Addressing the high price of International Mobile Roaming (IMR) in Southeast Asia

Page created by Janet Ayala
 
CONTINUE READING
Addressing the high price of International Mobile Roaming (IMR) in Southeast Asia
Addressing the high price of International Mobile Roaming
                     (IMR) in Southeast Asia
                     J. Scott Marcus, Robert G. Clarke, and Georgios Petropoulos1

                                                   Abstract
International Mobile Roaming (IMR) is a service whereby a user who subscribes to mobile
telecommunications services in one country is able to use his or her mobile device in other
countries. For many years, governments around the world have expressed concerns that the price
of IMR services seemed to be unreasonably high in comparison with the price of domestic
telecommunications services, with demonstrable negative impact on societal welfare. These
concerns have led to numerous regulatory proceedings, and to wholesale and retail IMR price
controls in the European Union and also among the six countries of the Gulf Cooperation
Council (GCC).2 Other regions have implemented various arrangements that seek to lower the
cost of IMR, but in most cases it is too early to assess the results of those arrangements.
Our focus in this paper is on the ten Southeast Asian countries that comprise the regional
organisation ASEAN:3 Brunei, Cambodia, Indonesia, Laos, Malaysia, Myanmar, the Philippines,
Singapore, Thailand, and Vietnam. In the ASEAN region, high prices for roaming have been a
concern for many years; however, concrete measures to address high prices have been rather
limited.
This paper seeks (1) to provide guidance to regulatory authorities and network operators in the
ASEAN region based on experience in other regions, especially Europe and the GCC; and (2) to
provide context for these discussions based both on economic theory and on global experience.

1
 Authors’ addresses: J. Scott Marcus and Georgios, both Bruegel, Rue de la Charité 33, 1210 Saint-Josse-ten-
Noode, Belgium, scott.marcus@bruegel.org and georgios.petropoulos@bruegel.org. Robert G. Clarke, Clarke
Mosby Mehta, robert@clarkemosby.com.
2
  The European Union issued Regulations in 2007, 2009, 2012, and most recently in 2015. The Regulations
approved by the ministers of Gulf Cooperation Council (GCC) countries (Bahrain, Kuwait, Oman, Qatar, Saudi
Arabia, and the UAE) in 2015 implement wholesale and retails controls that are broadly similar to those in effect in
Europe in 2012, as explained in GCC Roaming Working Group, J. Scott Marcus, Christin-Isabel Gries and Robert
Clarke (2015), “Consultation and Report on International Mobile Roaming (IMR) across the Gulf Corporation
Council (GCC) Region” prepared for the Gulf Corporation Council (GCC) Roaming Working Group, available at
http://www.cra.gov.qa/sites/default/files/Final%20Consultation%20Report%20%20-
%20International%20Mobile%20Roaming%20%28IMR%29%2020150621.pdf.
3
  ASEAN, or the Association of Southeast Asian Nations, is a regional organisation comprising ten Southeast Asian
states which promotes intergovernmental cooperation and facilitates economic integration amongst its members.

                       Electronic copy available at: https://ssrn.com/abstract=3529908
Addressing the high price of International Mobile Roaming (IMR) in Southeast Asia
Addressing the high price of IMR for Southeast Asia (the ASEAN region)

Contents
1.     Introduction ............................................................................................................................. 4
     1.1      The overall policy context ................................................................................................ 4
     1.2      Rationale for a taking a public policy intervention .......................................................... 5
     1.3      Measures taken in ASEAN to date ................................................................................... 6
     1.4      Global and regional developments ................................................................................... 9
     1.5      Structure of this paper ...................................................................................................... 9
2.     IMR data flows and payment flows ....................................................................................... 10
3.     Insights from the literature review......................................................................................... 13
     3.1      General policy assessments of International Mobile Roaming (IMR) ........................... 13
     3.2      High prices can be expected with IMR, and are challenging to remedy ........................ 14
     3.3      Wholesale payments influence the retail price ............................................................... 16
     3.4      Societal welfare implications ......................................................................................... 17
4.     Learning from experience in other regions............................................................................ 18
     4.1      National, bilateral, multi-lateral, regional, or global? .................................................... 19
     4.2      Price cap levels ............................................................................................................... 21
     4.3      Staggering wholesale versus retail price caps ................................................................ 22
     4.4      Collection of statistics .................................................................................................... 22
     4.5      Bill shock prevention ..................................................................................................... 25
     4.6      General Agreement on Trade in Services (GATS) considerations ................................ 26
     4.7      Price caps, structural solutions, Roam Like a Local, Roam Like at Home (RLAH) ..... 27
5.     Concluding observations ....................................................................................................... 29
References ..................................................................................................................................... 31

Figures
Figure 1. Charges for various IMR services among various ASEAN MNOs (in euros as of 2013).
......................................................................................................................................................... 7
Figure 2: Cash flows in IMR: calls made ..................................................................................... 10
Figure 3. Cash flows in IMR: calls received................................................................................. 12
Figure 4. Messages sent via mobile handsets by service type, worldwide, 2010–2018. .............. 18
Figure 5. Average price per minute for wholesale non-group roaming voice calls (in euro, EEA).
....................................................................................................................................................... 23
Figure 6. EEA average retail price per minute for intra-EEA roaming voice calls made (in euro).
....................................................................................................................................................... 24

                                                                                                                                   Page 2

                               Electronic copy available at: https://ssrn.com/abstract=3529908
Addressing the high price of International Mobile Roaming (IMR) in Southeast Asia
Addressing the high price of IMR for Southeast Asia (the ASEAN region)

Figure 7. Prepaid versus postpaid subscribers among ASEAN countries. ................................... 26

Tables
Table 1. Availability of Roam Like at Home (RLAH) in the region. ............................................. 8

                                                                                              Page 3

                      Electronic copy available at: https://ssrn.com/abstract=3529908
Addressing the high price of International Mobile Roaming (IMR) in Southeast Asia
Addressing the high price of IMR for Southeast Asia (the ASEAN region)

    1. Introduction
International Mobile Roaming (IMR) is a service whereby a user who subscribes to mobile
telecommunications services in one country is able to use his or her mobile device in other
countries. For many years, governments around the world have expressed concerns that the price
of IMR services seemed to be unreasonably high in comparison with the price of domestic
telecommunications services. These concerns have led to numerous regulatory proceedings, and
to wholesale and retail IMR price controls in the European Union and also among the six
countries of the Gulf Cooperation Council (GCC).4 Other regions have implemented various
arrangements that seek to lower the cost of IMR, but in most cases it is too early to assess the
results of those arrangements.
Among the ten Southeast Asian countries that comprise the regional organisation ASEAN
(namely Brunei, Cambodia, Indonesia, Laos, Malaysia, Myanmar, the Philippines, Singapore,
Thailand, and Vietnam),5 roaming has been a concern for many years, as we explain shortly;
however, concrete measures to address high prices have been rather limited.
This paper seeks (1) to provide guidance to ASEAN regulatory authorities and network operators
based on experience in other regions, especially Europe and the GCC; and (2) to provide a
broader policy context for these discussions.

