direct economic benefits from access to the submarine cable are substantially greater than from GMPCS. Almost all the benefit flows to the national operator in the form of lower costs of transmitting international calls. These benefits are likely to be $7-8 million a year in the coastal country in the first years of operation of the cable scheme. The cost savings may well diminish in later years as the costs of using other technologies, such as new satellite systems, falls over time.

Users benefit as charges for international calls and for international circuits fall. The rate of reduction in these charges is affected by tariff regulation; the Governments of both countries prefer that international call charges are reduced gradually. So, for a period charges are high relative to costs and the cross-subsidy of domestic services is preserved.

For the national operator in the coastal country, the cost savings are sufficient for its investment of $30 million in a cable landing station to pay off, yielding an internal rate of return of over 18%.

For the landlocked country, the cost of connecting to the landing station on the coast reduces the net benefits of using the cable, which are estimated at about $2 million a year. Even with the additional link, cable remains cheaper for all destinations except other neighbouring countries.

Neither GMPCS nor the submarine cable pose a significant competitive threat. Rather, they serve in practice to reinforce the market position of the leading operator in the mobile and fixed markets respectively. The commercial strategy of the GMPCS operators is to associate with leading land mobile operators in order to gain rapid access to users. They will offer dual mode handsets that will route calls via the land mobile networks where available. This should give the established mobile operator in each country a significant marketing edge. For cable schemes, only national operators can commit the large volumes of traffic necessary to make them viable. Other operators obtain access to the cable, though probably not on the same terms. So while the cable scheme reduces costs for all operators, the national operator will benefit more than others.

The case studies suggest, however, that these competitive effects are relatively modest and may not justify regulatory intervention. Regulation may be required to ensure that other operators are able to secure non-discriminatory access to the cable and satellite schemes. It would, for example, be open to a regulatory authority to require a GMPCS operator to co-operate with all land mobile service providers. However, such a policy might deter the leading mobile operators from promoting satellite services by removing much of the commercial incentive to co-operate. Moreover, probably more than one GMPCS service will launch, giving at least the second mobile operator in each market a satellite option. It seems likely that regulatory authorities will wait to see what impact satellite services will have before deciding whether specific non-discrimination rules are required.

In this section, a summary is given of the assumptions made in the case studies about the cable and GMPCS projects. These assumptions are common to both cases studies, and are derived from the same data as has been used in the preparation of the Briefing Document. In their own applications of the Toolkit, users can vary any of these assumptions as desired.

The GMPCS operators will offer from 1999 such services as global mobile telephony, fixed telephony, paging and data communications. The mobile satellite services considered in this case study is focussed on providing mobile telephony services, with fixed services to payphones as a second line of business in areas where the mobile service does not absorb available capacity. The first generation of GMPCS operators are not designed to provide substantial international transmission capacity for fixed operators. The second generation of GMPCS systems, starting service in 2001 or later, may well be able to offer bulk transmission capacity and be able to carry substantial quantities of data traffic, but it is too soon to predict volumes of such traffic or the prices at will it be carried.

The GMPCS operators plan to secure a relatively small share of the mobile telephony market in each country, and to build a large business by offering service in many countries. Their impact on the telecommunications market of any one country will be relatively small, and will mainly increase the scale and especially the coverage of mobile telephony services.

Prices charged by the GMPCS operator for its fixed and mobile services in each country are the same. as shown in Table 2.1. Users must purchase their own handset, which for the mobile service is expected to cost around $1,000. It is assumed that service providers do not cross-subsidise handsets.

GMPCS both complements and competes with land cellular mobile services, but at the start of satellite services its complementarity with land mobile services is likely to be more significant. The GMPCS operators plan to offer satellite-only handsets and dual-mode handsets, which will automatically hand over calls to existing terrestrial services where they are available. It is predicted that about 70% of the GMPCS users will purchase the dual-mode handset and so also subscribe to a land mobile service. This is a necessary part of their commercial strategy because calls via satellite will be charged at a substantial premium, so most users will only want calls routed via satellite if absolutely necessary.

