5 The estimation of costs

5.1 Introduction

Here we explain our general approach to the estimation of costs, the process used to draw up the costs for use in the cost-benefit model, and how the numbers in the economic evaluation model were derived.

5.2 General approach

To date no country has introduced NP into mobile networks, and so there are no direct experiences from which real costs can be drawn. While cost benefit analyses have been done in a few countries of mobile NP, the cost figures used are still broad estimates.

There are more challenges in the identification of costs:

• In many cases individual activities serve multiple objectives and therefore their costs have to be allocated between these different objectives. The allocations may be arbitrary and will not necessarily reflect the decisions that would be made by a profit maximising business.

• The costs borne by operators may be the prices that they can negotiate with suppliers. If the nature of the supply is special, the price will be set by negotiation without any recognised price expectation. A critical factor in the negotiation will be the extent to which the operator can consider alternative suppliers. If he is locked into his suppliers then the price may contain a very high element of profit.

•  The cost benefit analysis is concerned with the costs and benefits to the UK as a whole and therefore a distinction needs to be made between additional work occasioned inside the UK and additional work occasioned outside.

•  The nature of the costs in many cases is a high initial investment in the development of a capability and a lower recurring cost per item (eg software)

•  The range of costs for software development is very high, because the difference in productivity between different software programmers and system designers can be more than an order of magnitude.

•  Costs for some items may change significantly as a result of technological developments and from economies of scale gained by suppliers form a greater number of countries applying NP following the requirements proposed by the European Commission. (The evaluation compares the early introduction of number portability with no number portability at all, ie it ignores the effect of a requirement from the Commission)

•  There may be opportunity costs. The main potential areas are the use of scarce resources to support number portability development when these resources could be deployed on work that would yield greater benefits to the UK as a whole (not greater profits to the operator concerned), and any effect that the choice of solution has in inhibiting the development and implementation of new services and features that would be of benefit to customers.

• Cost will be lower the later they are incurred because of scope for achieving economies of scope in combining the activity with other activities.

In order to overcome these difficulties, we have taken the following approach:

• where systems and functions will be provided anyway, we include only the marginal increase in costs necessary for the support of number portability;

• where there is no established price expectation and prices are set by negotiation, we use an estimate that takes account of the development cost and the level of competition;

• where a product or function has been developed already within the UK with significant development cost but zero production cost, the cost to be included in the analysis is zero, because the analysis is concerned with the additional cost to the UK as a whole. Thus the acquisition of additional software by a network operator may cost the operator a great deal of money but if the supplier has already developed the product and is in the UK, the cost in terms of the analysis will be zero, or rather just the cost of handling the sale, installation and testing;

• where a product is developed within the UK specifically for number portability, the costs to be included should be based on an estimate of the man hours and typical labour costs, plus a reasonable profit. Excess profits, which may well exist and be very large, should not be included because they stay within the UK;

• where a product is developed outside the UK, excess profits are included because they are a cost to the UK as a whole;

• opportunity costs related to additional labour costs are included, but other opportunity costs could not be included because no operator was able to give us a reasonable guide to the possible size of these opportunity costs;

• costs that occur after start-up are reduced by 5% for year that they occur after start-up to take account of technological developments. This is not to be confused with discounting in determining the Net Present Value.

TACS

Porting from TACS to GSM has already been introduced by the GSM operators. The scale of the facilities are expected to be designed to allow for the great majority of existing customers to port to GSM. If they decide to port from TACS to DCS-1800, they can port via GSM. The costs of porting to GSM are already incurred, and the costs of porting from GSM to DCS-1800 are the same as porting between different GSM networks. Therefore the cost of porting from TACS to the GSM or DCS-1800 service of another network will be the same as porting between GSM networks.

Spectrum

Spectrum for DCS-1800 has slightly different propagation characteristics to that of GSM. These characteristics will result in greater network costs and give it the potential for greater re-use of spectrum from its shorter range. We therefore take the view that the extra network costs and benefits will be reflected in the valuation of the spectrum and so both effects can be ignored.

5.3 Method of providing portability in the networks

The basic assumption is that portability will involve the cessation of the account with the donor service provider and network and the establishment of a new account with the recipient service provider and network. A new IMSI will be issued and a new SIM fitted to the mobile.

The method assumed for handling incoming calls is signalling relay, where the original GMSC obtains an MSRN from the HLR of the recipient network. Signalling relay is understood to be the approach preferred by the operators.

A method is also needed for handling of outgoing calls from the recipient network to mobiles with imported numbers. Either all outgoing calls must be screened or calls to imported numbers will be tromboned to the original network and back, possibly with dropback. We assume that tromboning will be used to avoid increased set-up costs. At some stage a screening solution may become more cost effective but we expect that if the operators are content with the commercial implications of signalling relay then they will accept tromboning. This approach may produce a slight overestimate of costs.