    1.1      The overall policy context
Throughout the world, roaming prices tend to be far in excess of cost in the absence of
regulation. The economic drivers for these high prices are explained in Section 3. At the risk of
over-simplifying, the high prices are difficult to remedy because they result from rational profit-
taking by two different networks in two different countries – the incentives of the Home Network
(HN) and the Visited Network (VN) tend not to be aligned, and they are moreover regulated by
different national regulatory authorities whose interests may also not be aligned.
Policymakers have sought to bring roaming prices down through various voluntary and
mandatory mechanisms. For reasons just noted, many of the voluntary approaches have had little
or no effect;6 however, the Malaysia / Singapore and Singapore / Brunei agreements seem to
have been somewhat effective (see Section 1.3).

4
  The European Union issued Regulations in 2007, 2009, 2012, and most recently in 2015. The Regulations
approved by the ministers of Gulf Cooperation Council (GCC) countries (Bahrain, Kuwait, Oman, Qatar, Saudi
Arabia, and the UAE) in 2015 implement wholesale and retails controls that are broadly similar to those in effect in
Europe in 2012, as explained in GCC Roaming Working Group, J. Scott Marcus, Christin-Isabel Gries and Robert
Clarke (2015), “Consultation and Report on International Mobile Roaming (IMR) across the Gulf Corporation
Council (GCC) Region” prepared for the Gulf Corporation Council (GCC) Roaming Working Group, available at
http://www.cra.gov.qa/sites/default/files/Final%20Consultation%20Report%20%20-
%20International%20Mobile%20Roaming%20%28IMR%29%2020150621.pdf.
5
  ASEAN, or the Association of Southeast Asian Nations, is a regional organisation comprising ten Southeast Asian
states which promotes intergovernmental cooperation and facilitates economic integration amongst its members.
6
  For example, the Russian government prepared a memorandum of understanding for signature by itself, its
operators, and governments and operators from foreign countries who were willing to cooperate. Upon signing the
memorandum, operators indicate their agreement that IOTs between Russia and the country concerned should be
reduced, and commit to negotiating such reductions, without any specific price points in mind. Russia and its
operators have signed such a memorandum with their Finnish counterparts. A similar agreement was put in place in
2011 between Russia and Poland. The agreements have reportedly had no effect whatsoever on the price of IMR

                                                                                                  Page 4

                       Electronic copy available at: https://ssrn.com/abstract=3529908
Addressing the high price of International Mobile Roaming (IMR) in Southeast Asia
Addressing the high price of IMR for Southeast Asia (the ASEAN region)

Substitutes to IMR have always existed, but they have never offered the simplicity, ubiquity or
convenience of conventional IMR. Use of Wi-Fi when traveling is perhaps the most important of
these, but it tends to be limited to hotels, restaurants, and transportation centres. Acquisition of a
SIM in the Visited Country (“plastic roaming”) continues to be common in developing countries,
but the observation from EU experience that the use of plastic roaming disappears as soon as
roaming rates drop to acceptable levels suggests that consumers consider it to be an inferior
solution. The emerging use of electronic SIMs (eSIM) may make a variant of plastic roaming
more attractive, but we assume once again that they will prove to be an imperfect substitute for
IMR.

    1.2     Rationale for a taking a public policy intervention
There are two primary policy rationales for taking action to reduce roaming prices to levels more
nearly in line with underlying costs:
    •   The over-pricing of IMR services relative both to domestic prices and to real cost
        depresses consumption with negative impact on societal welfare (see Section 3). There is
        a classical economic rationale that reducing artificially inflated prices to cost-based levels
        not only transfers welfare from producers to consumers, but also promotes more usage
        thanks to the price elasticity of demand (PED) (the tendency of consumers to buy more
        of a good or service when the price drops), thereby reducing societal economic
        inefficiency in the form of deadweight loss.
    •   The achievement of low and uniform IMR prices arguably strengthens regional cohesion
        and integration. Regional cohesion can be of interest in regions of the world comprised
        of countries that seek to promote intergovernmental cooperation and to facilitate their
        economic integration, as is the case with the EU, the six GCC countries, and the ten
        ASEAN countries.
Correcting the over-pricing of international mobile roaming can be viewed as a public good.
Public goods benefit the public generally, but are available to all such that the consumption by
one individual does not necessarily reduce the availability for other individuals. Public goods
benefit society as a whole, but it is often the case that no single private organisation is motivated
to provide them on a commercial basis. Because they are typically non-excludable and non-
rivalrous (i.e. available to all who want them without limitation), it is usually impractical for
private entities to fully monetise them. In consequence, it is usually necessary for government to
play some role in ensuring that public goods are made available. (Arce M & Sandler, 2002)
The geographic scope of a public good could be national, regional, or global. (Arce M &
Sandler, 2002) There is good reason to consider the availability of affordable international
mobile roaming as a regional public good.
No single national regulatory authority can remedy over-pricing for IMR. Because of the
interdependency between wholesale payments and retail payments in two different countries (see
Section 2), concerted action is needed in both the Visited Country and the Home Country. This
implies that addressing over-pricing of IMR is not simply a national public good.

between Russia and either Poland or Finland. Tony Shortall (2013), International Mobile Roaming Agreements,
OECD DSTI/ICCP/CISP(2012)2/FINAL, 3 June 2013.

