Content uploaded by Justus Haucap
Author content
All content in this area was uploaded by Justus Haucap on Jul 20, 2018
Content may be subject to copyright.
Telecommunications Policy 30 (2006) 385–399
Mobile number portability in Europe
Stefan Buehler
a
, Ralf Dewenter
b
, Justus Haucap
c,
a
Socioeconomic Institute, University of Zurich & ENCORE, Blu
¨mlisalpstr. 10, CH-8006 Zurich, Switzerland
b
Department of Economics, Helmut-Schmidt-University, Holstenhofweg 85, D-22043 Hamburg, Germany
c
Department of Economics, Ruhr-University of Bochum, Universita
¨tsstr. 150, GC3/62, D-44780 Bochum, Germany
Abstract
This paper examines the causes and effects of mobile number portability (MNP) and provides a survey of its
implementation in Europe. It first examines the competitive effects and costs of introducing MNP. Next, it discusses how
to charge for MNP. It argues that a price cap regime starting from the average cost of porting is likely to provide
appropriate incentives. Finally, it reviews recent experience with implementing MNP in Europe. Differences in the speed of
porting and porting charges appear to explain part of the differences in the use of MNP across countries.
r2006 Elsevier Ltd. All rights reserved.
Keywords: Mobile; Number portability; Competition; Customers; Market entry
1. Introduction
Traditionally, consumers of mobile telecommunications services were required to give up their number
when switching providers. Consumers were thus hesitant to switch from an incumbent to competing
operators, thereby inhibiting more effective competition in mobile telecommunications. Lately, the picture has
changed dramatically, as mobile number portability (MNP) has been implemented in the European Union
(EU) and many other countries around the world.
According to the EU’s Universal Service Directive of 7 March 2002, which became effective on 25 July 2003,
MNP means that customers are given the right to keep their mobile telephone number when switching
between mobile telecommunications service providers. Under Article 30 of this Directive ‘‘Member States
shall ensure that all subscribers to publicly available telephone services, including mobile services, who so
request can retain their number(s) independently of the undertaking providing the service.’’
1
Similar
approaches towards introducing MNP have been adopted elsewhere.
From a property rights perspective, the introduction of MNP reallocates the property rights in mobile
telephone numbers from operators to consumers. The main rationale for this re-allocation is the enhancement
ARTICLE IN PRESS
www.elsevierbusinessandmanagement.com/locate/telpol
0308-5961/$ - see front matter r2006 Elsevier Ltd. All rights reserved.
doi:10.1016/j.telpol.2006.04.001
Corresponding author. Fax: +49 234 32 14311.
E-mail addresses: sbuehler@soi.unizh.ch (S. Buehler),ralf.dewenter@hsu-hh.de (R. Dewenter),justus.haucap@rub.de (J. Haucap).
1
Directive 2002/22/EC of the European Parliament and of the Council of 7 March 2002 on universal service and users’ rights relating to
electronic communications networks and services (Universal Service Directive), Official Journal of the European Communities, 24 April
2002, L108/51-77.
of competition in mobile telecommunications. As the EU argues, ‘‘number portability is a key facilitator of
consumer choice and effective competition in a competitive telecommunications environment’’ (see European
Parliament, 2002, p. 57). Accordingly, national regulatory authorities (NRAs) shall ensure that (a) charges for
MNP are cost-oriented and ‘‘that direct charges to subscribers, if any, do not act as a disincentive for the use
of these facilities’’ (Art. 30 (2) Universal Service Directive), and (b) retail charges for MNP do not distort
competition (Art. 30 (3) Universal Service Directive).
This paper presents a non-technical analysis of the causes and effects of MNP and reviews the recent
experience gained with MNP in European countries. It begins with a discussion of the competitive effects and
the costs of introducing MNP (Sections 2 and 3). Section 4 provides an analysis of appropriate charges for
MNP, an aspect that has largely been ignored in previous literature. It then proceeds to evaluate recent
experience with MNP in Europe (Section 5). Finally, a set of conclusions is provided.
2. On the competitive effects of MNP
2.1. The case for MNP
The rationale of introducing mandatory MNP is simple: It is expected to bring about considerable benefits
to consumers of mobile services (see, e.g., Ovum, 2000). Adopting the classification originally proposed by
NERA/Smith (1998), five potential benefits of introducing MNP to consumers of mobile services are discussed
(see Table 1).
With switching costs, customers that actually switch (and thus give up their numbers) incur a utility loss.
Also, switching costs induce some consumers to stick to a provider which is not their preferred choice.
Introducing MNP benefits both types of consumers: Consumers who switch even in the absence of MNP can
retain their number (benefit 1A) and consumers who switch only with MNP are more likely to obtain services
from their preferred operator (benefit 1B).
2
While MNP thus benefits consumers that actually switch, there are
also benefits to non-switching consumers resulting from more intense competition among providers of mobile
telecommunications services (benefit 2).
3
Furthermore, introducing MNP benefits consumers who place calls
to ported numbers (benefit 3). Without MNP, these consumers have to update their address books, may be
unable to call a particular user, etc.
4
Finally, introducing MNP benefits mobile customers because of the
reallocation of property rights (benefit 4). The fact that MNP reallocates property rights in telephone numbers
is especially important for so-called vanity numbers. If customers advertise their telephone number, this
increases the number’s value and may be seen as a specific investment into the number’s value. Hence, a
telephone number’s value is to some extent endogenous.
5
The incomplete contracts literature suggests that
underinvestment results if the customer making the investment does not hold the property right in the
number.
6
Hence, the reallocation of property rights strengthens the customers’ investment incentive. This
thickening of consumers’ property rights benefits all consumers—whether they actually port their number or
not. The option to port one’s number(s) is decisive here and this option is given to every telephone user with
MNP.
The paper now proceeds with a more detailed analysis of the competitive effects of MNP, focusing on the
elimination of switching costs associated with MNP.
ARTICLE IN PRESS
2
In a model with differentiated products, this means that consumers can switch more easily to their preferred provider, so that an
increase in allocative efficiency results (see, e.g., Buehler & Haucap, 2004).
