On the Impact of Call-Back Competition on International Telephony

7/31/2019 On the Impact of Call-Back Competition on International Telephony

1/21

Journal of Regulatory Economics; 20:1 2141, 2001

# 2001 Kluwer Academic Publishers. Manufactured in The Netherlands.

On the Impact of ``Callback'' Competition onInternational Telephony*

FABIO M. MANENTIUniversity of York (U.K.),

Department of Economics,

Universita di Padova (Italy){. Dipartimento di Scienze Economiche ``M. Fanno''

Abstract

In this paper we build a simple three-country model to evaluate the impact of ``callback'' on international

telephony. The effects on both accounting rates and collection prices are studied. Callback rms exploit

arbitrage opportunities in collection prices among countries, rerouting calls that originate in countries

with high prices for international phone calls via countries with low prices. Contrary to what is commonly

perceived, we show that callback tends to magnify the distortions associated with the current accounting

rate regime. In particular, callback puts upward pressure on low price countries' accounting rates and on

collection charges. Callback companies are assumed to enjoy a volume price discount on each rerouted

call; we show that the larger the price discount offered to callback companies, the higher the prices for

international calls in the country hosting callback.

1. Introduction

The aim of this paper is to analyze and evaluate the impact of ``callback'' on retail and

interconnection charges in international telecommunications. Callback is an alternative

way of placing international calls (alternative calling procedure1 in the industry jargon)

that new technologies have made available to customers for international voice

communications. Until recent market liberalization decisions by many national

* Paper presented at the 25th EARIE conference, Copenhagen and at the 1998 ASSET Meeting, Bologna. I

would like to thank Gianni De Fraja for his continuous and constructive help. Gabriella Chiesa, Riccardo

Faini, Indrajit Ray, Carlo Scarpa, Paola Valbonesi, Julian Wright and the seminar audiences at the

University of Brescia and Pavia provided many helpful comments on earlier versions of the paper. Two

anonymous referees and the editor, Michael Crew, are also acknowledged. I am particularly indebted to

the Alcuin college of the University of York (U.K.) for the nancial support for my studies.

{ Address for correspondence: Dipartimento di Scienze Economiche, Universita di Padova, Via del Santo

33, 35123 Padova (Italy). E-mail: [email protected]

1 ACP hereafter.


2/21

governments are effectively put in place, ACPs, and callback in particular, are one of the

main challenges to the monopolistic power of national carriers.2

Callback is based on a very simple idea: suppose a customer in Italy wants to speak with

a friend in a different country (the U.S.). If the price of a call from the U.S. to Italy is lower

than the price of a call in the opposite direction, then it may be cheaper for the Italian

resident to be called rather than to call the friend located in the U.S. He will then

compensate the friend for the cost of the call. Instead, he can use a public callback service

and get a computer to do the call-rerouting.

In the typical callback call, rst a customer in one country places a free call to the

callback operator equipment (usually just an automated system) in a second country. The

operator, which detects the caller's identity without answering the call, calls the customer

back at a predesignated number providing the customer with a dial tone in the secondcountry and connects the customer to a number in the second or third country. The original

call is now ``rerouted'' via the country where the callback rm operates.

Most callback enterprises are located in the U.S. where lower prices for international

phone calls provide wide margins for arbitrage opportunities; callback calls are usually

provided at a rate 2050% cheaper then that of the carriers. Unfortunately, the FCC does

not compile explicit data on callback trafc. Callback falls under the umbrella of ``resale''

trafc and there is no way for the FCC to track callback trafc versus other types of resale

trafc.3

Although price competition and deregulation together with the emerging of new

services, namely voice over IP, have made callback less protable than it was when this

service was introduced in 1992, global callback generated in 1999 estimated revenues of

around $800 million.4 It is widely accepted that the demand for callback services will

remain high in developing nations where international calls are still expensive. Nowadayscustomers place their calls using callback services in more than 200 countries.5

Despite its growth, callback is still a very controversial phenomenon. On the one hand

the FCC, the U.S. regulatory authority that provides the licenses to U.S.-based callback

rms, and international organizations such us the OECD and, more recently, the ITU

(International Telecommunications Union) have argued in favor of callback.6

Accordingly, callback should be encouraged because it would engineer downward

pressure on prices for international phone calls. On the other hand, 100 countries around

2 An exhaustive description of the ACPs is provided in OECD (1995). The most common practices of

ACPs are: reverse charge calling, credit card calls, country direct service, country and beyond service,

callback, international free-phone service, international leased lines and rele (hubbing).