                                                                                          Page 5

                     Electronic copy available at: https://ssrn.com/abstract=3529908
Addressing the high price of IMR for Southeast Asia (the ASEAN region)

At the same time, it is not helpful to view IMR as a global public good. Reducing the price of
IMR is indeed a global concern, but the prospects of a global solution in the near term seem low.
This implies that measures to address the challenges of IMR over-pricing are best understood as
a regional public good. Given that multi-national action is needed, but global action is unlikely,
it is clear that the greatest prospects for addressing the problem lie in regional groupings of
countries, and this has indeed been the experience. Most notably, as we explain subsequently:
      •    A comprehensive regulatory solution to high IMR prices has been in place since 2007 in
           the European Economic Area (EEA), which comprises the European Union (EU) plus
           Norway, Liechtenstein and Iceland, and has been progressively strengthened.7
           Arrangements to date have performed well; however, some experts have concerns about
           the impact in the medium term of Roam Like at Home (RLAH) provisions that are came
           into effect in 2017. (Marcus and Petropoulos, 2016)
      •    A regulatory solution has been in place since 2012 in the Gulf Cooperation Council
           (GCC) region, comprised of Bahrain, Kuwait, Oman, Qatar, Saudi Arabia, and the UAE,
           and was substantially expanded and strengthened in January and April of 2016. (GCC
           Roaming Working Group et al., 2015) Regulatory authorities in the region inform us that
           the new regulation is performing well.
      •    The countries of the East Africa Community region (EACO) agreed to take measures to
           reduce the price of IMR, but no mandatory framework was put in place, and actual
           implementation by network operators has been very uneven. Some IMR prices have been
           substantially reduced, while others remain high. (ITU-D, 2016)
      •    In a few instances (notably Singapore / Malaysia and Singapore / Brunei), bilateral
           agreements have had discernible positive effect.

      1.3      Measures taken in ASEAN to date
The concern that IMR prices are higher in ASEAN than in Europe, where regulation of IMR has
been effective since 2007, is long-standing. Current data is limited, but a number of assessments
were undertaken in the recent past. The price of all forms of IMR in ASEAN is clearly higher in
general than in Europe (where prices are regulated), as is evident in Figure 1; at the same time,
the variation from one ASEAN country to the next, and from one Mobile Network Operator
(MNO) to the next, has been considerable.

7
    The European Union issued Regulations in 2007, 2009, 2012, and most recently in 2015.

                                                                                            Page 6

                         Electronic copy available at: https://ssrn.com/abstract=3529908
Addressing the high price of IMR for Southeast Asia (the ASEAN region)

Figure 1. Charges for various IMR services among various ASEAN MNOs (in euros as of 2013).

Source: Bain & Company and CIMB Asian Research Institute (2013): “Connectivity: Lifting the
Barriers Roundtables”8

Prices have continued to evolve in the region. ASEAN itself reports that “about half of all the
mobile telecommunication operators (“Operators”) in ASEAN offer their subscribers
international mobile data roaming service on a daily flat-rate basis”. (ASEAN (2017))
Roam like at Home offers (or RLAH, where the roamer pays the same price for mobile service in
selected countries is the same as that which he or she would pay at home) have also increased in
the region. We infer that these are premium offers targeting frequent business travellers. A
forthcoming OECD report provides the following summary.

8
 Base on MNO web sites. They note that charges are for usage within or among ASEAN countries; that charges
vary based on call or usage destination; and that charges exclude VAT.

                                                                                         Page 7

                     Electronic copy available at: https://ssrn.com/abstract=3529908
Addressing the high price of IMR for Southeast Asia (the ASEAN region)

Table 1. Availability of Roam Like at Home (RLAH) in the region.
         Country     Operator   Destination in SEA                      Details
         Cambodia    Metfone    Laos, Myanmar and Vietnam           Applies only when connected to the
                                                                         network of Viettel or its subsidiary in
                                                                         each country
         Laos        Unitel     Cambodia, Myanmar and               Applies only when connected to the
                                Vietnam                                  network of Viettel or its subsidiary in
                                                                         each country
         Malaysia    Digi       Brunei, Indonesia, Myanmar,         Free with Digi postpaid 110.
                                Philippines, Singapore, Thailand.   60min Call, 5GB/month.
                                                                    Applies only when connected to preferred
                                                                         carrier in destination
                     U-mobile   Cambodia, Indonesia, Laos,          Free with certain post-paid plans
                                Philippines, Singapore, and         Up to local data quota
                                Thailand.                           Applies only when connected to preferred
                                                                         carrier in destination
         Myanmar     Mytel      Cambodia, Laos and Vietnam          Applies only when connected to the
                                                                         network of Viettel or its subsidiary in
                                                                         each country
         Singapore   Starhub    Malaysia                            Prepaid Customer Only
         Vietnam     Viettel    Cambodia, Laos and Myanmar          Applies only when connected to the
                                                                         network of a subsidiary of Viettel in
                                                                         each country
Source: Mobile Network Operator websites, as assessed in Porciuncula (2019, forthcoming)

Among pricing developments in the region, the mobile “local access” solutions that have been
adopted by some Chinese operators in cooperation with operators based in Hong Kong and
Macau are also worth noting. Under these solutions, a Visited Network offers a “local access”
service to the foreign Home Network, which the Home Network can then resell as an ‘add-on’ to
its customers. The customer then has access during his or her journey to a local number as well
as to his or her home country number. (3 Hong Kong (2019))
International mobile roaming has been a long-standing topic of study for policymakers in the
ASEAN region. The Singapore IDA’s 2009 study called for which called for ASEAN regulators
to implement collective measures to lower international mobile roaming charges. In 2012,
ASEAN and the EU also jointly organised a workshop to share experiences on how to lower
intra-ASEAN roaming charges. (IDA Singapore (2009))
As a Singapore proposal to the ASEAN Telecommunication Regulators' Council (ATRC) and to
APEC Telmin (regional intergovernmental associations for ASEAN and for the Asia-Pacific
region, respectively) notes, “There has been some minor progress as can be seen from the
bilateral mobile roaming arrangements between Singapore and Malaysia, and between Singapore
and Brunei.9 But more could be done.”10