3
See, e.g., Aoki and Small (1999),Galbi (2001), and Gans and King (2001).
4
However, most cost benefit analyses have estimated this effect to be relatively small or almost negligible (see, e.g. Oftel, 1997,Schwarz-
Schilling & Stumpf, 1999).
5
Some numbers that are easy to remember already have a high exogenous value. This is illustrated by the fact that in the Chinese
province of Sichuan the telephone number 8888 8888 obtained a price of 2.33 million yuan ($282,000) in an auction in August 2003. The
reason is that many Chinese people consider the number eight to be lucky because it sounds similar to the Mandarin and Cantonese word
for getting rich. So an eight-digit number containing only the number eight is considered especially auspicious (see BBC, 2003).
6
The classic reference is Grossman and Hart (1986).Haucap (2003) provides an application to telephone numbers.
S. Buehler et al. / Telecommunications Policy 30 (2006) 385–399386
2.2. The effects of eliminating switching costs
Consumers of mobile telecommunications services typically face switching costs which derive from the real
and psychological costs that consumers confront when changing suppliers (see, e.g., Klemperer, 1987a, b,
1988, 1995). These switching costs are endogenous if they emanate from customer loyalty programs (such as
Deutsche Telekom’s so-called Happy Digits program) or contractual clauses that make the change of suppliers
more costly (such as contract termination penalties). There are also exogenous switching costs resulting from
the transaction costs associated with switching providers (e.g. for changing the network assignment of a given
number). Introducing MNP eliminates at least part of these switching costs. In the following, some important
static and dynamic effects of introducing MNP are described.
2.3. Static effects
Previous literature has mainly focused on the static effects of introducing MNP, largely abstracting from
market entry and investment. This section briefly surveys some of the literature’s key findings.
2.3.1. Retail prices
It is well known that, in the presence of switching costs, firms may exploit their market power over captured
customers. For instance, with linear retail prices, an incumbent firm benefits from a wedge driven between its
price and the prices of new entrants, allowing the incumbent to charge a higher price than would otherwise be
possible. To see this, suppose that the customers of an established provider Aface switching costs S40. Firm
Ais thus able to charge a higher retail price than its competitor Bwithout inducing its customers to switch.
That is, the customers of provider Awill switch only if pA4pBþS. Introducing MNP should thus be expected
to reduce linear retail prices, benefiting mobile customers.
With nonlinear retail prices, introducing MNP is still likely to benefit mobile customers (provided they do
not suffer from the so-called ‘‘customer ignorance problem’’
7
). However, the argument is slightly more
complex. Using a simple model with differentiated networks, Buehler and Haucap (2004) show that an
incumbent’s customers benefit from lower fixed fees with MNP, whereas competitors’ customers suffer from
higher fixed fees. Since the beneficial effects on the incumbent’s customers dominate the adverse effects on the
competitors’ customers, the overall effect of MNP on mobile customers is positive.
8
2.3.2. Price elasticities
The above arguments suggest that it is more difficult to gain market share in the presence of switching costs,
as undercutting needs to be more severe. Technically speaking, the firms’ perceived price elasticity of demand
is smaller and equilibrium prices should thus be expected to be higher than with MNP. Moreover, the smaller
price elasticity of demand helps to stabilize collusive arrangements (see Schwarz-Schilling & Stumpf, 1999), as
the extra profits from deviating from collusive behavior will be relatively small.
ARTICLE IN PRESS
Table 1
Benefits of introducing MNP
Type of benefit Applies to Benefit
1A Users who switch even without MNP Avoided costs of number change (e.g., informing users, missed calls)
1B Users who only switch with MNP Benefits of moving to a more preferred operator
2 All users Intensified competition
3 Callers Avoided costs of finding changed numbers
4 All users Increased investment in number value due to re-allocation of property rights
7
We discuss the customer ignorance problem in more detail below (see Section 3.2).
8
A referee correctly pointed out, though, that the overall effects on prices in a more general model are less clear, as MNP reduces the
firms’ possibilities to discriminate between on-net and off-net tariffs.
S. Buehler et al. / Telecommunications Policy 30 (2006) 385–399 387
2.3.3. Termination charges
The effects of introducing MNP on termination charges crucially depend on the occurrence of the
‘‘customer ignorance problem’’ (Gans & King, 2000). If customers can identify the network assignment of
each individual number even after introducing MNP, termination charges should remain unaffected.
However, if customers are no longer able to determine which mobile network they are calling when
placing a call, termination charges are likely to increase. Intuitively, this follows from the firms’ incentives
to increase their charges when customers only take notice of average charges (see Buehler & Haucap,
2004).
2.3.4. Market shares
In the presence of switching costs, the market shares of incumbents and new competitors will typically be
asymmetric. More specifically, incumbents will have large market shares, whereas new competitors will have
smaller market shares. In standard network competition models, introducing MNP will eliminate this
asymmetry as new competitors are no longer forced to offer a discount relative to the incumbents to attract
customers and market shares will thus be aligned. If the providers’ profits are convex in own market share (as,
e.g., in Buehler & Haucap, 2004), introducing MNP will reduce aggregate profits. That is, the incumbents’ loss
is larger than the extra profit awarded to new competitors.
2.4. Dynamic effects
In addition to static effects, introducing MNP will generate dynamic effects. In particular, MNP might
affect entry and investment decisions, as is now briefly discussed.
2.4.1. Entry
It has been noted above that, with switching costs, entrants have to price aggressively to steal business from
the incumbent. Introducing MNP will alleviate the need to price aggressively, thereby facilitating entry.
However, there may be countervailing effects. For instance, if incumbent mobile operators have a large
captured customer base thanks to switching costs, they are less likely to fight entry by aggressively cutting
prices due to the so-called fat-cat effect (see Fudenberg & Tirole, 1984) in the absence of MNP. In contrast,
MNP makes incumbents also more aggressive so that market entry may become less attractive for new
operators. The net effect of introducing MNP on entry is thus ambiguous.
2.4.2. Investment
Introducing MNP is likely to affect the investment incentives of both incumbents and potential competitors.