3 In 1997 pure-resale trafc accounted for 15% of the whole outgoing U.S. trafc; this gure is expected

to grow even more with time. See Lande and Blake (1997).

4 See Nye (1999).

5 See Choi et al. (1999).

6 After more than two years of legal disputes, the FCC gave its nal assent to the provision of callback

services in 1995 (10 FCC Rcd 95-40). See Propp (1996) for a discussion of the legal issues related to the

FCC decision to authorize callback operators in the U.S. But note that not all types of callback are

allowed: the so called ``call-bombardment'' and ``answer suppression'' are fraudulent since they interfere

with the billing mechanism of the national operators.

22 FABIO M. MANENTI


3/21

the world, mainly developing countries, have so far prohibited callback in their territories 7

and are putting pressure on the FCC to stop licensing of new callback operators. These

countries perceive callback as a problem because it deprives the local operator of revenues

from the international telecommunications sector.8 This controversy has been, and still is,

a source of strong debate among operators.

The surprising growth and development of this particular form of competition has been

made possible by the way in which international telecommunications is organized. The

typical method of governing interactions among national telecom operators is the

accounting rate system. Consider, for example, a call from Italy to France: the Italian

carrier keeps the revenue for this call but, for this call to be terminated (i.e., received by a

user in France), it must pay an access fee to the French operator for the use of its network.

The same fee applies to each call that goes from France to Italy; this fee (called theaccounting rate or the interconnection charge) is arranged in bilateral negotiations

between the two operators. As suggested by Frieden (1997), under this regime, carriers

have failed to negotiate adequate interconnection charges. Even though accounting rates

are falling over time, they are still well above the cost of providing interconnection.9 This

distortion in the access charges has kept prices for international phone calls at an

articially high level, preventing them from reecting the dramatic costs reduction

implied by the development of new technologies.

Callback advocates, such as the OECD, argue that this form of competition between

international operators creates downward pressure both on retail prices and accounting

rates. According to Frieden (1997), the more nations authorizing callback, the more

accounting rates and collection charges10 decrease in consequence.

In a recent paper, Choi et al. (1999) provide a formal treatment of the callback

phenomenon. Interestingly, they show that one of the main reasons for the existing ofcallback services is the inefciency of the accounting rate system. Nonetheless, in their

paper, the inefciency is not fully explained; in addition, the authors use a two country

model, while most callback operators reroute calls between two high price countries and

are located in a third country with low international rates. We use a three country model in

order to capture the basic structure of a callback call; furthermore, by making endogenous

the determination of the accounting rates between countries, we are able to highlight the

consequences of callback both on accounting and retail charges. We show that with

callback, accounting rates and collection prices on different routes, otherwise

uncorrelated, become interconnected. As a consequence, the effects of the accounting

rate regime are no longer restricted to a few specic routes, but rather are extended to the

interconnection agreements between the countries involved in the calls rerouting.

This paper demonstrates that callback puts upward pressure on the accounting ratesbetween ``high price'' countries (called target countries), which in turn pushes up

7 The list of these countries is kept up-to-date on the ITU web-site (http://www.itu.int/).

8 Arguments for and against callback are summarized in Kelly (1996).

9 Data about accounting rates are condential and, apart from the U.S. and the U.K., they are not

published; estimates of the accounting rates can be found in OECD (1997).

10 According to the industry jargon, the price for a phone call is also dened as ``collection charge''.

IMPACT OF ``CALLBACK'' COMPETITION 23


4/21

collection prices for calls between the same countries. In addition, it is shown that the

impact of callback on the prices charged by the low price carrier (called the host carrier)

depends upon the price discount offered to callback rms. Quite surprisingly, it is found

that the larger the discount, the higher the host carrier mark-up.

Although in a different context, these results conrm similar arguments presented in

Alleman (1998) and are easily understood once the effects of the reciprocal accounting

rate regime are fully considered. Indeed, instead of undermining it, callback exacerbates

the distortions that this regime has on accounting rates and collection charges. Consider

target countries: carriers in target countries may nd it optimal to push up prices for

international calls placed between their customers by raising the reciprocal interconnec-

tion fee; this stimulates alternative callback rerouting and increases trafc imbalances

between target and host countries. According to the accounting rate regime, this impliesthat carriers in target countries receive higher settlement payments from the country

hosting the callback rms. On the other hand, also the host country carrier may be induced

to increase its tariffs for calls directed to target countries; this is to reduce the amount of

calls rerouted via callback and to lower the settlement payment due to the carriers in target

countries.