9
 According to the Singtel web site, for instance, incoming roaming calls in Malaysia or Brunei or outbound roaming
calls within the Visited Country cost $ 0.35 SGD, while calls home to Singapore from those two countries cost only
$ 0.46 SGD. No other roaming call costs less than $ 1.50 SGD. http://info.singtel.com/personal/phones-
plans/international-calling/travelling-overseas/roaming-rates viewed 18 April 2016. See also Tony Shortall (2013),
“International Mobile Roaming Agreements”, OECD, DSTI/ICCP/CISP(2012)2/FINAL, pages 20-21: “From 1 May
2011, Singapore and Malaysia mobile phone subscribers saw price reductions of up to 30% for voice calls and 50%

                                                                                                        Page 8

                      Electronic copy available at: https://ssrn.com/abstract=3529908
Addressing the high price of IMR for Southeast Asia (the ASEAN region)

Most recently, ASEAN adopted a Framework for International Mobile Roaming (AFIMR) in
2017. (ASEAN (2017)) This demonstrates continuing interest in the region, and also
demonstrates a reluctance to prematurely adopt an overly regulatory response. The measure is
likely to enhance transparency for roaming prices in the region, but our assessment is that it is
unlikely to have much impact on the level of IMR pricing (see Section 4.1).

     1.4    Global and regional developments
Regulation of IMR at regional level is either under study or under way in multiple regional
groupings around the world. The GCC region has had regulation in place for IMR calls made
since 2013, and is (as previously noted) in an advanced stage of implementation of a new,
comprehensive roaming regulation. Initiatives are ongoing at various stages in the Southern
Africa Development Community (SADC), the East African Community (EACO), in Latin
America, and in the West Balkans.
The International Telecommunications Union (ITU) issued Recommendation D.98 on IMR in
2012, (ITU, 2012) a technical paper on analysis of the cost roaming, (ITU, 2015), a set of
strategic guidelines, (ITU, 2017a) and more. These ITU activities may have generated interest
around the world, but they do not appear to have led to much in the way of concrete actions.
The OECD also issued a Recommendation on IMR in 2012, (OECD (2018)), and several reports
over the years. As summarised in Bourassa et al. (2016), the OECD Recommendation focuses
on:
     •   Promoting transparent information on roaming services
     •   Promoting awareness of roaming prices and substitutes
     •   Facilitating trans-national networks and alliances
     •   Transparency of inter operator tariffs
     •   Facilitating access to wholesale mobile services on local terms and conditions
     •   Wholesale price regulation
     •   Retail price regulation
     •   Assessment of costs and benefits

     1.5    Structure of this paper
Section 2 provides background on IMR data flows and payment flows, while Section 3 provides
insights from the relevant literature. Section 4 discusses how experience in the European Union
and European Economic Area, and also in the Gulf Cooperation Council (GCC) region, might
help address the specific challenges associated with IMR in the ASEAN region. Finally, Section
5 closes with brief concluding observations.

for SMS, when they use the mobile roaming service provided by all mobile operators in Malaysia and Singapore
respectively.”
10
   Ibid. It also notes that 17th ATRC adopted the Addendum on Intra-ASEAN Mobile Roaming Rates Roaming
Rates (MRR) to the Record of Intent (ROI) in 2011, which encouraged member states (AMS) to work together,
bilaterally or multilaterally, to lower international mobile roaming rates.

                                                                                           Page 9

                     Electronic copy available at: https://ssrn.com/abstract=3529908
Addressing the high price of IMR for Southeast Asia (the ASEAN region)

     2. IMR data flows and payment flows
IMR has generally been studied in terms of calls placed, calls received, SMS sent, and roaming
data. The technical details and the payment flows among these four distinct roaming services
differ from one another in important ways. Today, roaming data is arguably the most important
and future-oriented of the services, but we address the technical aspects and payment flows in all
four services since all four are relevant to the discussion.
The most intensively studied form of IMR, and perhaps the easiest to explain, is the case where a
roamer subscribed to Home Network (HN) A in Country A is visiting Country B and uses
MNO B in Country B as a Visited Network (VN) to place a call (see Figure 2). The roamer pays
his or her Home Network in Country A. Mobile Network Operator (MNO) B actually places the
call on the roamer’s behalf, thus incurring both origination and termination costs, much as if its
own subscriber had placed the same call.
MNO B (the VN) receives no retail revenue for this service; however, it receives a payment at
wholesale level from MNO A (the HN). This wholesale payment is often referred to in the
literature as an Inter-Operator Tariff (IOT).11

Figure 2: Cash flows in IMR: calls made

Source: ARCEP (France), 2006.

The same basic principles apply to SMS messages sent while roaming, and to data; however,
each service has its own idiosyncrasies.

11
   MNOs sometimes distinguish between the nominal wholesale payment rate versus bilaterally agreed discounted
rates, in which case the IOT generally refer to the nominal rate.

                                                                                           Page 10

                     Electronic copy available at: https://ssrn.com/abstract=3529908
Addressing the high price of IMR for Southeast Asia (the ASEAN region)

SMS messages are subject to the same general payment arrangements as calls made; however,
the SMS is sent to from the VN to the HN as a control message (i.e. not as bearer traffic), and is
then sent to its end destination by the HN.12
Roaming data is subject to similar payment arrangements; however, the data is physically sent
from the VN to the HN (often using the services of an interconnection point known as a GRX or
IPX) before being shipped wherever it is needed by means of Internet Protocol (IP) transit
services.13 This enables the HN to account for the data, and thus to bill for it. It may also enable
the HN to apply rules, for instance to block content that is impermissible in the Home Country.
For calls received (see Figure 3), a significantly differently flow of traffic and payments comes
into play. The call is effectively forwarded from the HN to the VN. The VN receives a normal
payment of the international Mobile Termination Rate (MTR), but receives no additional IOT
payment (which suggests that the incremental cost of providing this IMR service cannot be very
great).
For calls received (whether domestic or international), the HN generally receives a Mobile
Termination Rate (MTR) payment from the caller’s network.14 The difference between the MTR
received and the MTR paid is thus crucial in understanding the HN’s costs. This difference is
small among EU Member States, where MTRs are stringently regulated, but internationally this
difference can be much larger.
In most countries, the party receiving a domestic or international call typically does not make a
retail payment for it; however, roaming calls received have historically been associated with
retail charges to the party receiving the call.15

12
     In most countries, there are no charges for receiving an SMS message while roaming.
13
   In the future, LTE local break-out services (a technical capability that is related to but not synonymous with the
local break-out (LBO) regulatory option introduced with Europe’s Roaming Regulation of 2012) might make it
possible for the VN to ship IMR data off directly via IP transit services, without incurring a rather unproductive
transit shipment to the HN.
14
  If the caller is on-net (i.e. is also a customer of the called party’s MNO), the MNO receives retail revenue instead
of an MTR.
15
     Under Regulation 2015/2120 (Roam like at Home), these retail charges have been prohibited.