However, to the best of the authors’ knowledge, there is no systematic analysis of the effects of MNP on the
mobile service providers’ investment incentives. Standard arguments suggest that introducing MNP will
reduce the incumbents’ incentive to make cost-reducing investments, as the cost-reduction applies to a reduced
customer base. Conversely, the competitors’ incentives to make cost-reducing investment should be expected
to increase, with ambiguous net effect. The aggregate effects on demand-enhancing investment, such as
infrastructure quality or product innovation, are even less clear.
2.5. Summary
Overall, the competitive effects of introducing MNP are fairly complex. MNP is likely to affect
retail prices, termination charges, price elasticities, market shares, as well as entry and investment decisions.
So far, it is fair to say that most analyses of MNP have supported the notion that, on the whole,
MNP intensifies competition in mobile telecommunications. Available empirical evidence on the portability
of premium rate numbers appears to support this conclusion (Viard, 2004). Yet, it is unlikely that
introducing MNP reduces all prices. In fact, standard models suggest that handset prices will increase
as the value of a captured customer decreases, whereas prices for mobile services will decrease as
competition intensifies (Buehler & Haucap, 2004). Furthermore, the pro-competitive effects of MNP
ARTICLE IN PRESS
S. Buehler et al. / Telecommunications Policy 30 (2006) 385–399388
are likely to vary across countries, depending on the degree of competition achieved before introducing
MNP.
9
3. The costs of MNP
Obviously, the introduction of MNP generates costs (otherwise, the decision whether or not to introduce
MNP would be trivial). Generally speaking, it is possible to distinguish between direct and indirect costs of
introducing MNP.
3.1. Direct costs
The direct costs of implementing and running an MNP system consist of
costs of developing and implementing a MNP system (set-up costs),
costs per actual porting process, and
additional conveyance costs.
As will be further discussed below, the magnitude and structure of these costs depend heavily on the
technical solution used to implement MNP. The set-up costs are the non-recurring costs of developing and
implementing a particular MNP system. These costs include changeover costs which arise when changing from
an existing to a new system (e.g. when changing from an on-switch (ONS) to an off-switch (IN) solution, see
Smith/Arcome/NERA, 1997, p. 62 f.). More specifically, set-up costs include costs for developing software,
upgrading switch software, installing company procedural and operational methods and developing an
operational support system. These costs are fixed, that is, they do not depend on the number of actual portings
or the traffic routed over a network. In general, these fixed costs are relatively high for so-called IN solutions,
while they are relatively low for ONS solutions.
In addition to fixed set-up costs, there are variable costs associated with the porting process. These are
mainly the costs of carrying out the porting, e.g. advising the customer, communications between the donor
and the receiver network, administrative work related to the number switch and so on. These costs are
essentially personnel costs, and they depend on the specific administrative and technical procedures put in
place.
Finally, there may be conveyance costs, which also depend on the technical solution chosen to implement
MNP. Since simple technical solutions (ONS) lead to an inefficient use of network resources, these costs are
much lower with more advanced IN solutions. With simple solutions, it becomes difficult to determine the
exact costs of a connection and the problem is exacerbated when the number of portings and networks
involved is high. As a result, simple solutions for the implementation of MNP are usually regarded as
inefficient temporary solutions at best (see Smith/Arcome/NERA, 1997, p. 67).
Comparing the direct costs of the various technical solutions, the authors note that ONS solutions are
characterized by comparatively low fixed set-up costs and high variable costs. In contrast, IN solutions have
relatively high fixed set-up costs, whereas their variable costs are low. Hence, IN solutions are cost-efficient if
the expected number of portings is relatively high, whereas less advanced solutions (such as call forwarding)
are efficient as long as the number of portings is low. The stylized average cost functions for these two
technologies, ONS and IN, are depicted below, as is the threshold number of portings after which an IN
solution becomes more efficient than an ONS solution (Fig. 1).
ARTICLE IN PRESS
9
For example, a NERA/Smith study of the costs and benefits of introducing MNP in Hong Kong estimated the additional benefit from
increasing competition to amount to 1 Euro per customer over a period of 10 years, as competition was already quite intense even before
MNP was introduced (see NERA/Smith, 1998). In contrast, Oftel (1997) estimated the additional benefits of increased competition
resulting from MNP to lie around 69 Euros per customer over a 10 year period—quite a significant difference which is due to Oftel’s
assumption that competition would be significantly more intense with MNP than without MNP (see Oftel, 1997).
S. Buehler et al. / Telecommunications Policy 30 (2006) 385–399 389
3.2. Indirect costs of MNP
There are also indirect costs of introducing MNP, resulting from the potential loss in tariff transparency—
an effect that has largely been ignored in the debate on MNP. The loss of tariff transparency results from the
fact that MNP can make it more difficult for consumers to distinguish between different networks when
placing a call. In the absence of MNP, consumers can usually distinguish between different mobile networks
through the number prefix. When MNP is introduced, however, the number prefix no longer indicates the
network assignment of a given number. Thus, if calling prices differ across networks, customers will be
unaware of the exact charges for placing calls to mobile networks. As Ovum (2000, p. 14/15) acknowledges in
its cost–benefit analysis of MNP in Ireland, ‘‘the first three digits of the called number no longer indicate the
network operator of the called subscriber reliably. Full tariff transparency is therefore lost and, unless prices
change, callers end up paying a lot more than expected for certain calls. [yBut] it is difficult to quantify these
effects and we have excluded them from our cost–benefit analysis.’’
In the academic literature, the ‘‘customer ignorance problem’’ has been explored by Gans and King (2000)
and Wright (2002). These authors show that mobile operators may have incentives to increase their
termination charges if consumers only take notice of average prices. Dewenter and Haucap (2005) provide
empirical support for this finding, and Buehler and Haucap (2004) analyze the tradeoffs related to the
introduction of MNP.
However, the loss of tariff transparency may be overcome. In Finland and Germany, for example,
consumers can call a toll-free number to identify a particular number’s network assignment. In Portugal, an
acoustic signal alerts consumers when placing off-net calls. Yet, such mechanisms generate costs on their own,
and they are often considered a nuisance by many consumers.