These results must be evaluated considering the evolving context of the market for

international telephony. Many factors are inducing a decline in prices for international

phone calls; technologies are more efcient, in many countries markets are now open to

competition and new communications services (typically, the Internet) are eroding the

market for traditional telephony;11 our main argument is that callback instead of

accelerating the decline of collection prices has indeed prevented them from converging

down to costs.

The paper is organized as follows. In section 2 we present the model of callback. Section3 evaluates and discusses the solution of the game and section 4 concludes the paper.

2. The Model

We model a world with three countries A, B and C. For the sake of simplicity, we assume

that in each country, telecommunications services are provided by only one operator

(carrier). This assumption may appear rather unrealistic, especially in relation to the

country hosting the callback rms. In the real world, price differences occur mainly

because some countries have opened the market to competition. Nonetheless, as it will

become clear later, the main arguments of the paper are not affected by this assumption. Inthis stylized world only international telephone services are provided, that is, calls

originating in one country and terminating (i.e., received by a user) in another country. We

make the following assumptions:

11 In Manenti (2000) it is estimated that the demand for international telecommunications in Italy for the

period 199197 is signicantly and negatively affected by the dimension of the Internet; this suggests

that the new Internet-based communications channels are substitutes for standard telephony services.

24 FABIO M. MANENTI


5/21

(i) National carriers in countries A and B have the same cost conditions. We denote

with c the cost incurred by these operators in collecting a call on their local

network and sending it to the ``half-way point''.12 Instead, carrier C (C for

cheaper) is more efcient: it incurs a lower cost per outgoing call: cc5c. Given its

superior technology, carrierC can set lower prices for calls directed to countries A

and B: prices for international calls differ according to the direction of the call (i.e.

a call from C to A costs less than a call in the opposite direction).

(ii) Each call directed to a country must transit on that country's network in order to

reach the nal user. The transmitting carrier pays an access fee for each call to the

receiving carrier for the use of its network. This fee, called the ``accounting rate'',

is reciprocal (i.e. the same access charge for a call from A to B applies to a call

from B to A) and is negotiated between the carriers involved in the calltransmission. The accounting rate is determined according to a Nash-bargaining

process.

(iii) The cost of providing access (the cost of delivering other carriers' calls) is co and is

the same across countries.

(iv) Callback rms (CB hereafter) are located in country C. The CB sector is

characterized by price competition with free-entry. CB enterprises exploit

arbitrage opportunities based on calls rerouting: a CB rm located in C, the

cheap country, can reroute a call from A to B (or vice versa) transforming it into

two outgoing calls from C. For these reasons, we denote A and B as target

countries and C as the host country.

2.1. The Demand for International Telecommunications

Our main purpose is to investigate the impact of CB on high price countries' reciprocal

accounting rates and tariffs. For this reason we assume that CB services are provided only

on the main route AB: a customer located in a target country can place his calls to users in

the other target country either by demanding the services of his national operator or via

callback. In order to differentiate these two kinds of calls, we dene the former as

``standard calls'' and the latter as ``callback calls''.

We rule out the possibility that CB rms offer to reverse calls from A orB to C. This is a

simplication because, usually, once the customer located in the target country (for

example A), is connected with the CB terminal equipment, he can then place a call either

in the host country C or to a foreign destination (B).13

CB rms act symmetrically with respect to the target carriers: each CB rm offers to

reroute calls both from A to B and from B to A. This is what happens in practice in the CBmarket: unless a country has banned CB, then CB rms provide connection in both the

directions of the call.

12 This is called the imaginary point of interconnection between national networks.

13 However, if markets are separated and there are no economies of scale, the introduction of CB even

between target and host countries would not affect our results. For an analysis of CB between target and

host countries, see Choi et al. (1999).



6/21

There are n CB rms, indexed by h 1; . . . ; n, all located in the host country C. Theyhave two main effects: on the one hand they compete with networks A and B by lowering

the amount of calls that they place with each other and, on the other hand, they increase the

number of calls originating in C.

We model standard calls and CB calls as imperfect substitutes: the demand for CB

services increases with the price for standard calls charged by the national operator.