                                                                                                  Page 11

                         Electronic copy available at: https://ssrn.com/abstract=3529908
Addressing the high price of IMR for Southeast Asia (the ASEAN region)

Figure 3. Cash flows in IMR: calls received

Source: ARCEP, 2006.

Numerous roaming specific costs can be relevant. These have been analysed in numerous
studies,16 but the results are often less than fully conclusive. The information about these costs is
generally known only to the MNOs, if it is known at all.17 In practice, the MNOs may not have a
full understanding of their own “soft” costs, and also may not wish to report them.
Different costs are applicable to the HN versus the VN, and not all IMR costs are relevant to
each IMR service.18 Among the costs relevant to the HN are:
       •   retail costs (including a proportionate share of the cost of customer acquisition and
           customer care);19
       •   wholesale payments (IOTs) from the Home Network to the Visited Network;
       •   roaming overhead costs;
       •   signalling;
       •   international transit (where relevant);

16
   See for instance Chapter 5 of GCC Roaming Working Group, J. Scott Marcus, Christin-Isabel Gries and Robert
Clarke (2015), “Consultation and Report on International Mobile Roaming (IMR) across the Gulf Corporation
Council (GCC) Region”, op. cit.; and Imme Philbeck, Jasper Mikkelsen, and Werner Neu, “Trans-Tasman Roaming:
Service Costs”, a study for the Australian Department of Broadband, Communications and the Digital Economy and
the New Zealand Ministry of Business, Innovation and Employment (MBIE), 30 May 2012, available at:
http://www.dbcde.gov.au/mobile_services/mobile_roaming/trans-tasman_mobile_roaming.
17
   See TERA (2016), “Assessment of the cost of providing wholesale roaming services in the EU”, study for the
European Commission, Section 3.2. TERA made a serious attempt to assess these costs for wholesale roaming
services in the EU based on questionnaires to the MNOs, but it is clear that large uncertainties and gaps remain.
18
     The discussion in this section closely follows the analysis in GCC Roaming Working Group et al., 2015.
19
  Some would argue that this cost is not relevant, inasmuch as it is a general cost of running the business that is
unlikely to influence IMR-specific decisions. Others would argue that it is a cost that needs to be recovered.

                                                                                                  Page 12

                         Electronic copy available at: https://ssrn.com/abstract=3529908
Addressing the high price of IMR for Southeast Asia (the ASEAN region)

     •   taxes (if relevant);20
     •   origination, termination, and other traffic related costs (where relevant).
It is generally impractical to differentially allocate retail costs to individual services;
consequently, they are often assumed to represent either a constant percentage of retail revenue
for all services, or else as a constant percentage of the cost of generating that retail revenue
Roaming overhead costs consist of a variety of administrative and network-related components
that are entailed in maintaining the roaming service. They can be said to include negotiation of
agreements, testing (IREG, TADIG), operations and maintenance (including accounting,
payments, revenue assurance, fraud prevention, dedicated staff costs, software and systems for
roaming operations), data clearing, financial clearing, and hubbing.
The cost of signalling is in principle relevant to all roaming services in both Home and Visited
Networks; however, this cost is small. (Marcus et al., 2012)
Some of these costs can be viewed as fixed annual costs, largely independent of traffic volumes.
Their impact on unit costs can be greater for MNOs whose volume of IMR traffic is low.

     3. Insights from the literature review
Numerous studies of IMR have been conducted over the past ten years or so. We are not aware
of any overall modelling in terms of economic theory, but good econometrics have been done
(see for instance Section 3.4)
In this section, we summarise some of the most relevant or important results in term of general
analysis (Section 3.1), the tendency of IMR to be priced at levels well in excess of cost (Section
3.2), the linkage of wholesale prices to retail prices (Section 3.3), and the societal welfare
implications of the over-pricing of IMR (Section 3.4).

     3.1     General policy assessments of International Mobile Roaming (IMR)
Several studies have analysed the impacts of the regulation of IMR in the European Economic
Area (EEA) / European Union (EU), and have explored alternative approaches to addressing
over-pricing of IMR services. Particularly noteworthy in this group is the Commission’s Impact
Assessment in preparation for the Roaming Regulation of 2012, (European Commission, 2012)
which includes a thoughtful analysis derived from a study to which Steffen Hörnig contributed.
(see also Marcus et al. (2010) and Marcus et al. (2012))
As previously noted, a number of studies have attempted to quantify the actual wholesale and
retail costs of providing IMR services, including Philbeck et al. (2012), GCC Roaming Working
Group et al. (2015), and Tera (2016).
As regards the current situation in Europe, BEREC (2014) can also be viewed as a significant
contribution to the literature. BEREC (2014) starts from the obvious but sometimes overlooked
proposition that (1) wholesale revenues to the Visited Network should be at least as great as the
associated costs, and that (2) retail revenues should be at least as great as the payments that the
Home Network makes to the Visited Network. Fulfilling these desiderata under an RLAH regime

20
  Taxes rarely enter into the European IMR discussion, but different tax rates can be a significant consideration in
other parts of the world.