An alternative method to avoid the customer ignorance problem would be the introduction of the so-called
receiving party pays regime (RPP), as in the United States, in Canada and some Asian countries. Since under
RPP the calling party is charged for the origination but not for the termination of off-net calls, tariff
transparency becomes irrelevant. That is, under RPP, the customer ignorance problem vanishes.
4. Charges for MNP
The paper has argued that there are various costs and benefits of introducing MNP. However, virtually all
cost–benefit analyses have concluded that the overall effect of MNP on welfare is likely to be positive. Since
MNP is not costless, at least two questions remain: Who should pay for MNP? And what is the efficient
charge for porting a number? In the following subsections, we discuss potential answers to these questions.
4.1. Inefficiency of free porting
Economic efficiency typically requires that users of a given resource or service pay for their usage if the
resource under consideration is scarce (i.e. carries an opportunity cost). If users are not made to pay even
ARTICLE IN PRESS
Number of Portin
g
s X
ACON
S
ACIN
Fig. 1. Direct costs of MNP using ONS or IN technology.
S. Buehler et al. / Telecommunications Policy 30 (2006) 385–399390
though their usage causes costs, an inefficient over-use of the resource will result. This problem is also known
as the ‘‘tragedy of the commons’’. More generally, a below-cost usage fee leads to inefficiencies. Making a
resource available free of charge is thus only efficient if it is either not scarce or a public good, where the latter
is characterized by two distinctive features:
(a) it is economically not feasible to exclude people from using a good or service (non-excludability); and
(b) additional users do not cause additional costs (non-rivalry in use).
It should be clear that MNP is not a (pure) public good, as (a) people unwilling to pay for MNP can be
excluded and (b) additional portings cause additional costs, i.e. the incremental cost of an additional porting is
positive.
However, as has been seen, there are various benefits associated with MNP, some of which are private and
some of which are public. While type 1A and 1B benefits are private, type 2 and 4 benefits resulting from the
strengthening of competition and increased investment incentives are public, as they do not only accrue to
those users who actually port their number. In particular, it is not possible to exclude any user from the
competitive benefits that arise from the introduction of MNP. Moreover, type 2 and 4 benefits do not actually
arise from porting a number but from the option to do so. The possibility to port one’s number strengthens
users’ positions vis-a
`-vis their provider, and this possibility is decisive for competition.
10
Existing cost–benefit analyses suggest that type 1A and 1B benefits are at least as large as type 2 benefits.
Furthermore, type 2 benefits are expected to be the smaller the more competitive a particular market was
before introducing MNP. Hence, at least in highly competitive markets the benefits of MNP will be largely
private, while in less competitive markets the benefits are more likely to contain public benefits.
While private benefits and positive incremental costs per porting call for positive charges, there are also
positive externalities to be considered. These are the benefits to potential callers, i.e. type 3 benefits. On the
other hand, there are negative externalities, as MNP can reduce tariff transparency. In our view neither the
positive nor the negative externalities should be overemphasized, as their magnitude has not been estimated to
be significant. From this, it can be concluded that making MNP available free of charge will be inefficient:
Free porting will induce users to port their number even if their number’s valuation is smaller than the
incremental cost of porting. Therefore, an avoidable deadweight loss will result.
Furthermore, MNP can be viewed as a new service offered to consumers. In a competitive market, new
services will be offered if providers expect the total revenues from these services to cover the costs. Hence, even
in a competitive environment, MNP would not be offered free of charge, but at a positive price. Accordingly,
donor and/or recipient networks should be allowed to charge for porting.
4.2. Regulating charges for MNP
The next step is to look for an efficient charge for porting a number. Economic intuition suggests that
market forces are unlikely to lead to efficient charges for captured customers, as the respective operators are
the only providers of MNP for their customers and thus enjoy monopoly power. Note that most customers
have received their mobile telephone number well before MNP has been introduced.
The picture is less sombre for new customers, where the porting charge is just one element of the service
packages offered by the different providers. Here, it is conceivable that providers voluntarily offer MNP at
efficient prices so as to convince consumers to sign a contract.
11
In particular, with nonlinear tariffs, service
providers have an incentive to charge efficient variable charges and extract the costumer surplus using the
fixed fee. In this case, the outcome may be efficient, even though the consumer surplus is redistributed to the
providers. Overall, it seems nevertheless likely that, at least in the short run, market prices for MNP are at
inefficiently high levels, as mobile operators have monopoly power in setting MNP charges for the majority of
ARTICLE IN PRESS
10
Also note that users cannot be excluded from the option to port their number, even if users who do not wish to pay can be excluded
from actually porting the number.
11
This argument is related to Farrell and Gallini (1988) who show that producers may voluntarily commit to keep markets competitive
(by opening up second sources) in order to convince consumers that they will not be exploited at a later stage.
S. Buehler et al. / Telecommunications Policy 30 (2006) 385–399 391
their customers since MNP was not available when most customers’ received their current mobile telephone
number.
12
This suggests that regulating the charges for MNP is desirable (at least for a transition period after
introducing MNP).
For determining the efficient charge for porting a number, it is important to note that MNP has both fixed
and variable costs. In particular, IN solutions are characterized by decreasing average costs. This means that
determining efficient charges for MNP gives rise to similar problems as the determination of efficient charges
in natural monopoly settings. It is well known from standard economic theory that the efficient price for
natural monopoly services is at the marginal or incremental cost level (see, e.g. Viscusi, Vernon, & Harrington,
2000). The resulting deficit should ideally be compensated through government subsidies. Hence, one could
argue that porting charges should be set at incremental or variable cost and the government should cover the
fixed cost associated with the development and implementation of mandatory MNP.
13
One might be tempted to argue that the operators should bear the fixed cost and set prices at incremental
costs in order to guarantee allocative efficiency if the government is not willing to cover the fixed cost.
However, this solution distorts the operators’ choice of technology: If operators are forced to cover the fixed
cost and the technology is not predetermined, operators are likely to choose a technology with relatively low
fixed and high variable costs (such as an ONS solution), which may lead to productive inefficiencies. That is,
providers may deliberately install an inefficient technology with relatively high variable costs, attempting to
recover a larger part of the total cost from MNP. In addition, higher MNP charges imply higher consumer
switching costs. This again allows incumbent providers to charge less aggressively without losing their current
customers. Hence, incumbent providers are likely to set rather high MNP charges in order to avoid losing
customers.