Developments are also needed for the SMS service to provide portability. SMS is a partially standardised service that is not fully interconnected within the UK at present. No system is standardised for forwarding messages between SMS Centres but such a facility could be developed relatively easily. The SMS service is complicated by the fact that a customer of a UK network may be registered with an SMSC outside the UK. The tariffing and billing arrangements for SMS are not yet fully developed, and the service is being used beyond the current boundaries of its design. For the purpose of this analysis, we assume that the SMSCs are interconnected because this is an expected development of this service.

In the longer term a modified version of signalling relay may be introduced as a result of standardisation work that it just beginning within ETSI. It is not possible to predict the outcome of that work, but alternative methods are unlikely to be introduced unless they improve on the cost benefits of the earlier methods.

Market structure

Services are normally provided to end customers through service providers and dealers. Service providers resell airtime from network operators, and have the continuing relationship with the customer. In the case of DCS-1800 (PCN) the service provider function may be part of the network operator's own organisation. Dealers handle the sales to customers and establish contracts between the customers and the service providers for airtime. Customers are billed by the service providers and do not have direct contracts with the network operators. Network operators compete with respect to end users by direct advertising and branding and by offering special features and quality. They compete with respect to service providers mainly by offering special bonus arrangements.

Implications for Dealers and Service Providers

Number portability will have significant implications for dealers and service providers. Some customers will change service provider coincidentally with changing network operators, others will not. It is not yet clear how service providers will handle number portability and whether some form of one-stop shopping will develop where a recipient service provider both opens the new account and handles the closure of the old account on behalf of the customer.

The administrative systems of service providers are in most cases connected to the administrative systems of the network operators so that customers can be brought on-line almost immediately.

A procedure needs to be designed for the porting of a number from one operator to another and the synchronisation of this procedure with the issue of a new SIM and the withdrawal of the old SIM. These procedures have not yet been designed and agreed. In the fixed networks, faxes, and more recently EDI messages, are sent between operators. In contrast the operators in the Netherlands are studying the development of a distributed transactional database to handle the transactions.

5.4 Cost elements

We attach to this section a schedule setting out the various cost elements and the estimates for each type. As illustrated in Figure 5.1, the bulk of these costs lie in the customer costs, which consist of new handsets and SIM cards. The percentage figures in this diagram are Net Present Values from the flow of costs over a ten year period. The system set-up costs, often the costs that attract much of the attention, are a small part of total costs (at 4% of the total). The caller costs category is too small to be shown.

Figure 5.1: Cost categories

System set up costs

We define these as the costs necessary in order to carry the first ported call. They include costs required to condition the networks, develop software, and to change procedures for inter-working between the network operators, service providers and dealers.

In Figure 5.2 we show our estimates for the set up costs. The detailed calculations are shown in the Appendix.

Figure 5.2: System set up costs for operators, service providers and dealers

In Figure 5.3 we illustrate how the set up costs are distributed between the operators, service providers and dealers. The bulk of these costs will be borne by the network operators.

Figure 5.3: Distribution of set up costs

Customer transfer costs

Each time a customer ports, service providers will have to close down an account, open up a new account, and co-ordinate the network operators in the switching over of the mobile number and routing of calls. These will be one-off costs, which we estimate to be £19.00 per port.

We have not applied all of the costs equally to Type A and Type B customers. The costs of closing and opening accounts will be incurred by Type A customers whether or not NP is available, and so should not be included. We have therefore estimated that the additional costs involved in porting for Type A customers are only £5 per port, and we have applied the figure of £2 only to Type B customers.

In addition network operators may have to screen outgoing calls so that calls made to a number that has been imported into their network are kept in the network rather than sent to the donor network. We have allowed an estimate for this of £1.50 per ported number. In practice operators may accept tromboning for these calls if this is a cheaper solution. This would then make this figure an over-estimate.

We calculate that the total customer transfer costs, after a reduction of 5% per year to reflect cost improvements, and discounting at 8% per year, will be £49.5m. In Figure 5.4 we show how these are distributed over the ten year period.

Figure 5.4: Customer transfer costs

Additional traffic costs

Calls to ported numbers have to be re-routed from the donor network to the new network. In Figure 5.5 we show the process used to estimate these costs. They are applicable to calls to all ported numbers, whether or not they are Type A or to Type B customers. We have to allow for customers who port more than once (although they are counted as a new port in our estimates, their calls must not be double counted). These numbers are then accumulated year by year, and using estimates of the call minutes received by customers (1.36 minutes per mobile per day was provided by the network operators), we are able to calculate the total volume of call minutes to ported customers.