Formally, denoting by Qji the demand for standard calls from i to j, by Y

ji the demand for

CB services on the same route, and by pji and v

ji the prices for these calls, the following

assumption is made:

Assumption 1:

(a)qQ

j

i

pj

i

;vj

i

qpj

i50

qYj

i

vj

i

;pj

i

qvj

i50

qYj

i

vj

i

;pj

i

qpj

i40 i;j A; B and i T j.

(b) The demand for standard calls and the demand for callback calls are the same

across countries and have constant price elasticity, denoted by Z andE respectively:

Z qQ

ji

qpji

pji

Qji

i;j A; B; C and i T j

e qY

ji

qvji

vji

Yji

i;j A; B and i T j

(c) e4Z41.

The assumption of imperfect substitutability (1a) is taken on practical grounds: on the onehand, although CB is becoming a very popular telecommunication service, its use is still

conned to a restricted share of well informed customers. In addition, calls placed via CB

are often of a lower quality when compared to standard calls (delays in the conversation,

echoes, noise etc.): quality considerations may reduce the attractiveness of these services

affecting customers' demand. According to these observations, it is even natural to assume

that the demand for CB calls is more reactive to change in prices than the demand for

standard calls (assumption 1c): although we cannot empirically support this assumption,

we believe it is reasonable to think of CB users more concerned with changes in prices

than the customers of standard calls.14

2.1.1. Trafc Flows

Only standard calls are placed between target countries but, as explained, the demandfor calls between target carriers depends not only on the collection price p

ji charged by the

local monopolist, but also on the price for a CB call vji : Q

ji Q

ji p

ji ; v

ji .

For each call that is rerouted from the main line AB via C, two outgoing calls from C

are placed, one directed to each country. This implies that the amount of calls delivered on

14 The condition Z41 is necessary to guarantee positive equilibrium prices and is supported by some

studies of elasticity of demand for international calling. See OECD (1997) or Lande and Blake (1997).

26 FABIO M. MANENTI


7/21

carrier C's network and directed to country i is the combination between the amount of

standard calls Q ic and CB rerouting Yji . Therefore, the trafc ow between host and target

carriers is given by

Traffic Flow

TFic ?

standard calls

Qicpic

callback rerouting

Yji v

ji ;p

ji Y

ijv

ij;p

ij

with i A, B and i T j.According to our simplication, CB rms reverse calls only on the route AB. This

implies that the trafc of calls from countries A and B to country C is not affected by CB.

It is useful to provide a graphical representation of the trafc ows between the three

countries; see gure 1.

2.2. Fundamentals of the Reciprocal Accounting Rate Regime

The current system of multilateral collaboration in the transmission of international

calls is the accounting rate regime. According to this regime, the transmitting carrier pays

for each call sent to the foreign carrier an access fee (the accounting rate) for the use of its

network and receives an equal payment for each incoming call from this operator. The

accounting rate is negotiated at bilateral meetings between telecom carriers.

We assume that these negotiations take the form of a Nash-bargaining process in which

the contracting parties have the same bargaining power.15 Formally, the accounting rate

between country i and j ai;j is given by

ai;j argmax p1=2i p1=2jh i

1

where pi is the carrier i's prot function. For the sake of simplicity, we assume that the

disagreement point of this bargaining process is that there is no interconnection agreement,

namely the carriers cannot provide international services between the two countries and

both their prots are zero. This is a simplifying assumption since each carrier always has

the option of rerouting calls through the third country.16

a is the accounting rate between target carriers (A and B) while ti is the accounting rate

between carrier C and carrier i. Table 1 summarizes the adopted notation.

The analysis of the current accounting rate regime is beyond the scope of this work.

Nonetheless, as it will become clear later, since CB is intrinsically related to the process

15 This is realistic in this framework, in which international telephony services are supplied by monopolies

but it applies even when competition is considered. In this case we can think of a regulatory policy that

assigns to a representative carrier a monopoly equivalent bargaining power.

16 We wish to thank an anonymous referee for this observation. Nonetheless, taking into account for each

carrier the option of rerouting calls via a third country would require modeling transit tariffs for rerouted

calls and also the additional bargaining processes between the countries of transit in which these tariffs

are determined. This would make the analysis much more complicated without, at the same time, adding

any new relevant insight to the model. The assumption of zero prot disagreement point is frequently

adopted in the existing literature; see Wright (1999) and De Fraja and Valbonesi (2001).



