Chapter 5. Call termination – Pockets of monopoly power

Face - Contents - Bottom - Previous/Next

"Telecompetition"

Chapter 5. Call termination – Pockets of monopoly power

Networks need to connect to other networks. Fixed networks need to connect to mobile networks. Mobile networks needs to connect to other mobile networks and the fixed network. National networks need to connect to international networks. The reasons are straightforward.

The more people you can reach through a given network, the more valuable the network is. A network can increase its value by attracting more subscribers to the network and it can increase the value of being a member, if members are able to call members of other networks. The latter is achieved through commercially negotiated interconnection agreements.

Interconnection agreements are wholesale agreements between networks regulating the conditions under which members of the two networks can connect to each other. A key part of an interconnection agreement is the call termination rate. The call termination rate is the wholesale price a network pays to be able to terminate a call originating from a caller in its own network to a receiver in the other network. An agreement must specify at least two prices for call termination in each of the two networks. These prices are often symmetric (similar).

Call termination rates for fixed networks have long been regulated, because regulators have realised that any entrant to the fixed or mobile voice market needs to connect to the fixed network to be able to compete for customers on the market. Under these circumstances, allowing the fixed network operator freedom to determine the conditions for access to its network can hardly be expected to end up with socially acceptable solutions. Call termination rates for fixed networks in the Nordic countries were in the range 1-3 €-cent per minute in 2002.

In contrast, most mobile networks have commercial freedom to enter into interconnection agreements with other domestic or international networks and agree on mutual call termination rates. Only a limited number of mobile networks with SMP status (Significant Market Power), including the incumbent mobile networks in Sweden and Norway, are required to use costbased call termination charges. We can observe that the call termination rate for mobile network in the Nordic countries in 2002 were in the range 8-15 times the price of terminating a call on a fixed network, cf. Figure 5.1. Clearly, this drives up the retail price of using all networks, the fixed network in particular, potentially reducing and distorting demand for telephone services.

Figure 5.1: Call termination rates in Nordic fixed and mobile networks

Source: European Commission, 9th Report on the Implementation of the Telecommunications Regulatory Package, 2003

High call termination rates may be justified if the underlying marginal costs of call termination on a mobile network are significantly larger than for call termination on fixed networks. However, it may also be a consequence of lack of competition as each network in reality has a monopoly on terminating calls to their own subscribers. While any caller can choose alternative networks for call origination if prices become too large, for example by call selection, he has no choice in call termination. If he needs to connect to a specific person, he has no choice but to accept the implicit call termination price on the network to which the person currently belongs.

International interconnection agreements are in principle no different from domestic interconnection agreements. International interconnection agreements are agreements that determine call termination rates between two mobile networks that happen to be of different nationality.

They only appear different because these agreements quote accounting and settlement rates, not call termination rates. We cannot observe accounting and settlement rates directly, but we can infer their relative size by comparing retail prices for domestic and international calls. Indeed, retail prices for international calls, even inter-Nordic calls, on both fixed and mobile networks seem to be disproportionately large compared to domestic calls. Prices for interNordic calls are 2-8 times larger than the price for domestic calls depending on the exact call constellation (fixed-fixed, mobile-to-fixed or mobile-to-mobile). As before, the high prices may be explained by high underlying marginal costs that are proportionately larger than the costs of domestic calls. However, it may also be due to lack of competition as each network in reality has a monopoly on call termination to subscribers in their country or region.

We now investigate in some detail the determination of call termination rates in different competitive settings in order to properly understand whether companies in the market have incentives to charge call termination rates in excess of marginal costs. If this is the case, and if we cannot identify sufficient countervailing arguments, there may be a case for regulation of mobile call termination rates. See also Appendix 2: Two way access for an extended and more theoretical exposition.