To avoid this problem, regulatory authorities could prescribe the technology to be used. However, this
approach is unlikely to give rise to an efficient outcome, as operators typically have better information about
the efficient use of technologies than regulators. It is well known from principal-agent analysis that the
regulator (the principal) has to reward the operators (the agents) to induce an efficient technology choice in the
presence of asymmetric information. If operators anticipate that they will have to lower their porting charges
if they use an efficient instead of an inefficient technology, an inefficient technology will result. An obvious
way to provide incentives is a price cap regulation for porting charges. As a starting point, one may use current
average cost, but also some other figure exceeding incremental cost (to avoid the over-use problem). To avoid
potential ‘‘gold plating’’ by operators, one may use a ceiling for the average cost, based on the most efficient
technology in use today.
There are essentially two arguments in favor of a price cap regime starting from current average costs. First,
such a regime leads to efficient ‘‘make or buy’’ decisions: If operators are allowed to recover variable costs
only, they may strategically outsource parts of their business in order to substitute variable for fixed costs.
Since an external provider of MNP-related services will charge a price for its services that at least covers
average cost, mobile operators will outsource excessively in order to reduce their fixed cost (if they cannot
recover them through MNP fees). In contrast, under a price cap regime starting from current average costs,
operators will only outsource if an external provider of MNP services can offer these services at lower prices
than the corresponding costs of in-house production. Secondly, starting from current average costs allows
operators to implement MNP without incurring losses. If, instead, operators are forced to offer MNP at prices
below average cost, this will be regarded as a government hold-up or expropriation. Since license fees, network
investments and customer acquisition costs are all specific investments, operators are vulnerable to
expropriation through renegotiation of the regulatory contract (see Goldberg, 1976;Sidak & Spulber, 1999).
In dynamic and innovative industries such as mobile communications, dynamic efficiency aspects are highly
ARTICLE IN PRESS
12
It should thus not come as a surprise that the German regulatory authority (RegTP) felt it had to step in and to regulate charges for
MNP ex post after two small service providers decided to charge their customers a price of 116.00 hfor porting their number (see http://
www.regtp.de/aktuelles/pm/03140/index.html). Most other European countries have regulated MNP charges ex ante.
13
Introducing two-part tariffs for MNP is not very useful, as most consumers have only one mobile number to port (i.e., one unit of
consumption). Alternatively, one might consider a tariff involving a (fixed) fee for buying an option to port one’s number and another
charge when exercising this option (the incremental cost of porting). Note, however, that it is virtually impossible to install different
technologies for different customers on the same network. Hence, in practice, the MNP option can only be installed jointly for all
customers and it is not possible to exclude consumers from this option.
S. Buehler et al. / Telecommunications Policy 30 (2006) 385–399392
relevant. Therefore, regulators should be hesitant to introduce regulations adversely affecting investment and
innovation incentives.
Summing up, it is argued here that the market is unlikely to generate efficient charges for MNP, as service
providers have monopoly power over their captured customers as MNP became available after most
consumers had been allocated a mobile number. Regulating the charges for MNP thus seems desirable as we
have argued above, (at least in a transition period after introducing MNP), provided that the regulated
charges cover not only the variable costs, but also parts of the fixed costs (a) in order to avoid an ex post hold-
up which adversely affects investment incentives and (b) to provide incentives to implement an efficient MNP
technology. A price cap regime starting from the current average cost of the most efficient technology in use
should be expected to provide appropriate incentives.
5. The European experience
As noted at the outset, the EU’s Universal Service Directive requires member states to implement number
portability for publicly available telephone services (including mobile services). The provision of number
portability should be cost oriented and charges should not prevent subscribers from porting their numbers.
The implementation of MNP may be abandoned if it either leads to negligible positive effects for end-users
and competition or is technically non-feasible. The directive furthermore stresses the importance of
transparent tariff information and instructs NRA’s to facilitate tariff transparency when implementing
number portability.
The introduction of MNP and its success in European countries and abroad is now reviewed. In particular,
the actual use of MNP, porting charges, porting speeds and regulatory interventions to foster price
transparency for off-net-calls,
14
is the focus of what follows.
5.1. Introduction of MNP
In 1997, Singapore was the world’s first country to implement (limited)
15
MNP. The United Kingdom and
the Netherlands first implemented MNP in Europe in 1999. Countries such as Spain (2000), Sweden and
Denmark (all 2001), Belgium, Italy, Germany and Portugal (all 2002) followed suit. Most recently, Estonia
implemented MNP due to regulatory intervention. Only a few EU member states have not yet introduced
MNP (see Table 2).
The widespread implementation of MNP in Europe thus took about six years. A number of countries
postponed the implementation of MNP for various reasons. For instance, Germany delayed the introduction
of MNP due to the lack of an adequate technical solution. Also, Austria postponed the introduction of MNP
several times (see Table 3): While smaller Austrian operators such as tele.ring and Tele2 supported MNP, the
larger operators mobilkom and T-Mobile had reservations about MNP.
16
Similar delays occurred in Non-
European countries such as Australia, where MNP started with a delay of no less than 50 months (see Ovum,
2005).
5.2. The use of MNP
Since MNP lowers switching costs, churn rates should be expected to increase, thereby intensifying
competition. Unfortunately, there are no reliable statistics on churn rates available for most European
countries. The paper therefore focuses on the number of portings as a proxy for the success of MNP.
17
Table 4
illustrates that the use of MNP differs widely across countries. In the UK, Spain and Italy, more than 2 million
numbers were ported during the last few years. In countries such as Germany and France, however, relatively
ARTICLE IN PRESS
14
European regulatory authorities were surveyed by the authors in fall 2003 with respect to regulatory frameworks and their first
experiences with MNP. In the following, we present some of this survey’s results.
15
Even though implemented in 1997, MNP has been limited to voice telephony without the ability to support data services such as SMS
and MMS.
16
See http://www.heise.de/newsticker/meldung/45307 (18 June 2004).