8/21

through which collection charges are determined, it is important to recall briey how this

regime works; in particular two main effects are worth noting: 17

(1) When carriers face the same demand and cost conditions (perfect symmetry), trafc

ows between countries are balanced (outgoing traffic incoming trafc) and thebargained access price is set equal to the cost of providing access. In our case, since

carriers A and B are identical, then a co.(2) When there is asymmetry between countries either at the demand or at the cost

level, trafc is no longer balanced and the bargained access price is shown to be

above the cost of providing access. In this model, due to the technological

asymmetry, trafc ows between host and target countries are unbalanced; thisimplies that ti4co.

Perfect symmetry is a very special situation. In more general cases trafc ows among

countries are unbalanced. That is, due to the reciprocity of the accounting rate regime, the

carrier that terminates more calls than it originates receives more than it pays to the foreign

carrier for interconnection to its network.18

When the trafc is unbalanced, carriers fail to set the reciprocal access charge at the cost

level. This is considered one of the reasons why international calls charges are higher

when compared to national long distance calls. The accounting rate is ``de facto'' an

additional cost per call that each carrier must pay to deliver its calls. The higher the

accounting rate, the higher the cost per call and the higher the collection prices.19

Figure 1. Trafc ows.

17 For a formal treatment, we refer to an earlier version of this paper; in Manenti (1997) we study the

consequences of demand and cost asymmetries on bargained accounting rates.

18 The growing imbalances in trafc ows between countries have generated large decits in countries like

the U.S., Sweden or Australia and large surpluses in other countries such as Mexico or Germany. Data on

trafc ows are in ITU (1999).

19 For these reasons many governments and organizations are promoting the introduction of new

interconnection arrangements. See De Fraja and Valbonesi (2001) for a discussion of the proposed

reforms of the accounting rate regime.

28 FABIO M. MANENTI


9/21

2.3. Firms' Prots

2.3.1. Callback Firm's Prot

CB rms are assumed to be identical and, as stated above, they are established in

country C. For each rerouted call from one target country to the other, they must pay the

hosting carrier for the price of two outgoing calls. In practice, CB rms enjoy lower prices

than residential customers since they fully exploit the volume discounts offered by the host

carrier: assuming that CB rms do not face additional costs for each call, the CB cost per

call is equal to a discounted sum of two outgoing calls from C. We denote the price

charged to CB companies by pcb, with pcb ypac pbc , where y [ 0; 1 represents thevolume discount. For the moment, let us assume that y is exogenous. Note that pcb is

independent of the direction of the rerouted call.

Since each CB rm reroutes calls both from A to B and from B to A, the prot function

for the h CB rm is then

phCB

i;j A;B

vji;h pcbYji;h ? h 1; . . . ; n and i T j 2

where vji;h is the price for calling from i to j charged by the h CB operator and Y

ji;h is the

demand it faces.

2.3.2. Carriers' Prots

The prot functions for each established carrier are given by

pi pji c a

h iQ

ji ? a coQ

ij ? p

ci c ti Q

ci ?

ti coQic ? Y

ji ? Y

ij ? i;j A; B and i T j 3

Table 1. Notation

Name Description (i;j A;B and i T j, h 1; . . . ; n)

a Accounting rate between A and B

ta Accounting rate between C and A

tb Accounting rate between C and B

pji Price for a call from country i to country jvia domestic carrier

pic Price for a call from the host country C to country i

vji;h Price for a callback call from country i to country jvia operator h

Qji ? Demand for standard calls from target country i to target country j

Qic ? Demand for standard calls from country C to target country j

Yj

i;h?

Demand for callback calls from country i to country jvia operator h



10/21

pc pcb 2cc ta tb Yba ? Y

ab ?

i;j A;B

pic cc ti

Qic ? ti coQci ? 4

where pji is the price for a call between target countries via established carriers i;j A;

B, while pic and pci are the prices forCs outgoing and incoming calls respectively. a and ti

are the bargained accounting rates between target countries and between the host country

and country i respectively. In (4), the rst term represents the prot from selling calls to

CB operators. Each rerouted call implies two outgoing calls from C; each pair of calls is

charged pcb and costs 2cc ta tb to the host carrier.Using the denition of trafc ow between host and target countries, TFic, and

rearranging, it is useful to rewrite carriers A and B prot as follows:

pi pji c coQ

ji a coQ

ij Q

ji

pci c coQci ti co TF

ic Q

ci

where the rst and the third terms represent the revenues from collecting standard calls and

sending them to the called country, while the second and the fourth terms represent the net

interconnection payments with the other national operators. These are known as the

``settlement payments''.