Figure 5.5: Estimating additional traffic costs

We have estimated the additional cost of carrying traffic to ported numbers as 0.65 p per minute. This figures include 0.05 p for additional signalling costs. We explain below how we calculated these costs.

Incoming calls from external networks are handled by a signalling interaction with the recipient network and are then normally routed across to the recipient network. If the basic form of signalling relay is implemented, then a MSRN will be issued through the original GMSC, and if the called party is roaming in another network, or has set an unconditional divert to an off-net number, the call will be routed directly and not enter the recipient network.

The cost calculations assume that in all cases the call will be routed on to the recipient network which therefore overestimate the costs involved. There are two elements of additional conveyance cost:

• the additional traffic costs involving one extra inter network link (switching element plus transmission element)

• the additional signalling

The additional traffic costs were calculated in three ways:

• From the Oftel November 1996 Determination of BT's interconnection charges.

• From confidential figures submitted by Vodafone to Oftel in May 1996 following an audit by Touche Ross.

• From a confidential cost benefit analysis performed by NERA in 1994.

The calculations gave the results shown in Figure 5.6

Figure 5.6: Calculation of additional traffic costs

Because the Oftel Determination applied to BT's fixed network and costs for switches for mobiles may be greater than for fixed, and in order to err on the side of over-estimating the costs, a figure of 0.6 p/min was used for the additional traffic.

Oftel's January 1997 Determination of the additional conveyance cost for dropback between the fixed trunk and local exchanges was 0.119 p/min. In terms of switch activity and transmission, dropback involves much more activity than support of signalling relay. We therefore estimated that a figure of 0.05p/min will provide a more than adequate allowance for the additional signalling costs.

This gave the total additional conveyance costs as 0.65p/min.

In Figure 5.7 we show how these costs are distributed over the ten year period. We estimate that the total cost over ten years, after a reduction of 5% per year to reflect cost improvements, and discounting at 8% per year, is £37m.

Figure 5.7: Additional traffic costs

Customer costs

We have included two costs under this heading:

• the costs of new handsets for Type B customers who are moving from analogue to digital networks, from GSM to DCS networks, and from DCS networks to GSM

• the costs of new SIM cards for all other Type B porting customers.

We show the method of calculating customer costs in Figure 5.8.

Figure 5.8: Estimating customer costs

In order to calculate the costs of new handsets, we have taken the number of Type B customers porting from analogue networks as all of these will need new handsets. At present all customers moving between DCS and GSM networks will also need new handsets, but the dual band handsets (which are being introduced from 1997) will remove the need for new handsets. We have assumed that 2.5% of customers porting between the digital networks will have dual band handsets in 1999, and that this figure will have risen to 100% by the year 2004.

We have used a figure of £200 per handset as the resource cost. Prices to dealers range from under £100 for analogue budget handsets to over £400 for premium digital handsets. Although most handsets are available at a much lower price to the retail customer, we have ignored the subsidy element. We have allowed a figure for the increased utility of the handset to the customer, which may be in the form of greater functionality, improved appearance, or battery life etc. We have allowed £50 for this improvement, thus giving a figure of £150 per handset for the economic evaluation.

All Type B customers porting within the GSM and the DCS networks do not need to change their handsets, but will need a new SIM card. We have allowed a resource cost of £10 per SIM card for this.

The total customer costs are £93.7m, after a reduction of 5% per year to reflect cost improvements, and discounting at 8% per year. We show the year by year costs in Figure 5.9.

Figure 5.9: Customer costs

Caller costs

Calls to ported numbers will suffer a small delay during the set up process because of the relaying of the call to the new location of the ported number. We estimate that this will be similar to that found in the fixed network - 0.3 sec per call. This will be experienced on all calls to ported numbers, whether or not it is to a Type A or to a Type B customer.

We have used 482 per year as the number of calls received by mobile customers, based on data provided by the mobile operators. We do not know whether this number will rise or fall in the future. While call volumes are likely to rise as a consequence of rising income and greater use of telephony, the increasing proportion of personal customers may result in the average falling. We have therefore used a constant figure over the ten year period.

We have taken as the value of time £8.70 per hour, the average wage rate in the UK in 1996, and added 25% to cover the costs of National Health Insurance etc. This gives a figure of £10.88 for the model. We have assumed that the same value of time is applicable to personal callers as to business callers. Other studies have used a lower rate for personal callers, but we have made this assumption because:

• it avoids the lengthy debates over the valuation of personal time;

• it may be that business calls to mobiles are made by higher paid staff, and the use of average wage rates underestimates this;

• it is a simpler approach, and avoids having to make dubious assumptions about whether callers are making calls in personal or business time.