Again for commercial reasons, the GMPCS operator chooses to co-operate with the market leader, as this will give access to the greatest number of potential users. To persuade the market leader to co-operate, the GMPCS operator offers it an exclusive relationship. If, as in the case study, only one satellite mobile service is launched, this will give the leading land mobile service provider a significant marketing advantage. The land mobile service provider undertakes the larger part of the retailing of the satellite service.

The payphone business is assumed to be operated by domestic service providers, rather than the GMPCS operator. In the two countries examined here, it is the national operator who runs the payphones. In other cases, private service providers such as shopkeepers might operate them. Although the tariff charged by the GMPCS operator for calls from fixed telephones is much less than from mobiles, at $0.60 per minute it is still high relative to the average $0.25 per minute normally charged for long distance calls. To cover the costs of connection to the GMPCS service, monthly subscription, and to recover equipment costs, the payphone service provider must charge about $0.25 per minute on top of the $0.60. That is, if it is to break even, the average charge to users of the payphone service has to be about $0.85 per minute.

As the payphone service provider, the national operator could choose to set satellite payphone prices at a lower level, for example at $0.25 per minute – the same as the standard price for trunk calls – and absorb the extra cost. If so, then each payphone installed would make an annual loss of $19,200. Losses might be greater as lower prices would tend to stimulate demand.

The GMPCS service produces additional revenues for the national fixed and mobile telephony operators through the interconnection of calls generated by the GMPCS system. For calls made to GMPCS users, the prices charged by fixed and mobile operators exceed the interconnection charges paid to the GMPCS operator. For calls handed over by the GMPCS operator, it is assumed that the fixed and mobile operator receive their standard interconnection charge. As shown below, that means the land mobile services receive more per incoming call than do fixed services.

The cable project used in these case studies is assumed to be an optical fibre submarine cable system that will ring Africa and start service in 1999. It has 45 landing points on the continent and connects with other cables and continents, so providing global connectivity. The national operator of the coastal country decides to invest in the project by financing construction of a landing point, at a cost of $30m. The landlocked country connects to the submarine cable via a terrestrial optical fibre cable. Both countries acquire long term rights to use capacity on the submarine cable.

The cable has a nominal 40 Gbit/s capacity with a 25 year life. This capacity is more than adequate to carry all the projected telecommunications traffic to and from Africa in the immediate future. The capital and operating costs of the cable project are summarised in Table 2.3. On the basis of these costs, it is estimated that international calls can be transmitted over the cable at a cost of 11 cents (US) per minute. Strictly, this is the average cost per minute at which the project is just viable (i.e. investors will just get their money back plus interest) given the traffic forecast of $2.7 billion minutes of traffic a year. In this calculation, all the costs are attributed to voice telephony services. In effect, it is being assumed that the investors will make profits on their investment in excess of interest charges by selling or using capacity for services other voice telephony.

The national operators in both the coastal and landlocked countries will not pay for transmission on a per minute basis, but will purchase bulk capacity on a long term basis. If they order too much capacity relative to usage, the average annual cost will exceed 11 cents per minute.

To the basic transmission cost of 11 cents per minute must be added:

Costs per minute

Costs of originating and terminating calls ("national extension") 5 cents

Cost of accessing cable (for landlocked country) 4 cents

Cost of connecting calls via other cable schemes 5 cents

Allowing for these additional cost elements, the end-to-end cost of transmitting an international call is estimated at 21 cents per minute for the national operator of the coastal country, and at 26 cents per minute for the national operator of the landlocked country.

The second operator in the coastal country can choose whether to pay an interconnection charge per call minute (which would exceed 21 cents per minute as interconnection charges for international calls are based on accounting rates rather than actual costs) or to buy bulk capacity. The terms on which the second operator can buy bulk capacity are likely to be less favourable than obtained by the national operator, as the second operator will order less capacity and will place its order later. As the second operator grows fast from a small market share, it may well plan to buy some bulk capacity and use interconnection with the national operator for the excess.

Prior to the advent of the cable, the cost of international transmission via satellite is estimated at 29 cents per minute for both countries, and the end-to-end cost at 34 cents per minute. Accordingly, the submarine cable is cheaper than the alternative, by 13 cents a minute for the coastal country and by 8 cents a minute for the landlocked country.