                                                                                                 Page 13

                       Electronic copy available at: https://ssrn.com/abstract=3529908
Addressing the high price of IMR for Southeast Asia (the ASEAN region)

seems doubtful. BEREC (2014) therefore attempts to systematically explore ways to mitigate the
economic imbalances that flow from a pure RLAH regime, primarily by (1) reducing wholesale
payments to levels that are low but still in excess of marginal cost; or (2) imposing Fair Use
Limits (FULs) that limit the losses being imposed on MNOs, or by means of a combination of
the two.
Finally, TERA (2016) has attempted to model wholesale roaming costs in the EU. The work
appears to be conscientiously done, but a great many questions remain. Notably, transit costs
were not addressed, and retail costs were not within the scope of the study. A new study for the
European Commission is currently is progress.
Shortall (2013) (for the OECD) is in a somewhat different direction. It explores principles that
could form the basis for good practices in the establishment of IMR agreements between two or
more countries. In much the same vein, Clarke (2014) (for the ITU) seeks to provide guidance to
regulators as to practical considerations in addressing the challenges of IMR.
The OECD subsequently reported on developments in regard to IMR as part of their monitoring
of their 2012 Recommendation. The found that “there have been a marked reduction in
international roaming prices and a range of new service offers, in particular for mobile data
roaming, by mobile network operators (MNOs) that have aimed at responding to the demand of
roaming customers … Despite the reduction in prices, however, roaming prices in many
countries are far from competitive.” (Bourassa et al. (2016))

       3.2      High prices can be expected with IMR, and are challenging to remedy
In Bourassa et al. (2016), the OECD summarises the problem succinctly. In several previous
assessments, they had “concluded that price levels for international mobile roaming services
were unreasonably high. The causes for high roaming prices included the non-competitive
characteristics of the roaming market which led to high wholesale charges and, in turn, high
roaming retail charges. For example, in some cases the wholesale rates charged by foreign
operators could account for up to three quarters of the retail rate. Other factors leading to high
retail roaming prices included the fact that consumers did not take into account roaming services
when choosing a bundled mobile offer, lack of market contestability and low consumer
awareness of roaming prices.”
We are aware of no theoretical economics model of IMR, but the tendency to over-pricing that is
obvious in practice is in line with the literature. Where two firms that both possess market power
have vertically related offerings (in this case the HN and the VN as we explain shortly), a
tendency to dual marginalisation can be expected, leading to prices to end-users that are even
higher than those that a vertically integrated profit-seeking monopolist would choose.21 Consider
international calls between two countries, each with a monopoly operator (and thus no
competition for the same end-users, which corresponds to the typical nominal roaming situation
between the VN and the HN).22 “A noncooperative setting of access charges would necessarily

21
  See Laffont and Tirole (2000), Competition in Telecommunications, op.cit., Section 5.3 (in a chapter written with
Patrick Rey). They refer to this form of over-pricing as the chain of monopolies or pancaking problem.
22
     As we explain shortly, this condition is not fully met, but competition for roaming services is weak.

                                                                                                     Page 14

                          Electronic copy available at: https://ssrn.com/abstract=3529908
Addressing the high price of IMR for Southeast Asia (the ASEAN region)

result in the well-known double marginalization, each country adding its monopoly [mark-up] to
its perceived marginal cost.”23
Dual marginalisation can lead to end-user prices that exceed even the monopoly price, and are
thus above the economically optimal price for the network operators. The problem is not easy to
correct, even for the network operators in question (which probably explains why high IMR
pricing has been difficult to correct in general). Dual marginalisation can be ameliorated
somewhat if one of the monopolists acquires the other, or if they are permitted to set prices to
one another cooperatively – in that case, the end-user pays “only” the monopoly price.
Otherwise, regulation would appear to be the most practical alternative to high prices.
These considerations also apply to IMR, albeit imperfectly. The VNs typically do not have
perfect monopoly control over visitors; nonetheless, they probably possess considerable pricing
power. In Europe (and in most countries around the world), it is rare to have more than four
mobile network operators (MNOs). Each HN could in principle choose one (or more) of these
four for each Visited Country to which its roamers travel. One might therefore well imagine that
prices charged to the HN would reflect oligopoly prices with up to four market participants,
which would not necessarily differ greatly from competitive prices. In practice, however, it is
often the case that one or more MNOs (1) have only limited geographic coverage, or (2) do not
support all of the mobile connectivity options desired by the HN’s customers (for instance,
support only 2G and 3G but not 4G mobile connectivity); or (3) possess networks that do not
offer the quality or speed that the HN desires. The number of realistic choices is therefore more
likely to reflect an oligopoly situation rather than perfect competition.
The practical reality may be considerably worse. Roaming can be thought of as a multi-stage
game, where the HN initially chooses its VN partner(s) in a given Visited Country, then the
roamer effectively chooses a VN from among those supported by its HN in the Visited Country
(in most cases, the handset makes this choice automatically based on steering preferences
downloaded from the HN to the handset), and only then is the end-user choosing how much use
to make of the service. Beyond this, the HN’s ability to steer traffic to its preferred VN is limited
by numerous practical considerations, and especially by any gaps in coverage of a particular VN.
Under these circumstances, the flow through of wholesale competition into end-user retail prices
might well be attenuated.
The effectiveness of wholesale bargaining is in any case not altogether clear. Market players
claim that there is intense negotiation over wholesale prices, sometimes leading to prices
discounted as much as 40% below regulated European wholesale price caps.24 The negotiations
between MNO multi-country groups could be complex, since each could offer to steer roamers to
the other in one country or another.
At the same time, assuming that the prices are symmetric (as appears to often be the case), then
the prices only matter for unpaired minutes. If MNO A is both an HN and a VN in Country A,
and MNO B an HN and VN in country B, and each steers some or all of its roamers to the other,
then the only minutes that result in a net payment are those to which roaming minutes consumed

23
  Laffont and Tirole (2000), page 184. They explain the underlying drivers, and reference an extensive literature
addressing this point.
24
  This emerged in multiple interviews, as reported in J. Scott Marcus and Imme Philbeck (2010), “Study on the
Options for addressing Competition Problems in the EU Roaming Market”.

                                                                                              Page 15

                      Electronic copy available at: https://ssrn.com/abstract=3529908
Addressing the high price of IMR for Southeast Asia (the ASEAN region)

by MNO A roamers in country B exceed those consumed by MNO B roamers in country A (or
vice versa) – since the other minutes net to a zero payment, the rate charged for them is
irrelevant. For purposes of the quarterly statistics that BEREC generates,25 the MNOs generally
report the paired minutes at a high rate (in Europe, typically the wholesale price cap rate), thus
making the average wholesale payments for all minutes appear higher than the unit prices for
unpaired minutes.26 In other words, wholesale negotiations may possibly be somewhat more
effective than most experts have assumed based on the BEREC statistics.
For any given pair of MNOs and countries, one MNO will tend to be a net payer, the other a net
receiver. Their negotiating interests are not aligned. The net payer will prefer a low wholesale
price, the net receiver a high price.
Taking all of this together, it is not surprising that both wholesale and retail prices are high in
practice. This is clear in the statistics published by BEREC where regulated wholesale and retail
prices tend to be just below the price caps, and unregulated “bargain” prices are actually on
average above the caps. Competitive forces appear to be weak. As BEREC notes, “… [f]or voice
roaming services, average EEA prices are close to the regulated caps. This suggests that
providers see little attraction in competing on Eurotariff rates, despite the fact that there is a
significant margin between typical wholesale prices and retail caps.”