We have not allowed for any real increases in the average wage rate over the ten year period. This keeps the approach simple, and in any case a change would have no significant impact on the total outcome.

The caller costs are relatively insignificant at £1.5m over the ten year period, after a reduction of 5% per year to reflect cost improvements, and discounting at 8% per year.

Total costs

The total costs for the economic evaluation model are therefore £190m, and in Figure 5.10 we show how these are distributed between the categories of costs and the various players (assuming that there are no mechanisms for recovering the costs of number portability from each other). We cannot be more specific about how the costs vary between the players for the customer transfer costs because we do not know how these will be divided between the network operators and the service providers. The distribution of the customer costs will depend on the prices to the customers and the cross-subsidy arrangements between the operators, service providers and the dealers.

Figure 5.10: Distribution of costs of number portability

Click here to return to contents

6. The results of the economic evaluation

6.1 Introduction

We describe in this section the results of bringing together the benefits of Section 4 and the costs of Section 5, along with the sensitivity analyses.

6.2 Timing issues

In preparing for the economic evaluation, we chose a central case, against which a number of changes in the basic assumptions can be made as part of the sensitivity analyses. One of the key assumptions in the central case concerned the timing of number portability.

We have chosen a ten year period for study in our central case starting in 1998 and finishing in 2007. We assumed that the set up costs would be incurred in the first year (1998), and that NP would be available from 1999. However we have not done any detailed analysis of the feasibility of these dates, and they are made for the purposes of simplifying this study only.

The relationship of these assumptions with the changes in the numbering plan announced by Oftel in early 1997 need explanation. As part of the changes to the numbering plan, all mobile numbers (which were unaffected by the introduction of the 01 code change in 1995) will be moved to the 07 range. Most mobile numbers will simply have the additional digit added (for example, 0836 will then become 07836). A few number ranges will have more extensive changes (for example 0385 will move to 07785). the programme of conversion is as follows:

• from July 1997 new mobile numbers will be in the 07 range

• from December 1998 a period of parallel running will start, and this will allow calls to both the old and the 07 new number to reach the same destination

• NP will be permitted only for 07 numbers

• after 31 December 2000, all mobile customers will be contactable only on their number in the 07 number range.

This raises two issues for this study:

• should the introduction of NP be delayed until 2001 because of these changes?

• should we exclude numbers that have not been moved to the 07 range?

Delay of NP

We concluded that our analysis should assume that NP will start in 1999 because parallel running will permit most of the benefits of NP to be retained. For callers, parallel running will allow a call to the old and the new number to be delivered, so the same benefits will be derived. After 2001, callers will be aware that mobile numbers start with 07 from the likely advertising campaigns and publicity, and so can make the adjustment themselves. Porting customers will have sufficient time to change their stationery, inform callers etc over the two year period in the course of stationery re-ordering, regular contact with callers, etc.

The only exception is for numbers where a double digit change is needed - these callers are less likely to guess the change after 2001. However there are no records to the numbers actually in use in these ranges because the allocation of individual numbers is done by the service providers. We therefore cannot say whether this will be a significant exception or not.

Restriction to 07 numbers

For similar reasons, we concluded that we should not exclude numbers in the old number ranges from the study. Although the number of a porting customer will have to be changed to the 07 range, the customer will retain all the other digits with NP. For the reasons explained above this will make no difference to callers, and the customer can phase in changes to stationery etc.

Short message service

As explained in Section 5, there are technical problems in the porting of the SMS service, and it will take time before a standardised solution is available. We have assumed that this will be resolved by 1999. Our surveys suggested that the availability of this service is important to some customers, and that if it is delayed beyond then, the demand for porting may be reduced. However, based on discussions with the industry, we have assumed that this problem would be resolved over the next one to two years.

6.3 The central case

Our key assumptions for the central case were:

• the overall views expressed by the industry on market growth and churn rates will apply (as set out in the base case and the switching model papers);

• the market size in the year 2007 will be 19.55m customers, giving a penetration rate of 33%;

• the levels of benefits derived from the survey results will apply

• the costs estimated in Section 5 will apply

Our analysis shows that a critical variable in the economic evaluation is the estimates of churn. The higher the rates of customers leaving networks, the greater the numbers that will join other networks, and hence the greater the costs and benefits of number portability. We have made separate estimates of churn rates for personal, SME and corporate markets, and we show our overall churn estimates for the central case in Figure 6.1. Over the last few years Oftel and the operators estimate that between 18 and 24% of customers have left mobile networks each year (Source: Market Information Update, Oftel). We believe that this rate will fall as the market matures (see Base Case paper in the Appendix).