Despite the cost advantage of cable, both countries are likely to retain satellite links, partly because the investment in an earth station has already been made, partly for calls to other African countries which are not accessible by cable. A further consideration is that, although the cable offers alternative routing in the event of a break, having access to an alternative technology is a sensible precaution. It is also quite possible that in future the cost advantage of cable may be overturned.

3. Case study 1: a coastal country

The coastal country is assumed to have a relatively large population and to be more developed than the average for sub-Saharan Africa (see Table 3.1). It has already made significant progress in restructuring its telecommunications sector, having privatised the national operator and licensed a second fixed operator. Cellular mobile services are also liberalised, with two operators in the market.

The national operator has a 100% share of the fixed voice telephony market in 1998. Having attracted a strategic investor, its total number of subscribers is now increasing at 10% a year. As a result of commitments to invest made at privatisation, the waiting list is starting to decline after rising rapidly in the recent past. However, despite the imminent arrival of a competitive network, prices charged for telephone service have yet to be adjusted towards cost, and so international services continue to contribute a high proportion of total revenue (over 70% in 1998). Average revenue per line is $650 a year and the business as a whole is reasonably profitable. Table 3.2 summarises the national operator’s profit and loss account for 1998.

As a result of privatisation, the national operator will in future pay dividends, starting in 2000. The dividend rate has been agreed at 25% of net income after interest payments. Retained earnings fall to about $ 42 million.

The national operator is subject to tariff controls. Under these controls, as well as the effect of liberalisation, tariffs are gradually rebalanced. Over the period to 2008, the average annual changes in tariffs in real terms are:

%

Monthly rentals +2

Local calls +3

Long distance calls -2

International calls -5

The second fixed operator will start service in 1999, when it expects to have 7,000 lines connected, all in the capital city. The second operator plans to grow quickly in the capital and other urban areas in order to achieve a commitment made at licence award to have 35,000 connected lines by 2003. Table 3.3 shows projected market shares for fixed voice telephony over the next ten years.

With these policies, teledensity in the coastal country more than doubles over ten years, from 0.6% in 1998 to 14% in 2008, despite rapidly growing population.

2. International services

At present, international calls are connected to other countries by a mix of terrestrial microwave (to neighbouring countries) and via an earth station working to Intelsat standards to more distant countries. The top twelve routes served by satellite are direct; several of these routes are also used for transitting traffic to other international destinations. The average cost of these arrangements is calculated at 29 cents per minute. The average accounting rate under current agreements is $1.00 per minute.

International calls to African countries are charged at the equivalent of $1.00 per minute, and to other countries at $3.00 per minute. The relatively high charges for intercontinental calls result in a substantial imbalance of traffic: incoming call minutes exceed outgoing by 50%. This imbalance in traffic produces a substantial net revenue for the national operator, on top of the profits made on outgoing international calls (see Table 3.4). Partly for this reason, the national operator is reluctant to agree to lower accounting rates. Nonetheless, accounting rates are declining at about 9% a year, which is the international average rate of reduction according to the ITU. As a result, the average accounting rate falls from $1.00 to 39 cents a minute by 2008.

As can be seen, the gross profits from international services exceed the net profits of the whole of the national operator’s business. Additional costs of about $6 million could be added to reflect the costs of the domestic network attributable to carrying outgoing and incoming international calls. Even then, it is evident that international services are contributing a substantial cross-subsidy to domestic telephone services.

Cellular radio service to the GSM standard has been provided for the past two and a half years. The cellular service provider, who is independent of the national operator, is operating profitably with 7,000 subscribers. Demand is strong and the number of subscribers is expected to increase by 100% within a year. A second GSM service provider was licensed at the same time as privatisation; it expects to achieve 2,000 subscribers in 1999, its first year of operation.

The tariffs charged for cellular radio service are significantly above those of the fixed service, though lower than the GMPCS mobile service (compare the prices in Table 3.5 with those in Table 2.1). Of course, GSM subscribers must also purchase a handset, at an average cost of $500.

By the year 2003, the GMPCS service is able to secure 2000 mobile subscribers in the coastal country rising to 5000 by 2008. These relatively low numbers are consistent with the market share assumptions of the systems expected to start service in the next few years. In addition to domestic subscribers, users roaming in the country generate additional call revenues. Finally, some 200 payphones are installed in rural areas by 2003, rising to 400 by 2008. Mobile and fixed users in the coastal country generate revenues for the GMPCS service provider of $2.6 million a year.