First, we consider how a mobile network competing with other mobile networks determine its own call termination rate vis-à-vis a fixed network, knowing that fixed and mobile networks do not compete for the same customers and that the call termination rate for the fixed network is regulated. Thus, the mobile network does not need to worry about any counter reaction from the fixed network.

We conclude that mobile networks, indeed, have an incentive to price call termination rates at a level (significantly) above marginal costs, even in the case when mobile networks compete fiercely for customers. However, competition indeed drives down profit to zero, as the monopoly profits on call termination rates is used to subsidise entry into the network, for example by subsidising the provision of hand-sets⁵⁸.

Interestingly, in markets with many small competing mobile networks call termination rates tend to be larger than in markets with only a few large mobile competing mobile networks. The reason is that large mobile networks tend to take much more into consideration the demanddampening effect on overall consumption from large call termination rates.

However, in markets with network externalities as telecom markets, the optimal call termination rate is indeed larger than marginal costs. It may be socially desirable to charge a mark-up above costs and use the revenue to subsidise access to the network and expand network penetration. We do not know the level of the optimal mark-up, but in this context it suffices to note that the optimal mark-up must be low in networks with very high penetration rates as the 2G-networks dominating the Nordic countries.

It may be argued that the mark-up does not matter, as all monopoly profits are recycled back to customers as hand-set subsidies. This may be correct, but still leaves the question whether it is socially desirable today to give preferential treatment to mobile services vis-à-vis fixed services financed by subscribers to the fixed network and giving rise to distorted decisions regarding the use of fixed versus mobile networks. Bomsel et al. (2003) has estimated the level of transfers from fixed subscribers to mobile subscribers to 5-10 percent of total mobile revenue in Germany, France and UK. In the Nordic countries this would correspond to about 300-600 m€ per year in Denmark, Finland, Norway and Sweden or about 20€ per year per inhabitant.

Second, we turn to the question how two mobile networks mutually would determine their call termination rate for calls originating from each other, knowing that their networks are rivals and compete for the same subscribers on the retail market. Now, the two mobile networks cannot unilaterally set their call termination rates, but need to agree with the competing network. The question is whether this makes a difference?

We conclude that most of the arguments from the previous section carry over, but some new aspects arise. It turns out that two networks negotiating mutual call termination rates for two reasons modify their rates compared to the unilateral case. First, if the two networks have some market power on the retail market and charge retail prices above costs, they realise that they have a mutual double marginalization problem and prefer to lower their call termination rates somewhat to reinstall an overall monopoly price. Second, the two networks also realise that high termination charges give the rival an incentive to compete more aggressively for subscribers and lower external call termination payments. In order to modify the competitive pressure the two networks prefer to agree on slightly lower rates than in the previous case.

Moreover, large networks may prefer to price discriminate between calls terminating in own network (on-net) and calls terminating in small rivals network (off-net). Price differentiation can harm actual entrants and deter potential entry for the same reasons as before. Both arguments level off as the market becomes more symmetric.

Finally, we note that high call termination rates make more sense for society in small networks with socially desirable growth perspectives than in large, mature networks with limited growth prospects. This could be the case for the young, struggling 3G-networks.

Lastly, we consider the special case of international interconnection and roaming agreements, being special not by nature, but because the settlement scheme is different from the one described previously. We conclude that, if there is a case for regulation of domestic interconnection, there is case for regulation of international connection, including roaming.

We also note that inter-Nordic prices on both fixed and mobile networks national are vastly above domestic prices to an extent that hardly can be explained by additional costs. We argue that increased competition in some cases can lower the prices, but it will never solve the fundamental problem of monopoly on call termination. We argue that there may be a case for coordinated Nordic action to regulate call termination in inter-Nordic interconnection agreements.

Unilateral setting of domestic call termination rates

We first consider a market with a monopoly provider of fixed voice services and several providers of mobile voice services. Fixed and mobile voice are not on the same market and do not compete for subscribers. In addition, the call termination rate for the fixed voice operator is regulated and determined by regulators.