Figure 6.1: Central case churn estimates

The central case gives a net present value of +£98m, and in Figure 6.2 we show the year by year cash flows. All the figures have been discounted at 8% per year, and we show the discounted figure in this table. The Type B benefits are negative in 1999 and 2000 because of the impact on industry profits.

Figure 6.2: The central case (£m)

In Figure 6.3 we illustrate the magnitude of the various costs and benefits of the central case.

Figure 6.3: The costs and benefits of the central case

The central case shows that the introduction of number portability into the mobile industry in the UK will be of national economic benefit, and therefore there is a good case that it should be introduced.

However we need to perform sensitivity tests to show how well this recommendation stands up to changes in the underlying assumptions. We first consider changes to the assumptions about growth and churn levels, and then see how far the costs and benefit figures in the central case have to change before the recommendation becomes questionable.

6.4 The low case

In this case we assume that the growth of the mobile market is less buoyant, and that churn levels are lower (both with and without NP), perhaps reflecting less competition between the operators and service providers. Our key assumptions are:

• the most pessimistic views expressed by the industry on market growth will apply;

• the market size in the year 2007 will be 17m customers, giving a penetration rate of 29%;

• the churn rates are reduced by 33%, and we show the consolidated rates for this case in Figure 6.4;

• the levels of benefits per porting customer derived from the survey results will apply

• the costs estimated in Section 5 will apply

Figure 6.4: Low case churn estimates

The NPV for the low case falls by £54m to £44m, and we show the year by year results in Figure 6.5. It is worth noting that almost all this fall derives from the change in assumptions about switching rates. In the model very little is attributable to the lower growth rates. Nevertheless, the NPV is still positive, suggesting that mobile NP is still worth implementing even if this scenario applies. We believe that it is very unlikely that churn rates will fall to 4% in a competitive consumer industry such as mobile telecommunications.

Figure 6.5: The low case (£m)

In Figure 6.6 we illustrate the magnitude of the various costs and benefits of the low case.

Figure 6.6: The costs and benefits of the low case

6.5 High case

Here we take the opposite view, and assume that the market becomes even more competitive, resulting in much faster growth and switching rates than expected by the industry. We assume that:

• the market size in the year 2007 will be 24m customers, giving a penetration rate of 41%;

• switching rates are increased by 33%, and we use the consolidated churn rates in Figure 6.7;

• the levels of benefits per porting customer derived from the survey results will apply;

• the costs estimated in Section 5 will apply.

Figure 6.7: High case churn estimates

Here we find that the NPV rises to £146m, an increase of £48m on the central case. Again most of the increase is due to the greater churn levels, rather than the changes in the number of mobile customers. However in this case there is a negative impact from the Type B customers in years 2 and 3 because of the impact on industry profits. Nevertheless under this scenario, mobile NP becomes a greater benefit to the nation.

Figure 6.8: The high case (£m)

 In Figure 6.9 we illustrate the magnitude of the various costs and benefits of the high case.

Figure 6.9: The costs and benefits of the high case

6.6 Sensitivity analysis of the central case

In this analysis, we change the main variables in the central case to see by how much they have to change in order for the NPV to approach zero. We accept that there will be some latitude in our estimates, and although our user surveys are probably the most extensive undertaken, the results are still subject to statistical error. We can then judge whether it is reasonable to expect these changes to take place. If it seems possible that the changes will occur, then the decision to introduce mobile NP must be questioned. If it is unlikely that the changes will occur, then a decision to introduce NP seems robust.

We therefore changed the following questions:

• how far do the benefits have to change?

• what happens if there are no Type B benefits?

• how far do switching rates have to fall?

• by how much do costs have to increase?

• what combinations of costs increase and benefits decline are critical?

We summarise the results of these analyses in Figure 6.10

Figure 6.10: Results of sensitivity analyses

This table suggests that there have to be substantial changes in the underlying assumptions before the NPV becomes close to zero. We therefore conclude that the economic case for mobile NP is robust, and that its introduction will be of overall benefit to the UK.

We also examined the results if cash flows for a period of only five years are considered. Hence all costs and benefits from the year 2003 to the year 2007 would be ignored. The NPV then falls to + £1 m. While this is still positive, this much figure would not stand up well to further sensitivity testing. However we do not consider it reasonable to exclude the costs and benefits after 2003 as they will exist. The discounting mechanism enables us to recognise that they will be less valuable than those incurred in the first five years.