GMPCS has a significant positive impact on the revenues of the leading provider of land mobile services. In 2003, the revenue of land mobile services benefits by about $2 million from the extra customers generated by the GMPCS option and from GMPCS calls. The net profit of the operator rises by more than $1 million. By 2008 the association with GMPCS is worth about $4 million extra revenue and $2 million extra profit. The positive impact on mobile services arises mainly from the higher take up of cellular services stimulated by GMPCS; higher levels of calling to and from GMPCS accounts for about 20% of the extra revenues of the land mobile services.

The impact of GMPCS on the second cellular operator is not significant. GMPCS generates some additional traffic for the second operator, but its competitive position is weakened by the lack of a universal coverage option. How significant the lack of this option is in the market place is hard to assess. The disadvantage is like to be short term, as the second operator can probably sign up with another GMPCS operator in a few years.

By 2003, the national operator is charging $0.45 per minute for calls from its customers to the GMPCS, and paying out an interconnection charge of $0.20, so that it retains $0.25 per minute from GMPCS calls. The national operator also receives an interconnection payment of $0.05 per minute for calls handed over by the GMPCS operator. The national operator has extra revenues arising from GMPCS calls net of interconnection charges of about $140,000 in 2003, rising to $270,000 in 2008.

About 20% of the extra revenues generated by GMPCS for the national operator come from the satellite payphone business. The significance of the payphone business lies more in the potential contribution that it can make in extending service into remote areas. The high costs of installation and maintenance of payphones in remote areas and the relatively high charges for GMPCS fixed services combine to limit the financial attractiveness of this option. An operator with an obligation to provide service nationally would be likely to use GMPCS payphones only in areas that could not economically be linked to the terrestrial network. Even then, operators might have difficulty in charging enough for payphone calls to cover costs, whether because of tariff controls or lack of customers able and willing to pay high call charges. If standard prices are charged, the subsidy required for a network of 200 satellite payphones might exceed $4 million a year.

As the discussion of the payphone business indicates, the prices charged by the GMPCS operator mean that the service is likely to have very little impact on the competitive interaction among fixed operators. In its early years, at least, the most efficient course for the second operator may be to interconnect with GMPCS via the national operator.

By choosing to have a cable landing station, the national operator of the coastal country is making a significant investment in the submarine cable scheme. The landing point costs $30 million to construct, has an expected life of 25 years and costs $1.35 million a year to operate. Depreciating the landing station over its life, and allowing for financing costs, the operator has annual costs of $6.3 million from its involvement in the scheme.

The return from this investment comes primarily in the form of lower costs of international transmission. The cable scheme itself may prove to be profitable, and yield dividends, but this will depend on a range of factors outside the national operator’s control, such as the cost advantage of cable over satellite services, whether rival cable schemes are launched, and so on. The risk of losses being incurred is very small, provided enough coastal countries have made long term commitments to take capacity. This will probably have been a pre-condition of the scheme being built in the first place.

If all outgoing international traffic carried by the national operator is transferred to cable, the reduction in international transmission costs would be about 18 cents per minute, worth $11.5 million in the first year. In practice, it may take two or three years to agree to re-route traffic, and probably not more than 80% of international traffic would be carried over the cable. Some landlocked countries in Africa would still be reached by satellite and some by terrestrial cable or microwave. In addition, the cost of transmission by satellite can also be expected to fall. Taking all these factors into account, the annual expected cost savings may fall to $7-8 million.

The national operator has to share the benefits of lower transmission costs with the second operator. But the charges made by the national operator for international transmission are based on agreed accounting rates, rather than actual costs. As accounting rates adjust slowly, there is a significant delay before the second operator also secures benefits from access to the cable scheme. So long as accounting rates are substantially above costs, the provision of capacity by the national operator to others is good business.

The landlocked country also hands over traffic at the landing point, and so the national operator in the coastal country also gains from transit fees charged to the landlocked country.

The landlocked country has a relatively small and dispersed population and is less developed than the average for sub-Saharan Africa (see Table 4.1). It has yet to make significant progress in restructuring its telecommunications sector. The national operator has a monopoly of fixed telephony and is 100% government owned. Cellular mobile services are also monopolised.