       3.3   Wholesale payments influence the retail price
If both (1) wholesale unit charges for IMR and (2) the volume of roaming traffic between MNOs
A and B in Countries A and B were symmetric, one might well imagine that the wholesale
charges would be totally irrelevant to retail pricing, since they would net to zero; in practice,
however, they are highly relevant to retail pricing. Suppose, for instance, that MNO A were to
choose to lower its roaming prices in order to win more business from country A roamers who
travel in Country B (i.e. to attract them when they choose to make a subscription). To HN MNO
A, the wholesale payment to VN MNO B is a real cost (assuming that the MNOs are not
affiliated with one another). MNO A will therefore tend to view its payment to MNO B as a key
component of its real marginal cost, and is unlikely to set its retail price below its real marginal
cost.
Assume (without loss of generality) that MNO A does not offer retail services in Country B, nor
MNO B in Country A. Under these circumstances, one can make a straightforward economic
argument that the societally optimal wholesale rate is equal to marginal cost (ignoring network
effects).27 If, however, one takes into account the fact that fixed and marginal costs must also be
recovered, then there is an argument to be made that higher mark-ups are needed to recover those
costs.

25
  See for instance BEREC (2016), “International Roaming: BEREC Benchmark Data Report: April – September
2015”, BoR (16) 28 Rev.1, at http://berec.europa.eu/files/document_register_store/2016/3/BoR(16)28-
1_Roaming_Benchmark_Data_Report_04-09_2015.pdf.
26
     Ibid.
27
   See Laffont and Tirole (2000), Competition in Telecommunications, op.cit., page 197. When one takes network
effects into account, the societally the optimal price (disregarding any dynamic impact on investment) is arguably
below marginal cost. See also Armstrong and Wright (2007), op. cit.

                                                                                               Page 16

                      Electronic copy available at: https://ssrn.com/abstract=3529908
Addressing the high price of IMR for Southeast Asia (the ASEAN region)

If the two MNOs happen to operate in the same country, and thus to compete for the same end-
user customers, then they tend to have an additional incentive to maintain high wholesale rates.
For reasons just noted, the retail price is unlikely to be less than the wholesale payment; thus,
high wholesale payments effectively raise a rival’s cost, and weaken competition between them
for the same end-user consumers.28 Under suitable assumptions, both benefit.

       3.4      Societal welfare implications
In European Commission (2011), Hörnig et al. provides a detailed econometric analysis of the
societal welfare impact of over-pricing of IMR services, and the impact of the Regulation put in
place to address over-pricing.
In doing so using the classic Harberger Triangle analysis, a key first step is to estimate the price
elasticity of demand of the service. Hörnig found the own price elasticity of demand for IMR
calls made, calls received, and SMS to all be in the -0.24 to -0.27 range, which is to say that
demand for the services is relatively inelastic; however, the own price elasticity of demand for
IMR roaming data was in the -1.23 range, which is to say that consumer demand for roaming
data is relatively elastic.29
It is possible that the price elasticity of demand for roaming voice services is higher today. In the
past, even with regulation, the price was perceived as being too high – most Europeans placed
calls while roaming only when unavoidable, both before and after the regulations came into
force. We have argued, however, that when the price of calls made while roaming approached
that of domestic calls, that the price elasticity of demand for placing roaming calls could also be
expected to be similar.30
The same impact assessment report by Hörnig et al. (in European Commission (2011)) also
assesses the impact on societal welfare (in terms of static economic effects) of the changes
already put in place up to that time. It compares a “business as usual” scenario, where the price
reductions already put in place by the 2009 Roaming Regulation continue without change, to an
alternative (counterfactual) scenario where the Roaming Regulation would have been
discontinued, and where prices would have returned to the previous unregulated levels.
Compared to the counterfactual scenario, the continuation of regulation over the years 2012
through 2014 would increase consumer surplus by € 18,600 million, while reducing industry
profits (producer surplus) by € 5,000 million. The difference of € 13,600 million represents a net
gain in societal welfare (i.e. a reduction in deadweight loss), and reflects increased consumption
of IMR services.31

28
     See for instance Laffont and Tirole (2000), pages 190-195.
29
     The mathematical specification of this work appears on pages 101 through 104 of the document.
 J. Scott Marcus and Imme Philbeck (2010), “Study on the Options for addressing Competition Problems in the EU
30

Roaming Market”, op. cit.
31
  The application of welfare economics to international mobile roaming is explained at some length in Section 8.4
of the previously cited report for the six GCC countries, GCC Roaming Working Group et al. (2015). See also
European Commission (2011).

                                                                                                 Page 17

                         Electronic copy available at: https://ssrn.com/abstract=3529908
Addressing the high price of IMR for Southeast Asia (the ASEAN region)

These may sound like large numbers, but it is important to bear in mind that in a pre-Brexit
European Union with some 508 million inhabitants, this works out to about €9 per inhabitant per
year.
As a related matter, the societal welfare impacts are more important for some services than for
others. Societal welfare gains come mainly from increased consumption thanks to the price
elasticity of demand. SMS is, however, a service that is in global decline (see Figure 4), possibly
due to competition from Over-the-Top services. (ITU (2017b))
Figure 4. Messages sent via mobile handsets by service type, worldwide, 2010–2018.

Source: Analysys Mason (2014)

The logic of promising increased usage of SMS is thus questionable. The same argument holds
for calls made or received while roaming, but with less force, because the tendency of OTT
services to substitute for traditional voice is primarily limited to relatively expensive
international voice calls. For calls made and calls received (and also for SMS), the impact of
lower prices on societal welfare is further limited by the fairly low price elasticity of demand.
For roaming data, however, there is the opportunity to facilitate use of a growing service,
arguably the service of the future; moreover, the impact of price reductions is likely to be large in
light of a relatively high price elasticity of demand.