In contrast, mobile-service provision is highly competitive, in the sense that competition between mobile-service providers for subscribers is intensive, yet each mobile network has a monopoly on termination on its own network. Therefore, potential competition problems arise not so much from lack of inter-brand competition as such, but more from the fact that residual pockets of market power remain. Therefore, pricing of call termination rates is less competitive.

First, we focus on a single mobile network and consider how it unilaterally would determine (wholesale) call termination rates for subscribers to the fixed network requesting access to its network. At this stage, we abstract from other providers of mobile voice services, cf. Figure 5.2.

Figure 5.2: Unilateral call termination rates between non-rival networks

We assume that the mobile network maximises profit from selling outbound mobile calls, covering internal costs of call origination and paying external costs on call termination for calls to the fixed network. In the simplest version, we assume that all calls originating in the mobile network is terminated in the fixed network. However, relaxing this assumption makes no difference to the results that we report below.

Under some simplifying assumptions, the setup implies that the mobile network will maximise termination revenues, irrespective of the intensity of competition between mobile networks (inter-brand rivalry). Profits are separable, in the sense that revenues from call termination are independent of retail revenues from outbound calls.

However, if services provided by the various mobile networks are identical, competition between mobile networks for subscribers drives overall profits to zero. Therefore termination revenues are fully recycled into the mobile subscriber fee. Hence, if we include the cost of the hand-set in the total cost of mobile services, we conclude immediately that termination revenues could potentially be used to subsidize hand-sets.

The main point is that the monopoly on call termination will generally induce even highly competitive mobile service providers to levy termination charges that differ from the marginal cost of termination, potentially far in excess of the socially optimal charges. Essentially, there are many separate relevant markets for termination in which even small and intensely competitive mobile service providers have monopolies. The competitive problem arises due to these residual pockets of monopoly power in markets for call termination, not due to lack of competition on the retail market.

Surprisingly, small mobile networks have larger incentives to raise termination rates than large networks. The reason is that the small network does not take into consideration that a jump in rates also increases the average price on the market and drives down demand. In contrast, a large mobile network is more likely to take into account this countervailing effect, hence, moderating its incentive to drive up call termination rates.

When the mobile network possesses network externalities, it may be privately and socially desirable to cross-subsidise handsets to attract more subscribers. A network externality arises if the value of being member of a network increases in the overall number of network members, or, more generally, members in all associated networks. In this case, it is optimal to impose a mark-up on call termination rates that generate the required revenue for crosssubsidisation. In an otherwise competitive mobile market, this is the main instrument available.

The case for a mark-up on call termination rates is especially strong in immature markets with limited membership and a large potential for further penetration; less so, if at all, in mature markets with high penetration rates. This argument is reinforced even more in the presence of switching costs between old, mature networks and new, immature and technologically advanced networks with high social priority, as e.g. the new 3G-markets.⁵⁹

Coordinated setting of domestic call termination rates

We now turn to considering how mobile voice providers jointly might agree on call termination rates. The mobile voice networks are on the same market and compete fiercely for the same subscribers, but the services they offer may be only partially substitutable leaving some room for mark-up pricing on the retail market. In this case we ignore the fixed network and focus exclusively on two mobile networks, cf. Figure 5.3.

Figure 5.3: Reciprocal call termination rates between rival or non-rival networks

We still assume that the mobile networks maximise profit from selling outbound mobile calls, covering internal costs of call origination and paying external costs on call termination on the other mobile networks. The set up still implies that the mobile networks will maximise termination revenues and tend to set high call termination rates.

Competition between the mobile networks for subscribers again drives down overall profits depending on the degree of substitutability between the mobile services. With full substitutability profit is driven down to zero, while partial substitutability leaves some room for mark-up retail pricing. As before termination revenues are recycled into the mobile subscriber fee, for example subsidizing hand-sets.