Note that if the access charge exceeds the marginal cost of giving access co, then the

carrier makes money on access (accounting revenues) only if it terminates more calls than

it originates. When volumes of incoming and outgoing trafc are unbalanced, then the

operator which generates more trafc pays for the difference to compensate the

terminating carrier, namely the settlement payment. For example, if a4co then carrier Areceives from B a positive settlement payment ifQab Q

ba40. The reciprocal accounting

rate regime implies that part of the ``high trafc'' carrier's prot is ``stolen'' by the ``low

trafc'' rm through the bargaining process.

It is therefore clear how this system reduces the incentives for a rm to cut its prices:

lower prices imply more outgoing calls and, as a consequence, a lower (even negative)

settlement payment.

2.4. The Timing

We model CB competition as a four stage game; the timeline is in gure 2.

In the rst stage, each monopolist negotiates the reciprocal accounting rates with each

of the other carriers. At this stage a, ta and tb are determined.

Figure 2. The timing.

30 FABIO M. MANENTI


11/21

Given these values, collection prices are dened in the second stage. This sequence of

events is justied on practical grounds: accounting rates are negotiated during periodical

meetings between telecom operators. Retail prices are then adjusted accordingly. In the

third stage CB competition occurs. CB rms exploit arbitrage opportunities between

carriers' prices. This means that CB rms observe the retail prices for international phone

calls charged by carriers and then compete in the market for CB calls. In the last stage,

rms' prots are realized.

3. The Impact of Callback Competition

CB is a textbook example of a market with no cost of entry. Indeed, to set up a CB rm and

to start competing, only switched equipment terminals are required. For this reason, in the

last few years there has been a proliferation of CB companies, especially in the U.S.

Accordingly, we analyze CB assuming price competition with free entry.

3.1. t 2: Callback CompetitionStarting from the last stage of the game, we solve the model by backward induction. At

date t 2, CB competition occurs: CB rms observe the collection prices charged by thecarriers at t 1 and then compete in the market for rerouted calls.

CB rms compete on prices: given the collection prices set by domestic carriers, each

company sets its prices vji;h, with h 1; . . . ; n. The prot function for the h CB rm is

given in (2). Note that at this stage of the game, carriers' collection prices are determined:

Yji;h ? is now function only of the prices charged by CB rms.Since the CB industry is characterized by price competition, then CB calls are priced at

marginal cost. The cost for each rerouted call is a discounted sum of the prices of two

outgoing calls from C; this implies that at the equilibrium:

v vji;h ypbc p

ac i;j A; B and h 1; . . . ; n: 5

We note that: (i) the equilibrium price for a CB call is independent of the direction of the

call, and (ii) since pic is a component of the marginal cost of the CB rms, then an increase

in Cs outgoing prices implies, for given y, a reduction in CB rerouting:

qYji

qpic qY

ji

qvdv

dpic50:

3.2. t 1: Second Stage Collection ChargesAt date t 1 carriers set their prices given the reciprocal access charges a and ti that

they have bilaterally negotiated at t 0. The following proposition presents the secondperiod mark-ups at the symmetric equilibrium.



12/21


13/21

trafc on any other route. With CB part of the trafc ows on one route is moved to a

different route; this makes prices and quantities of calls more correlated than otherwise.

Nevertheless, since at date t 1 retail prices are dened as functions of the negotiatedaccounting rates, in order to analyse the impact of CB services on carriers' collection

prices it is necessary to characterize the behavior of the interconnection charges between

networks. This analysis is conducted in the following section.

3.3. t 0: Accounting Rates

3.3.1. The Accounting Rate Between Target Countries

Interconnection charges (or accounting rates) are dened at date t 0 according to a

Nash-bargaining process. We assume that rms have equal bargaining power. Thefollowing proposition presents our main result:

Proposition 2: When callback rms reroute calls between target countries, the

interconnection charge between these countries increases with respect to the solution

without callback:

a4co:

Proof: See the Appendix. &

This is a surprising result: the presence of CB enterprises located in C pushes up the access

price between A and B. Proposition 2 has a clear explanation. Consider rst the solution

without CB. This situation is close to the analysis of ``symbiotic production'' reported inCarter and Wright (1994). In this case, perfect symmetry between operators A and Boccurs; these carriers have identical demand and cost conditions which implies that, at the

equilibrium, trafc on the route AB is perfectly balanced Qba Qab. As a consequence,

the access is not an issue for the carriers: with reciprocal access price, the interconnection

payment that one monopolist pays to the other is exactly the same it receives for delivering

the other carrier's calls.