6.7 Less quantifiable effects

Our economic evaluation has omitted several effects because they are more difficult to measure. This is not to say that they are not important, but that inclusion of them would reduce the credibility of the more easily quantifiable results which we have included in our model. Some of the possible effects were raised during our discussions, and they need to be considered as part of the overall evaluation.

"Type 2" benefits

In the NERA study of NP in the fixed network most of the benefits were attributed to the effects of increased competition that would result from the introduction of NP, which were defined as Type 2 benefits. NERA estimated that about £1,280m of benefits (some 69% of the total) would be derived from this source, mainly as a result of the alleged improved efficiency in BT due to NP. For reasons explained in Sections 2 and 3 we do not believe that the impact of this type of benefit is likely to be so marked in mobile networks. NERA's estimates caused extensive debate at the MMC Inquiry, and fortunately the NPV of the central case here is positive without having to rely on the more controversial Type 2 benefits.

Nevertheless, we believe that there will be some benefits to overall competition from the introduction of mobile NP, and that these benefits could be significant. There are three possible sources:

• NP may make the mobile networks more efficient. Figures submitted to Oftel by the mobile operators suggest that if NP resulted in one of the operators bringing their costs per minute down to the level of their main competitor two years earlier as a result of NP, there would be an efficiency gain of the order of £350m. We calculated the value by taking the difference in the fully distributed costs per minute, multiplying it by the number of call minutes over two years, and discounting for the value of the saving in year two. While this is not an accurate method (for example it does not include the costs of bringing forward by two years the necessary investments needed to secure the efficiency improvements) it illustrates that Type 2 benefits could be at least as large as the Type 1 and Type 3 benefits.

• NP may well increase competition between service providers, especially in terms of the customer service. It may also give service providers greater market power, enabling them to negotiate better deals with network operators, thus making them more efficient.

• It may also lead to a shake-out of the handset subsidy system, bringing prices charged to the customer more in line with costs. This rebalancing will be a significant overall national economic benefit.

Prices to callers

For some time BT has charged a lower price to fixed network customers calling the DCS networks than to those calling the GSM networks. A mobile customer porting from GSM to DCS will thereby confer a benefit on a caller of lower prices for fixed to mobile calls. This will, of course be offset by mobile customers moving in the opposite direction, although the evidence from the user surveys suggest that most of the movement will be from GSM to DCS.

We have chosen to ignore this potential benefit because operators find practical problems in changing the charges for calls coming into ported numbers. Moreover, it is likely that the differential will disappear in the next few years. Oftel has already put pressure on the operators to remove the differential, and it was reduced in late 1996. Indeed NP may well have the effect of undermining the main cause of the difference - the difference in interconnection prices - and hence we do not expect this benefit to exist much beyond 1999/2000.

Effect of personal numbering

Several companies are now offering a so-called "Lifetime" number, which can be used whether the called party is on a fixed phone, a mobile phone, or in a new location.

We do not see this as a substitute for NP - personal numbering is a premium service designed for almost daily changes of number. Moreover a change of number is necessary to gain the personal number, to some extent defeating the objective of NP. However some subscribers to personal numbers may well have been porting customers if they did not have personal numbers. Hence personal numbering may reduce the demand for NP.

However to date evidence about the actual take-up of personal numbers suggest that personal numbering will be a premium service rather than one that appeals to large numbers of people. We therefore believe that personal numbering will not have a significant impact on the demand for mobile NP.

Fixed-mobile convergence

Most experts predict that the mobile telephone will increasingly substitute for the fixed telephone, and there is evidence of this happening in the Scandinavian countries. This will make the mobile number more important, and hence increase the value of mobile NP.

We have also decided to ignore this effect because it is difficult to estimate at what speed this will happen, and by how much the value customers put on their numbers will increase.

Opportunity costs

The network operator and service providers point out that they will have to dedicate scarce resources to introducing NP, and that these resources would be used on other innovations if NP were not introduced. If this is so, there is an opportunity cost attached to the introduction of mobile NP.

To some extent these scarce resources can be augmented - for example by employing additional staff - and we have included these costs in our estimates. However some of the staff resources may not be expandable in this way, for example programmers who have an intimate understanding of the billing systems, or the number of times that software can be updated in a year may be limited.

We accept that there may be opportunity costs of this nature. However it has proved impossible to quantify them as the network operators were unable to indicate what would be forgone as a result of NP, and what the value of this would be. Hence we have to acknowledge the possible existence of opportunity costs, but cannot include them in our evaluation.

Hence there are a number of wider costs and wider benefits that we have excluded from our analysis. We believe that on balance their overall effect will be positive, thus adding to the argument for introducing mobile NP.