17
Note, however, that there may also be an overuse of MNP if porting charges are too low as argued above.
S. Buehler et al. / Telecommunications Policy 30 (2006) 385–399 393
ARTICLE IN PRESS
Table 3
Delays in implementing MNP
Country Original plan Date started Time lag (months) Reason
Austria Aug-03 Oct-04 15 Opposition of large MNOs
Czech Republic 2004 Not known Not known Not known
France Jan-01 Jun-03 30 Not known
Germany 1997 Nov-02 60 Awaiting international standards
Public consultations
Time needed to design solution
Ireland Nov-02 Jul-03 9 Lack of co-operation due to resources being
deployed in the transition to the Euro
Emphasis on full automation and very fast
porting times leading to complex
specifications.
The Netherlands Dec-98 Apr-99 4 Technical issues
UK Jul-98 Jan-99 6 Technical issues
Source:Ovum (2005), own inquiries.
Table 2
Introduction of MNP in Europe
Year Countries
1997 Singapore
1999 UK, Netherlands, Hong Kong
2000 Spain, Switzerland
2001 Sweden, Denmark, Norway, Australia
2002 Belgium, Italy, Germany, Portugal
2003 Finland, Luxembourg, Ireland, France
2004 Greece, Austria, Slovenia, Cyprus, Lithuania, Poland, Hungary, USA, South Korea
2005 Estonia, Latvia (planned) , Malta (31 July 2005)
Not clear Czech Republic, Slovakia, New Zealand, Japan, Mexico
Sources: Own inquiries, EEC (2004).
Table 4
Ported numbers in European countries
Country Period Ported numbers Avg. ported numbers
(monthly)
Percentage of all
subscribers
UK 1/1999–8/2004 3036863 44659.75 5.6
Italy 5/2002–8/2004 2500000 89285.71 4.5
Spain 12/2002–8/2004 2091515 99595.95 5.5
Finland 7/2003
a
–8/2004 993578 76429.07 20.8
Netherlands 4/1999–8/2004 925343 17459.30 6.9
Denmark 7/2001–8/2004 918000 35307.69 17.8
Belgium 10/2002–8/2004 500408 21756.86 6.2
Sweden 9/2001–8/2004 486936 13526.00 5.6
Germany 11/2002–8/2004 349000 15863.63 0.6
Ireland 7/2003
a
–8/2004 142414 10954.92 4.1
Lithuania 1/2004–8/2004 130000 16250.00 n.a.
France 7/2003–8/2004 100000 7142.85 0.2
Portugal 1/2002–8/2004 35032 1094.75 0.4
Hungary 5/2004–8/2004 13875 3468.75 n.a.
Cyprus 7/2004
b
–8/2004 98 65.33 o0.1
Source:EEC (2004), own inquiries (in 2004).
a
25 July 2003.
b
12 July 2004.
S. Buehler et al. / Telecommunications Policy 30 (2006) 385–399394
few subscribers ported their numbers. In terms of percentages of subscribers, Finland, Denmark, and the
Netherlands show the highest numbers.
Since many European countries have introduced MNP only recently, it is difficult to predict future adoption
processes. Furthermore, because of different introductory dates and considerable heterogeneity among the
countries, it is no trivial task to compare the number of portings. Many variables such as contract periods,
competitive environments, and switching costs affect the decision to use MNP services. A crucial factor,
however, should be the price for porting numbers. As charges for porting mobile numbers strongly vary across
countries, price differentials might be an adequate explanation.
5.3. Charges for porting: regulatory frameworks and empirical evidence
The EU’s Universal Service Directive requires that charges for porting mobile numbers are cost oriented
and most member states have thus implemented regulations that should prevent porting charges to exceed
costs. Nevertheless, the extent to which these charges are cost oriented is difficult to assess, since the evaluation
of costs is a daunting task. If mobile operators use IN-solutions with high set-up costs, for example,
uncertainty about the future number of portings could lead to a serious misinterpretation of costs. In
particular, underestimating the number of future portings would lead to charges above costs.
The survey here indicates that in many countries, either the donor network (as, e.g., in Sweden) or the
recipient network (as in Finland, Italy, and Norway) is allowed to charge fees for this service. A typical
situation is that the recipient network charges a fee to the customer’s account and, simultaneously, the donor
network charges a fee to the recipient network’s account. In Finland, for example, operators are allowed to
charge fees to the switching customers and maximum fees are not regulated. However, number porting is
typically free in practice, due to the competitive situation in Finland. Irish and Spanish operators, in contrast,
are not allowed to set charges that can be a disincentive to porting numbers. Table 5 summarizes the various
regulatory frameworks in European countries.
As can be seen from Table 6, there are a number of countries where networks do not charge customers for
porting numbers. For instance, in addition to Finland, MNP is typically free in the UK and in Ireland. In
Belgium, only pre-paid subscribers pay for porting their mobile number. German operators charge between
h22.50 and h29.95, whereas networks charge between h9 and h24 in other countries.
Relating the actual use of MNP (measured by the percentage of ported numbers) to the relevant charges
yields ambiguous results: In some countries with free porting (Ireland, Spain and UK) only a moderate
number of portings occurs, whereas Finland shows a high number (see Fig. 2). Similarly, in countries with
moderate fees the number of portings also varies from rather low rates in Italy to reasonably high numbers in
Denmark (with the Netherlands somewhere in between). Finally, German operators seem to have set nearly
prohibitively high charges. Overall, while other factors appear to affect the success of MNP, the relation
between porting fees and the number of portings (in percent of all numbers) seems to be negative. Higher
charges tend to lead to higher consumer switching costs and therefore to a lower use of MNP.
5.4. Speed of porting
During the porting process, the ported number cannot handle incoming or outgoing calls. Therefore,
increasing the speed of porting is crucial for fostering the use of MNP. Porting time depends both on the
technical porting systems and on the willingness of networks to speed up the porting process. Typically,
neither the donor nor the recipient network have strong incentives to quickly resolve technical problems in
porting, as it is less costly to let the other networks look for a solution.