However, we now have two monopoly providers of call termination selling inputs to each other and with some latitude for mark-up pricing on the retail market. This is in essence the double marginalization problem, where two monopoly mark-ups are layered upon each other raising the overall price in the market above the monopoly price and reducing demand excessively.

The two mobile networks have a clear own interest in jointly lowering call termination rates to moderate the impact on prices and reinstall joint optimal monopoly pricing. This tendency is in the interest of society, but the two networks still aim to negotiate a joint monopoly termination rate in excess of marginal costs. This rate is still not socially optimal and a case for regulation remains.

We can take this line of argument one step further in mature, symmetric markets with balanced calling pattern. High termination rates clearly increases average costs in the two networks and also increases the incentive to capture market shares from the competitor. Accordingly, very high rates increase the risk of initiating harsher price competition and may lead two networks to agree on slightly lower termination charges to strike the right balance between high termination revenue in the short run and less competitive pressure in the long run⁶⁰.

In asymmetric markets, the case is different. Here the large mobile network faces a potential entrant or an actual competitor with a small market share. As the entrant has a very large share of off-net calls, the large mobile network clearly favours high termination rates, as it weakens the financial position of the competitor. With retail price differentiation between on-net and off-net calls it may even be beneficial to agree on taxing off-net calls to increase the pressure on the entrant. However, if the entrant still has limited physical coverage, high termination rates may induce the entrant to build up coverage at a higher speed. This dampens the incumbent's interest in high termination rates.

The dynamic features of mobile termination charges carry over from the last section. If mobile telephony displays network externalities in the sense that fixed subscriber utility is increasing in mobile penetration, then it is socially optimal to let termination charges rise above costs in order to subsidise mobile subscription. Similarly, if mobile markets are immature with few subscribers, and if further market penetration for some reason is deemed socially valuable, but potential subscribers must overcome switching cost given whatever subscription they may currently have, then it is similarly socially valuable to allow termination charges to rise above costs to subsidize mobile subscriptions.

Both of these examples suggest that unregulated termination mark-ups are less of a problem in immature markets as 3G mobile markets than in mature markets. However, since basic 2G- mobile penetration throughout the Nordic countries is close to saturation, arguments for lighthanded regulation of mobile termination charges seem to lose weight at considerable pace.

This should induce one to pause to think whether the time is ripe for time-invariant marginalcost regulation of call termination on mobile networks from the fixed net. At the very least one should think carefully about how to regulate over time. Of course, innovations in the fixed sector might also be desirable, and we are back to some of the reasons to allow the fixed network incumbent a mark-up.

A proposal has been put forward to let regulated call termination charges follow a sliding scale contingent on the degree of market maturity. In the early stages: Light-handed regulation allowing for some mark-up on call termination charges. In the later stages: Stricter regulation with call termination charges closer to marginal costs. This proposal mirrors the proposal presented in the previous section. So do the comments!

Pricing of international calls

At last, we turn to the pricing of international calls. There is no fundamental difference between interconnection agreements for national and international calls. In both cases, two networks have to negotiate reciprocal conditions for being able to terminate calls in each others network. International prices are settled in international interconnection agreements on the basis of recommendations from International Telecommunications Union (ITU). We distinguish between international calls between fixed networks and international calls between mobile networks.

International fixed calls
An international call between two fixed networks can be illustrated as in Figure 5.4. The call originates in the fixed network of country A and is send to the networks' international gateway. From here the call is transported via cable, satellite or microwave to the international gateway in the terminating country B and further on to the receiver.

Figure 5.4: International fixed-to-fixed calls

International tariffs for fixed voice have traditionally been based on the century old accounting rate system. Accounting rates are wholesale rates, which are negotiated and determined freely in bilateral interconnection agreements between operators in different countries on basis of a number of recommendations from the International Telecommunication Union (ITU). For example ITU recommends that accounting rates should be cost-based.