Click here to return to contents

7. Charging principles

7.1 Introduction

In Sections 4 and 5 we discussed the benefits and costs created by the introduction of NP. However here, as in examples of NP in other countries, the costs and benefits are not distributed evenly, and there is considerable debate about who should pay for the costs of NP. This section sets out some principles and suggestions for the mobile industry in the UK.

7.2 The issues

The problem is essentially that many of the costs are borne by parties that do not gain the benefits. In particular, the operator losing the customer bears some administrative costs and the costs of re-routing calls to ported numbers, as well as losing revenue from the customer. Meanwhile the new operator has some costs of administration, but gains the benefit of the revenue stream from the new customer.

In Figure 7.1 we illustrate the mismatch between costs and benefits, but we make no attempt to estimate the scale of these. This illustrates that the main loser is likely to be the donor network and its customers or shareholders (who will have to pay for the excess of expenditure over income by higher charges, lower service quality, or decreased profits). The main gainers are the recipient network and its stakeholders, the porting customer and the callers to ported numbers. The majority of callers to mobile numbers are on the fixed network, and under the solution proposed in the UK, the fixed network will not have any additional costs, making the fixed network and its stakeholders net gainers from the introduction of NP into the mobile networks.

Figure 7.1: Distribution of costs and benefits

+ = benefit

- = cost

We can describe the distribution of costs and benefits summarised in Figure 7.1 in more detail:

• the system set up costs are incurred by all operators, and by permitting NP give benefits to porting customers, callers to ported numbers, and to all mobile customer (who will benefit from the wider impacts on NP of competition discussed in Section 6.7);

• the customer transfer costs are shared between the donor and recipient operator (we include the service providers and dealers in this term for the sake of simplicity), and the porting customer gains the utility of costs saved from NP or the benefits of switching supplier. The recipient network gains the revenue from the new customer;

• the additional traffic costs are borne by the donor network, to the benefit of the recipient network (which gains interconnection revenue), the caller (who can have the call routed without the costs of finding the new number) and the porting customer (who presumably gains some benefit from receiving the call);

• the customer costs (the new handset or SIM card) are shared between the new operator (including service provider and dealers) and the customer in some proportion. The porting customer gains a new handset and the recipient operator gains from the new rental payments;

• finally, the caller costs of additional call set up delay are borne by the callers, but they will gain from being able to complete the call successfully.

From this complex web of interactions, a wide range of views has developed on how to deal with the costs of number portability. There are four main options:

• Option A: the donor operator charges the receiving operator the costs caused by a customer porting his number. The receiving operator may, or may not pass on the charge to the customer;

• Option B: the donor operator charges the porting customer the costs caused by that customer porting his number

• Option C: operators bear all their own costs

• Option D: the donor and receiving operators share the costs in some proportion.

There are a number of different economic principles that can be applied to these options. Each can be backed up by good arguments and each has different financial implications.

7.3 Economic principles

During the MMC Inquiry into NP in the fixed network, the Commission adopted six principles for the allocation of costs in number portability. Although their application to mobile may give different results to those derived by the MMC for the fixed network, they provide a good framework for this discussion.

• cost causation: the costs of NP should be allocated to those who cause them. This gives cost causers the right price signal and encourages economically efficient behaviour from them. Cost causation is a fundamental principle in developing charging regimes in telecommunications. It forms the basis of interconnect charging in the UK and elsewhere. On its own it suggests that other operators should recover in full the customer transfer costs and additional traffic costs from the porting customer or from the receiving operator (as the customer's agent).

• cost minimisation: the costs of NP should be allocated so as to give operators an incentive to minimise the costs of NP. This suggests limiting the proportion of an operator's costs which it can recover from the receiving operator. Alternatively the price an operator charges for NP services can be fixed so that the donor operator keeps the residual cost savings as profit.

• distribution of benefits: benefits from NP accrue to callers and to users in general (through fiercer competition) as well as to those porting their numbers. Maximising these external benefits means stimulating demand for NP by limiting the prices which are charged for NP services - that is only partial cost recovery.

• effective competition: the methods of costs recovery should not frustrate the wider benefits that NP will bring to all customers. For example, a high charge to porting customers may deter most customers from porting, thus making it unnecessary for operators to reduce prices or improve services, which would benefit all their customers.

• reciprocity and symmetry: a rule for the costs of a customer porting from one operator to another should apply to a customer porting in the opposite direction.

• practicality: the rules should be easy to implement.

In Figure 7.2 we summarise how these were applied to the fixed network. In the first few lines we show how the six principles led to the allocation of costs either to each operator bearing its own costs, to costing sharing between the operators, or to the cost causer bearing the cost. In the three last lines we show how these were then applied to the system set up costs, the customer transfer costs (called the per line set up costs in the MMC Inquiry) and the additional traffic (conveyance) costs. The conveyance costs were to be shared for the call tromboning method, and then borne by the donor operator (mainly BT) once the call dropback method was used.