Nonetheless the interconnection rate a matters since, according to expression (9), it

affects collection prices. Without CB rms A and B set it equal to the cost of providing

access a co; in doing so each carrier enjoys the highest level of prot (monopolisticlevel) charging the monopolistic price

pM c co

Z

Z 1:

As explained in Carter and Wright (1994), collusion over the access price lowers

collection charges: both the rms increase their prot by lowering the access price. The

cost of providing interconnection co represents the lower bound below which it is not

optimal for both the carriers to set the accounting rate.

Consider now the solution with CB. Since CB acts symmetrically towards target

carriers, the relationship between networks A and B is still symmetric: trafc ow is

perfectly balanced and there is no settlement payment; the accounting rate is still a means



14/21

of collusion. Nevertheless now target carriers deal with the new CB effect: the higher the

price they charge for a call on the route AB, the more the calls rerouted via callback and

the higher the settlement payment that they receive from carrier C.

In other words, since t4co, rerouted calls are attractive because they increase the target

countries settlement payments. Target carriers may agree upon a higher interconnection

charge and, as a consequence, a higher collection price thus stimulating CB and increasing

their access revenues. Moreover, (12) implies that, other things being equal, the

accounting rate between countries A and B increases the more calls are rerouted via

callback.

This result gives evidence on how CB rerouting exacerbates the inefciency of the

current accounting rate regime: CB puts upward pressure on accounting rates that,

otherwise, would have been set at the cost level.

3.3.2. The Accounting Rate Between Host and Target Countries

The characterization of the impact of callback on the interconnection charge between A

(or B) and C is much more complex and cannot be unambiguously dened.

Given the asymmetry between host and target countries, we know from the previous

analysis that the reciprocal interconnection charges ta and tb are set above the cost of

providing access; at the symmetric equilibrium ta tb t and t4co.Nevertheless the interactions between prices, accounting rates and prots are very

complex. The presence of CB affects both the prices between target countries and between

target and host countries making the overall impact of Yon t undetermined.22

3.4. The Impact of Callback on Collection Prices

3.4.1. The Price for a Standard Call Between Target Countries

Applying the results of the previous section to the second stage prices given in section

3.2, we now discuss the impact of CB on retail prices. Let us start with the price for a

standard call between A and B; the following corollary holds:

Corollary 1: The collection price for a call between target countries is increased by

callback:

p c coZ

Z 1 t cor

Y

Q

Z 1

Z 14pM: 10

22 In an extended version of this paper, we show that, under very mild conditions, the accounting rate

increases with the asymmetry in the size of demand faced by the countries involved in the bargaining

process. This argument is conrmed in Wright (1999) where it is shown that the accounting rate between

two countries is higher the more the two countries differ in per capita income: other things being equal,

higher differences in income imply higher trafc imbalances which translate into higher accounting rates.

According to this argument, since CB rerouting increases the asymmetry in the size of demand between

host and target countries, then the bargained accounting rate should be xed at a higher level when CB

rms are in the market.

34 FABIO M. MANENTI


15/21


On the one hand, with CB, both the interconnection charges a and tput upward pressure on

A and B collection prices. As stated in Proposition 2, the access charge between A and B is

higher when CB rms are in the market: the cost per call incurred by each target carrier is

increased with respect to the equilibrium without CB thus putting pressure on retail prices.

In addition, now, even the interconnection charge between target and host countries

enters in the determination of the equilibrium price: all the other things equal, the higher t,

the higher the price for a call from A to B (or vice versa). On the other hand, CB reduces

the demand faced by target carriers thus pushing down collection prices; this effect is more

than compensated by the effect of t on both accounting and collection charges and the

equilibrium price increases with respect to the solution without CB. These arguments arestrengthened by the presence of the term Y/Q in expressions (12) and (10). The term Y/Q

represents the ratio between the amount of CB calls and the amount of standard calls on the

route AB. Ceteris paribus, the larger this ratio the higher the accounting rate bargained by

target carriers and, therefore, the collection price they charge for calls on the route AB.