18
A further obstacle to rapid porting is
the notice period. Subscribers must wait until the notice period in their contract expires before the donor
network releases the number.
19
ARTICLE IN PRESS
18
In particular, the donor network has little to gain from speeding up the process.
19
In Germany, for example, subscribers have to inform the donor network about their intention to port a mobile number about 2 weeks
before the contract expires.
S. Buehler et al. / Telecommunications Policy 30 (2006) 385–399 395
ARTICLE IN PRESS
Table 5
Regulation of one-time charges for porting mobile numbers
Country Permission of charging fees? Regulation of maximum fee? Basis
Austria YES Max. h19 —
Belgium YES—Only recipient network is allowed
to charge fees.
Max. h15 Cost oriented
Denmark YES—Customers pay typically the fee
that operators charge to their
competitors.
NO Cost oriented
Germany YES A unique fee based on the costs caused by
porting numbers.
Cost oriented
Finland YES—Only recipient network is allowed
to charge fees.
NO Cost oriented
Greece YES — —
Ireland YES—Charges shall not be an
disincentive for users to port their
number
Charges shall not be an disincentive for
users to port their number
Cost oriented
Italy YES—Only recipient network is allowed
to charge fees.
Max. h10.02 —
Netherlands YES—Only recipient network is allowed
to charge fees. Charge shall not exceed
administrative costs.
Max. h9—
Norway YES—Only recipient network is allowed
to charge fees.
NO Charge between
networks should
cover costs of donor
network.
Portugal YES NO —
Sweden YES—Only donor network is allowed to
charge fees.
NO (to customer), YES (donor to
recipient)
Cost oriented
(administrative and
porting costs)
Switzerland YES NO —
Spain YES Charges shall not be an disincentive for
users to port their number.
—
UK YES ‘‘adequate fees’’ Marginal costs
Source: Own inquiries (in 2003/2004).
Table 6
Charges for porting mobile numbers
Country Fee
Austria Recipient network charges h4–15.
Belgium Only pre-paid but not post-paid customers are charged for porting mobile numbers.
Denmark Operators committed to charge h9.60 per ported number to customers. The donor operator charges the same
amount to the new operator.
Germany O2 charges h22.50 and T-Mobile,Vodafone and E-Plus charge h24.95 to their customers. Some small service
providers charge h29.95.
Finland The donor network charges about h5–10 to the recipient operator. No fees for customers.
Ireland No fees for Customers.
Italy The donor operator charges h10.02 to the recipient operator. No fees for Customers.
Netherlands The recipient operator is allowed to charge the customer h9. Charges consist of administrative fees.
Sweden Only donor operators charge h4–24 fees to the recipient operator.
Spain No fees.
UK Typically no fees. Some operators charge £25.
Source: Own inquiries (in 2003/2004).
S. Buehler et al. / Telecommunications Policy 30 (2006) 385–399396
Table 7 gives the porting speed for various countries (for times within the period when porting is possible).
As one can see the speed of porting is quite heterogeneous across countries. While in some countries porting
time is extremely short—porting takes only two and a half hours in the US—operators from other countries
may need days, weeks or even months to port a number (see Ovum, 2005). Porting speeds have also varied
considerably over time. While many countries have partly used manually operated porting processes in the
beginning of MNP, nowadays most countries have installed automated porting systems. Although, porting
time has been reduced considerably, there are still obstacles to rapid porting. In the UK, for instance, porting
processes are donor led. Since the donor network has little incentive to speed up porting, this is likely to delay
porting. In Germany, subscribers only have a relatively small timeframe of 5 months in which they can
actually port their number (up to 4 months before their contract expires and until 1 month after expiry).
Within this timeframe, portings usually take 6 days. In the Netherlands, roughly 60% of porting requests are
not successful because of invalid customer information (see Ovum, 2005).
5.5. Tariff transparency
As noted above, once MNP is implemented, the prefix of a mobile number no longer indicates the network
of a subscriber and consumers may thus be unaware of exact charges for calls.
20
As a consequence,
ARTICLE IN PRESS
Spain
Ireland UK
Germany
Italy Netherlands Denmark
Finland
0
5
10
15
20
25
30
0 5 10 15 20 25
Percentage of Ported Numbers
Charge
Fig. 2. Comparison of charges and portings.
Table 7
Speed of porting
Country Speed of porting Porting process
Australia 3 h, 2005 Via recipient
Austria 3 days Via recipient
Germany 6 days, 2005 Via recipient
Hong Kong 36 h, 2005 Via recipient
48 h, 2004
Italy 15 days, 2003 —
Ireland 2 h, 2005 Via recipient
Netherlands 4 days, 2005 Via recipient
3–12 weeks, 2004
2 month, 1999
UK 5–7 days, 2005 Via donor
1 month, 1999
USA 21
2h, 2005 —
Sources:Ovum, 2005, Syniverse Technologies, 2004 and Wind, 2003.
20
Recall that this problem arises only when the so-called calling-party-pays regime applies as in most European countries.
S. Buehler et al. / Telecommunications Policy 30 (2006) 385–399 397
introducing MNP reduces tariff transparency. For this reason, the EU’s Universal Service Directive stresses
the importance of transparent tariff information and instructs NRA’s to facilitate tariff transparency when
implementing number portability.
There are different options to increase tariff transparency in the presence of MNP. Customers may be
informed by enquiry numbers or SMS services to learn about the network of a given number. Acoustic signals
may alert subscribers when placing off-net calls, or verbal announcements could inform about tariffs when
calling to different networks. Table 8 summarizes the methods used in European countries. In Finland,
consumers can call a toll-free number to learn about subscribers’ network association. In Germany, toll-free
numbers and a toll-free SMS service is available. In Portugal, Ireland and Belgium, consumers are informed
by an acoustic signal when placing off-net calls. Consequently, users are informed that they are placing an off-
net call, but they do not learn the price of the call. Obviously, a better system would inform subscribers about
prices when placing their calls.