Figure 7.2: Allocation of costs in the fixed network

In terms of the options outlined above for the distribution of costs (section 7.2) the MMC leant heavily towards Option C (operators bearing their own costs) and Options A ( charges to the receiving network) and D (cost sharing). A number of recipient operators do make one-off charges to the porting customer for the administration of porting.

7.4 Application to mobile networks

There are a number of important differences between mobile and fixed networks that lead to a rather different application of principles to the allocation of costs.

• the existence of the service provider and dealer levels in mobile add to the complexity of charging. There are many more organisations involved in porting, and it makes it more difficult to develop and enforce simple and practical charging arrangements.

• there is no one dominant operator in the mobile market, in the same way that BT has about 90% of the market in the fixed network. Whereas the direction of switching in the fixed network is almost totally from BT, although this is starting to change now, in mobile networks there will be a good deal of switching between the operators, some of which will cancel each other out. The net flows of cost charges may therefore be relatively small in comparison to the gross flows.

• for the same reasons, the benefits of NP will be more evenly spread between the network operators - a donor operator one day will be a recipient the next day. In the same way the wider Type 2 benefits (those to all customer of greater competition) will also be spread more evenly between the customers of the network operators.

• some beneficiaries of mobile NP, as explained above, are outside the domain of the mobile operators. Fixed network customers will gain from not having to search for changed numbers, etc, and from lower prices for fixed to mobile calls. However the size of these benefits are not large (see Sections 6.3 and 6.7).

We believe that these differences argue for greater emphasis to be put on each operator bearing its own costs. This is partly because the net flows of costs are likely to be small, and this will make for a more practical (and cheaper) system of operation.

7.5 Ovum's proposals

We suggest that the costs of mobile NP should be allocated according to these proposals:

System set up costs

Each network operator should bear its own costs

Each service provider should bear its own costs

We believe that the set up costs of NP, which are a relatively small proportion of the total costs, should be seen as part of the costs of doing business in a competitive industry. Furthermore this will encourage cost minimisation, and will be symmetrical between the operators. This is also a cheap and practical method of allocating costs.

Customer transfer costs

The donor operator/ service provider should pass on its reasonable costs to the recipient operator/service provider.

The recipient operator/service provider can take a commercial decision as to whether these costs should be passed on the porting customer

This enables the cost causer (the porting customer) to bear some of the cost of porting, and a charge may be desirable to discourage uneconomic porting. This imposes costs on the operators that are much greater than the benefits to the customers. This arrangement, which already exists in the fixed network, is also practical.

However there is scope for the donor network to make unrealistic charges into order to deter porting, and so we propose:

Oftel should publish a benchmark charge for the customer transfer costs.

This could be enforced through Oftel's powers to investigate anti-competitive practice. An aggrieved customer or recipient operator/service provider would complain to Oftel, which would then use its powers to investigate whether the donor operator/service provider was levying unreasonable charges so as to hinder the development of NP and competition. Alternatively, Oftel could make this charge determinable, and an operator/service provider which could not agree with the charge proposed by the donor network would request a determination.

Additional traffic costs

Each network operator should bear its own additional traffic costs.

We believe that this is appropriate because in time the net flows between the networks may not be large. Furthermore the application of this proposal will encourage operators to minimise costs. The proposal is also practical, and its application will be symmetrical.

We have no proposals for the allocation of the customer costs or the caller costs. The operators will have to decide how to allocate the costs of the new handsets or SIM cards in line with their commercial judgements. We see no need to reallocate the caller costs, which by themselves are minimal, and less than the benefits that callers gain from NP.

Although our rationale is somewhat different from that applied to the fixed network, the results of our proposals are the same as the MMC developed for the long term in the fixed network. We summarise our proposals in Figure 7.3.

Figure 7.3: Ovum's proposals for the allocation of costs in the mobile network

Glossary

DCS Digital communications systems

ETSI European Telecommunications Standards Institute

GMSC Gateway mobile switching centre

GSM Global system for mobile

HLR Home Location Register

IMSI International mobile subscriber identity

MMC Monopolies and Mergers Commission

MOU Minutes of use

MSRN Mobile station roaming number

NERA National Economic Research Associates

NP Number portability

NPV Net present value

Oftel Office of Telecommunications

PCN Personal communications network

PSTN Public switched telephone network

SIM Subscriber identity module

SME Small and medium enterprise

SMS Short message service

SMSC Short message service centre

TACS Total access communications system

UMTS Universal mobile telecommunications system