In a typical agreement, two operators agree on accounting rates for conveyance of calls from origination to termination. Accounting rates can be symmetric (same price in both directions) or asymmetric (different prices in the two directions) depending among others on historic differences in price structure. The accounting rate will mostly be divided equally between the operators⁶¹, although other arrangements have been seen.

The share of the accounting rate to be paid to the terminating operator is called the settlement rate. The settlement rate corresponds to the call termination rate as discussed above. The caller in the originating country pays the collection rate. This rate is often higher than the accounting rate – though not necessarily.

Historically, accounting and collection rates have been far above costs, also in the Nordic countries. However, with liberalization the accounting rate system has come under pressure, but has until now been very resistant to change. One reason could be common interests between networks in keeping up prices. As Figure 5.5 shows, retail rates for calls from the Swedish fixed network to fixed networks in other Nordic countries are still significantly larger than national prices. The same picture holds for the other Nordic countries.

Figure 5.5: Retail rates for fixed-to-fixed calls from Sweden to selected countries, 2004

Notes: All available peak and non-peak rates are collected, grouped according to call destination and ranked from lowest to highest. The overall average domestic price has an index value of 100.

Source: Teleprisguiden, http://www.cint.se/Sweden/asp20/tele_fast.asp

The pressure on the accounting rate system comes from the competition regime established nationally and internationally during the past years. Competing operators have an interest in lowering the international tariffs in order to gain customers and the margins for lowering the tariffs have been - and still are in many parts of the world – considerable.

Some competitors exploit the asymmetries in the international tariff structure. The end user price depends on the country from which the call is originated. This allows some service providers to offer low cost international rates by using Call Back Technology. When using a call back service, one has to dial a number in a country with low international rates. No one answers the call, but the service provider immediately calls back and provides a cheap line, now originating in the low cost country. IP-telephony, known as `phone card service', is also used extensively by dialling a local number with access and after verification of accounts having the possibility to place an international call via Internet.

For EU member states the principle of equal treatment⁶² applies for operators with significant market power (SMP), including all fixed network operators. They are obliged to offer the same call termination rates to both national and international operators. The rates have to be based on national call termination charges plus charges for the international link. The system, however, requires that the calling operator requests SMP charging. Nonetheless, the accounting rate system is still often used.

Even though, inter-Nordic tariffs are relatively low for fixed line telephony, retail prices for interNordic traffic still seem to be significantly above costs. For calls between directly neighbouring Nordic countries, for instance Sweden and Finland, international end user prices are twice as high as national prices. In a survey from the late 1990s the Independent Regulators Group (IRG), concluded that although the accounting rate system was showing some resistance and even though end user prices were still too high, it seemed as if international competition would bring prices down to an acceptable level and therefore no immediate regulatory actions were required.

International mobile calls (GSM) Routing of international calls from a mobile network to another mobile network is in principle similar to routing of international calls between fixed networks. However, the system must also be designed to keep track of the location of the calling and receiving mobile phones. This is done using directories for locations within own network and guest directories for locations outside own network where roaming is needed.

We can distinguish between four different scenarios, depending on the location of the calling and receiving mobile phone. Here illustrated using Danish and Norwegian phones:

1. Danish mobile in Denmark calls Norwegian mobile in Norway
2. Danish mobile in Denmark calls Danish mobile in Norway
3. Danish mobile in Norway calls Norwegian mobile in Norway
4. Danish mobile in Norway calls Danish mobile in Norway

The four scenarios are illustrated in Figure 5.6. The first scenario corresponds closely to calls between fixed networks. In the second scenario the Danish directory (HLR) contacts the Norwegian guest directory (VLR) to query the exact location of the second mobile (a*) on the foreign network. The second mobile needs access to a foreign network to be able to terminate the call. We say that the second mobile is roaming on a foreign network. When the information is received, the network automatically routes the call to Norway though international gateways. In principle, international roaming can be exposed to competition as the second mobile is not necessarily restricted to roam on a specific network, but can choose between the available networks contingent on the existence of an interconnection agreement.