The national operator has a 100% share of the fixed voice telephony market in 1998. Its network is very small, with less than 40,000 lines. Despite a large waiting list, the number of connected lines is increasing at only 3,000 a year. International services contribute a very high proportion of total revenue (over 70% in 1998). Average revenue per line is $450 a year and the business as a whole is profitable. Table 4.2 summarises the national operator’s profit and loss account for 1998.

The profits of the national operator’s business in the land locked country are less robust than is the case in the coastal country. The consequence is that the charges for long distance calls and for international calls cannot fall as fast, without endangering future profits. Specifically, long distance call charges are projected to fall at 1% a year in real terms (compared to 2% in real terms in the coastal country), and international calls to other continents to fall at 3.5% a year in real terms (compared to 5% a year in real terms in the coastal country). Even then, to maintain profitability, monthly rental charges increase by 5% a year in real terms.

At present, international calls are connected to other countries by a mix of terrestrial microwave (to neighbouring countries) and via an earth station working to Intelsat standards. The top twelve routes served by satellite are direct; several of these routes are used for transitting traffic to other international destinations. The average cost of these arrangements is calculated at 29 cents per minute. However, the average accounting rate under current agreements is $1.00 per minute

International calls to African countries are charged at the equivalent of $1.00 per minute, and to other countries at $3.00 per minute. The relatively high charges for intercontinental calls result in a substantial imbalance of traffic: incoming call minutes exceed outgoing by 50%. This imbalance produces significant net revenue for the national operator, on top of the profits made on outgoing international calls (see Table 4.3). Partly for this reason, the national operator is reluctant to agree to lower accounting rates.

The gross profits from international services exceed the net profits of the whole of the national operator’s business. Additional costs of about $1.5 million could be added to reflect the costs of the domestic network attributable to carrying outgoing and incoming international calls. Even then, it is evident that international services are contributing a substantial cross-subsidy to domestic telephone services.

Cellular radio service to the GSM standard has been provided for the past two and a half years. The cellular service provider, who is independent of the national operator, has just begun to operate profitably with 5,000 subscribers. Demand is strong and the number of subscribers is expected to increase by 50% within a year. The Government is considering licensing a second GSM operator, but has yet to take a decision.

The tariffs charged for cellular radio service are significantly above those of the fixed service, though lower than the GMPCS mobile service (compare the prices in Table 4.4 with those in Table 2.1). GSM subscribers must purchase a handset, at an average cost of $500.

The cellular mobile market in the landlocked country is less than a quarter the size of the market in the coastal country. The less favourable economic conditions tend also to reduce the potential market for GMPCS. This effect is more significant than the fact that the coverage of the land mobile service is poor and so satellite should be more attractive. The GMPCS service is able to secure only 500 mobile subscribers in the country by the year 2003 and 1000 by 2008. These users generate annual revenues for the GMPCS service provider of $0.7 million by 2003.

The impact of GMPCS use on the cellular operator is modest – extra revenue of $0.4 million in 2003 and $0.7 million by 2008.

Similarly, the national operator gains a small amount of revenue from GMPCS - $67,000 in 2003 rising to $130,000 by 2008.

The relatively low usage of the GMPCS system over the landlocked country creates an opportunity for greater fixed use. Accordingly, as many satellite payphones are installed in rural areas as in the coastal country – 200 by 2003 rising to 400 by 2008. The limiting factor in the use of satellite payphones in the landlocked country is demand, given the relatively high tariffs that are charged. Lower prices could be charged, and more payphones installed, if the national operator were willing to subsidise them. But the fragile state of its finances preclude this option.

To gain access to the round Africa cable, the landlocked country has to make a connection to the network of the coastal country. This link is via terrestrial cable between the two capital cities, and is some 400 kilometres in length. The estimated annualised cost of this link contributes 4 cents per minute to the cost of outgoing and incoming calls. Taking the cost of this link into account, the saving in transmission cost from routing international calls via the submarine cable is about 14 cents per minute. This produces an annual saving of about $1.5 – 2 million for the national operator. These benefits are not all passed on directly to users; the precarious profits of the business mean that international call charges can only fall at 3.5% a year.