   4. Learning from experience in other regions
At this point, a considerable body of experience in addressing the challenges of high IMR prices
has been accumulated. In seeking ways forward for ASEAN, it is possible to benefit from this
experience in many distinct dimensions.

                                                                                      Page 18

                    Electronic copy available at: https://ssrn.com/abstract=3529908
Addressing the high price of IMR for Southeast Asia (the ASEAN region)

Different regulatory approaches are possible, including (moving from less intrusive measures to
more intrusive) statistics collection, raising consumer awareness, bill shock prevention measures,
and wholesale and retail price controls. In this section, we discuss their various implications for
the ASEAN region. We take up wholesale and retail price controls in Section 4.1, 4.2, and 4.3;
statistics collection in Section 4.4; and bill shock prevention measures in Section 4.5.
Trade agreements including the General Agreement on Trade in Services (GATS) raise a number
or issues and risks that might need to be addressed if ASEAN were to consider a more activist
approach to IMR (see Section 4.6)
Europe has also implemented measures that are either unproven, risky, or else proven to be
problematic, as we explain in Section 4.7. They may be imprudent in a region where regulatory
institutions at regional level are not as well developed as those in the EU. We do not suggest that
these be attempted in the ASEAN region. These are:
     •   Structural solutions (Alternative Roaming Provider, Local Break-Out); (see Marcus,
         Wernick, Gries and Philbeck (2016))
     •   Roam Like at Home (RLAH). (see Marcus and Petropoulos (2016))

     4.1    National, bilateral, multi-lateral, regional, or global?
A first question entails the appropriate level at which policy or regulatory measures should be
undertaken: national, bilateral, multi-lateral, regional/supranational, or global?
Much of the IMR discussion has been about price controls, and the discussion is most acute in
that context. At the end of this section, we return to “softer” regulatory measures.
By definition, the Visited Network is in a different country than the Home Network. Price caps
that involve coordinated price regulation of both the HN and the VN thus always involve
cooperation among two or more governments.
If a country were to unilaterally regulate only the wholesale prices of its VNs, it would generate
no benefits for its own residents. If a country were to unilaterally regulate only the retail prices of
its HNs, great care would need to be taken to ensure that regulated prices were high enough to
cover the wholesale prices that the HN was obliged to pay to VNs in other countries. There is
thus little, if any, practical scope for price regulation of IMR solely at national level.
(Clarke (2014), pages 3-5)
Government initiatives can be said to fall in three broad categories: (1) purely voluntary
arrangements, (2) voluntary arrangements made under the threat of regulation, and (3) regulatory
arrangements. We consider these in turn.
Bilateral voluntary commercial arrangements have enjoyed mixed success. The bilateral
arrangements between Russia and Poland, for instance, appear to have produced no tangible
results at all. (Shortall (2013))
New Zealand and Australia in February 2013 announced an agreement to enhance their national
regulators’ powers in respect of international roaming services between their two countries.32

32
  See the press release of the New Zealand Prime Minister, “CER strengthened after Queenstown talks”, 10
February 2013, available at https://www.beehive.govt.nz/release/cer-strengthened-after-queenstown-talks

                                                                                      Page 19

                    Electronic copy available at: https://ssrn.com/abstract=3529908
Addressing the high price of IMR for Southeast Asia (the ASEAN region)

This included allowing the regulators, should they choose to regulate wholesale roaming rates, to
impose retail price controls as well. Although draft legislation was prepared, this approach was
never actually implemented for two main reasons. First, the announcement had the effect of
reducing trans-Tasman roaming rates significantly (including the introduction of daily flat-rate
pricing), meaning that the intervention arguably became unnecessary; and second, the September
2013 federal elections in Australia resulted in a new government with a more light-handed
approach to economic intervention.
In ASEAN, bilateral arrangements falling short of actual regulation have been in place between
Singapore and Malaysia (2011), and also between Singapore and Brunei (2014). These
arrangements, whose introduction relied on ‘discussions’ with operators with an implicit threat
of regulation, have achieved limited but positive results (see Section 1.3). Per OECD (2013),
“Singapore and Malaysia mobile phone subscribers saw price reductions of up to 30% for voice
calls and 50% for SMS, when they use the mobile roaming service provided by all mobile
operators in Malaysia and Singapore respectively.”33
More recently, at the regional level, ASEAN member states agreed in November 2017 to the
ASEAN Framework on International Mobile Roaming (AFIMR). (ASEAN (2017)) Under the
AFIMR, which is not legally binding, it is planned that national regulatory authorities will
encourage operators in their jurisdiction to “i) adopt measures that promote price awareness and
transparency of international mobile data roaming services to its subscribers, such as by sending
Welcome Short Message Service (SMS), providing accessible and easy-to-understand pricing
information, and informing consumers when they reach a certain limit for data roaming usage;
and ii) offer its subscribers at least one option of mobile data roaming service on a daily flat-rate
basis in the territory of the other [ASEAN member state] by 2020. The rate, terms and
conditions of such a service will however be mutually determined by the operators involved in
providing the service.”
It is rather soon to assess the concrete effects of the AFIMR. The AFIMR did not envision
collection of statistics, and we are not aware of publicly available progress reports. The
transparency mechanisms will probably have positive effect. For the rest, given the focus solely
on ensuring availability of daily flat-rate offers from at least one MNO (but not on their price),
and on the total absence of enforcement mechanisms, we question whether the measures will
have much practical effect.
The approach taken in the East Africa Community region (EACO) is based on multilateral
voluntary commercial arrangements. The countries agreed to take measures to reduce the price of
IMR. Positive effects are visible, but prices are far from uniform. Implementation by network
operators has been very uneven. Some IMR prices have been substantially reduced, while others
remain high.
In sum, purely voluntary arrangements have tended to have little or no effect on actual roaming
arrangements; nominally voluntary arrangements under the threat of regulation have had
somewhat greater effect, but still limited effect.

33
     This is consistent with the joint press release dated 20 April 2011 available                          at
https://www.imda.gov.sg/about/newsroom/archived/ida/media-releases/2011/singapore-and-malaysia-to-reduce-
mobile-roaming-rates.

                                                                                           Page 20

                     Electronic copy available at: https://ssrn.com/abstract=3529908
You can also read