In the two last scenarios the originating call in Norway is routed back to Denmark, where the query for the location of the second mobile starts. Once the other mobile has been located in Norway, the call is again routed back to Norway. In the last case international roaming is also needed to terminate the call.

Figure 5.6: International mobile-to-mobile calls

Note: The first example is an ordinary international connection. The three other examples all include elements of roaming, either roaming for call origination, roaming for call termination or roaming for both call origination and termination.

Optimal routing would avoid sending the call back to Denmark and, instead, send it directly to the mobile phone in Norway. The technology enabling this type of optimal routing has existed for the last decade or so. It has been implemented here and there, but it is not used by operators on a wide scale.

Wholesale prices for terminating calls on mobile networks abroad is also determined in bilateral commercial negotiations between operators. Mobile communication is a much newer market than fixed line communication and there is not the same century-long tradition for price setting. But the basic mechanisms are the same: There is a strong mutual interest among mobile operators for keeping international roaming prices high. For example, revenues from international roaming often constitute 15-25 percent of revenues for mobile network operators. The mechanisms are possibly even stronger than in the fixed line area, as independent service providers are not as aggressive in the international mobile calls as in international fixed calls.

Call termination on foreign networks could in principle be subject to WTO national treatment rules, requiring prices for international call termination to equal rates for national call termination roaming. However, as for fixed voice it has been agreed to defer this question to the next (present) round of WTO negotiations.

Consequently, international wholesale international mobile prices and, therefore, mobile retail prices for using mobile phones in foreign countries are exceedingly high, as is illustrated in Table 5.1 and Table 5.2.

Table 5.1: TDC mobile-to-mobile peak retail minute prices

Note: TDC Mobile Extra with `vacation module', all prices paid are in daytime. Prices paid are with the so-called `vacation module', which apply automatically for MobilExtra subscriptions without extra subscription fee. With a `vacation module' foreign countries are gathered into different areas with fixed prices. A call to Denmark from a TDC mobile phone at any time in any of the other Nordic countries costs 5.00 DKK per minute, and receiving a call from Denmark costs the roaming subscriber 3.00 DKK.

Table 5.2: TDC mobile-to-fixed peak retail minute prices, August 2004

Note: TDC Mobile Extra with `vacation module', all prices paid are in daytime. Prices differ when roaming on mobile networks of different operators in the individual countries. The operators chosen here are in Finland: FINNET; on Iceland: Siminn; in Norway: Telenor; in Sweden: TeliaSonera.

The issue of excessive roaming charges has been raised many times by the International Telecommunications User Group (INTUG) and has also been examined by the European Commission. DG Competition concluded in 2003 that international roaming suffers from significant competition problems, applying not only to voice services, but also to other services as SMS and GPRS.

Footnotes

58 Mobile networks may anyway end up with positive profits, if the services offered from the mobile networks are only partial substitutes.

59 Add to this that there may be a certain amount of ex post subscriber lock-in. This might push mobile service providers into the red in the short term, despite sizeable termination surpluses. This seems to be broadly consistent with recent empirical observations of the mobile-phone industry in particular. The mobile termination charges are tempered somewhat by network externalities. This, however, does not obviate the need for regulation.

60 With retail price differentiation between on-net (calls to own network) and off-net calls (calls to competitor's network) it may even be beneficial to agree on subsidising off-net calls to credibly enter into a collusive agreement.

61 The accounting rate is split differently, if the call is transmitted by sea cable or satellite or if it requires transit through other countries.

62 WTO rules for interconnection would require that call termination rates should be equal for all operators, national as international. However, it was agreed to defer the issue to the next (i.e. current) round of negotiations.

Version 1.0 October 2004 • © Danish Competition Authority.
Published by the Danish Competition Authority, www.ks.dk
Publication produced according to the standard for electronic publication set by the Government