6. Summary and conclusions
The paper has argued that the rationale of introducing MNP is simple: It is expected to eliminate or at least
reduce the costs of switching providers, making mobile telecommunications more competitive and thus
bringing about considerable gains to consumers of mobile services. Even though there are both direct and
indirect costs of introducing MNP (including reduced tariff transparency), virtually all cost–benefit studies
have concluded that the overall effect of MNP is positive. Interestingly, none of these studies has focused on
the question of how to best charge for porting a mobile number. The authors have argued that a price cap
regime starting from the current average cost of porting a number provides appropriate incentives for
operators.
This review of the experience with MNP in Europe indicates that the regulatory frameworks for MNP and
its actual use vary considerably across the continent. In countries such as Finland or Denmark, a large number
of subscribers have already ported their mobile number, whereas in other countries, including Germany,
France and Italy, relatively few subscribers have ported their number. To some extent, this may be explained
by the differences in (i) charges for porting numbers, and (ii) the speed of porting.
Acknowledgements
The authors thank two anonymous referees, Bernhard Albrecht, Jo
¨rn Kruse, Scott Minehane, Alexander
Zuser and participants at the 15th Biennial Conference of the International Telecommunications Society (ITS)
at Berlin 2004 for most helpful comments and suggestions. They are also grateful to the numerous individuals
and regulatory authorities who have responded to their survey on mobile number portability. Sandra
Grunewald, Tobias Hartwich and Jutta Kehrer provided valuable research assistance.
ARTICLE IN PRESS
Table 8
Methods of carrier identification.
Country Method
Austria Verbal announcement
Belgium Acoustic signal when placing off-net calls
Finland Toll-free enquiry numbers
Germany Toll-free enquiry numbers and toll-free SMS Service
Ireland Acoustic signal when placing off-net calls
Portugal Acoustic signal when placing off-net calls
Sources: Own inquiries (in 2003/2004).
S. Buehler et al. / Telecommunications Policy 30 (2006) 385–399398
References
Aoki, R., & Small, J. (1999). The economics of number portability:Switching costs and two-part tariffs. Working Paper, University of
Auckland, November 1999.
BBC. (2003). China’s Lucky Phone Number, online at: http://news.bbc.co.uk/1/hi/world/asia-pacific/3163951.stm
Buehler, S., & Haucap, J. (2004). Mobile number portability. Journal of Industry, Competition and Trade,4, 223–238.
Dewenter, R., & Haucap, J. (2005). The effects of regulating mobile termination rates for asymmetric networks. European Journal of Law
and Economics,20, 185–197.
EEC. (2004). European electronic communications regulations and markets 2004, 10th report Annex 3, Commission Staff Working Paper,
Brussels 2 December 2004.
European Parliament. (2002). Directive 2002/22/EC of the European Parliament and of the Council of 7 March 2002 on universal service
and users’ rights to electronic communications networks and services.
Farrell, J., & Gallini, N. (1988). Second-sourcing as commitment: Monopoly incentives to attract competition. Quarterly Journal of
Economics,103, 673–694.
Fudenberg, D., & Tirole, J. (1984). The fat-cat effect, the puppy-dog ploy and the lean and hungry look. American Economic Review,74,
361–368.
Galbi, D. A. (2001). Regulating prices for shifting between service providers. Information Economics and Policy,13, 181–198.
Gans, J. S., & King, S. P. (2000). Mobile network competition, customer ignorance and fixed-to-mobile call prices. Information Economics
and Policy,12, 301–328.
Gans, J. S., & King, S. P. (2001). Regulating endogenous customer switching costs, Contributions to Theoretical Economics 1(1), Article 1.
http://www.bepress.com/bejte/contributions/vol1/iss1/art1
Goldberg, V. P. (1976). Regulation and administered contracts. Bell Journal of Economics,7, 426–448.
Grossman, S. J., & Hart, O. (1986). The cost and benefits of ownership: A theory of vertical and lateral integration. Journal of Political
Economy,94, 691–719.
Haucap, J. (2003). Telephone number allocation: A property rights approach. European Journal of Law and Economics,15, 91–109.
Klemperer, P. (1987a). Markets with consumer switching costs. Quarterly Journal of Economics,102, 375–394.
Klemperer, P. (1987b). The competitiveness of markets with switching costs. RAND Journal of Economics,18, 139–150.
Klemperer, P. (1988). Welfare effects of entry into markets with switching costs. Journal of Industrial Economics,37, 159–165.
Klemperer, P. (1995). Competition when consumers have switching costs: An overview with applications to industrial organization,
macroeconomics, and international trade. Review of Economic Studies,62, 515–539.
NERA/Smith. (1998). Feasibility study and cost benefit analysis of number portability for mobile services in Hong Kong. Final report OFTA,
NERA/Smith, London.
Oftel. (1997). Economic evaluation of number portability in the UK mobile telephony market. London: Oftel.
Ovum. (2000). Mobile numbering and number portability in Ireland, A report to the ODTR. London: Ovum.
Ovum. (2005). Mobile number portability—An international benchmark, A report to MTN, Ovum Project CLM42. London: Ovum.
Schwarz-Schilling, C., & Stumpf, U. (1999). Netzbetreiberportabilita
¨t im Mobilfunk: Auswirkungen auf Wettbewerb und
Verbraucherinteressen. Discussion paper no. 199, WIK: Bad Honnef.
Sidak, J. G., & Spulber, D. F. (1999). Deregulatory takings and the regulatory contract: The competitive transformation of network industries
in the United States. Cambridge, MA: MIT Press.
Smith/Arcome/NERA. (1997). Technical options and costs for achieving number portability. Final report for the European Commission,
Brussels.
Syniverse Technologies. (2004). A global perspective on number portability. Syniverse Technologies Inc.
Viard, V. B. (2004). Do switching costs make markets more or less competitive? The case of 800-number portability. RAND Journal of
Economics, forthcoming.
Viscusi, K., Vernon, J., & Harrington, J. E. (2000). Economics of regulation and antitrust (3rd ed). Cambridge, MA: MIT-Press.
Wind. (2003). Mobile number portability, WIND telecommunications—Media relations.
Wright, J. (2002). Access pricing under competition: An application to cellular networks. Journal of Industrial Economics,50, 289–315.
ARTICLE IN PRESS
S. Buehler et al. / Telecommunications Policy 30 (2006) 385–399 399