Together with Proposition 2, this result is the central message of the paper. CB does not

put downward pressure on the accounting rate a between target countries and it does not

put pressure on collection prices pji i;j A; B either. Instead of undermining the

accounting rate regime, the presence of CB tends to expand its distortions. Our conclusion

provides a formal support to the intuition presented in Alleman (1998); in his paper,

Alleman raises many concerns about the widespread view of CB as a device to induce a

reduction in foreign collection rates and accounting charges. For this reason, the author

claims that foreign countries should welcome CB, not make it illegal because it can

improve their monopolists' revenue and prot.

3.4.2. The Price for a Call From Country C to Target Countries

Consider now the impact of CB on the price for calls from host to target countries;

although we don't know the exact impact of CB on the bargained accounting rates between

host and target countries, ti, we can still say something relevant on the impact of CB on

carrier C's mark-up.

Corollary 2: For a given accounting rate between target and host country, the mark-up

for calls that originate in C and terminate in A or B tends to decrease with the price

discount offered to CB rms; in particular:

(i) with full price discount, y 0, CB competition increases the mark-up;(ii) without any price discount, y 1, CB competition reduces the mark-up.


By offering a large price discount to CB companies (low y), the contribution of CB to

carrier C total prot is relatively small. At the same time, a large discount implies that

more calls between A and B are rerouted via C by CB companies. This entails an increase

in the trafc imbalance between C and foreign countries and, as a consequence, a higher



16/21

settlement decit for the host carrier. By increasing pc, the carrier limits the number of

outgoing calls; this reduces the trafc imbalance and the payment due to foreign carriers

for interconnection. However, with little or no volume discount (y close to 1), CB becomes

protable for carrierC, which now fully exploits the increased demand by reducing the

price for a call.23

These observations might become useful in a competitive marketplace. As mentioned,

in some countries, and in the U.S. in particular, prices are lower because of the pressure of

competitive forces. This creates the ideal environment for callback to proliferate. What

could become relevant in a competitive context, is the relationship between the price

charged by the host country carrier and the volume discount offered to CB companies.

Competition in the country C has two main effects. Firstly, it lowers the host country

collection prices which accentuates further the trafc imbalance between host and targetcountries. Secondly, as discussed in Choi et al. (1999), it increases the price discount

offered to CB companies: rival rms compete not only for standard market shares but even

to attract CB trafc: competition reduces y.24 According to our discussion, larger

discounts translate into higher equilibrium prices; this might prevent competition from

achieving its goal, namely driving prices down to costs.

Finally, our model does not say anything about the impact of CB on the price for calling

from target to host country. By assumption, there is no CB service for calls from country A

(respectively B) to country C and the retail price is the standard monopoly price given in

expression (9). Everything now depends on the impact of CB on the accounting rate t

which, as already discussed, is not dened.

4. Conclusions

Callback is an alternative calling procedure that allows customers located throughout the

world to place international calls at lower rates by rerouting calls via foreign and cheaper

23 For the sake of completeness, the endogenous choice ofy by the host carrier is

y e

e 1

1 2e YQc

Z 4e YQc

It is easy to see that y40. For sufciently high levels of the elasticity of demand for CB services, e, a

corner solution occurs, y 1.24 We must note that competition does not necessarily affect the bargaining process over the

interconnection terms. In the U.S., the market for incoming trafc is divided up in proportion to the U.S.

carrier's own share of outgoing trafc with the corresponding country. This follows the FCCs rules called

proportional return rules. These rules were introduced to prevent U.S. carriers competing against each

other for calls termination, when dealing with a monopoly foreign carrier. Such competition could

decrease the bargaining power of the U.S. carriers and it could also lead the incumbent U.S. carrier to

reach an exclusive deal with the foreign carrier, to prevent entry of other carriers into the U.S. market

(wipsawing). As discussed in Wright (1999), these rules imply that U.S. carriers act jointly in the

bargaining over the accounting rate; similarly, a representative host country operator can be assigned

with a monopolistic bargaining power thus keeping the process equivalent to the monopolistic case.

36 FABIO M. MANENTI


17/21

phone lines, usually located in the U.S. Although the international callback market has

been growing at a fast pace in the last ten years, it is still one of the least understood

telephony markets.

The use of a three country model allows us to evaluate the impact of callback rms that

reroute calls between two high price countries via a low price country. Contrary to what

has been claimed by many inuential observers, we show that callback puts upward

pressure on the interconnection rates between the high price carriers; furthermore, we

show that this form of call rerouting might induce a general increase in collection prices.

Transferring this into a real world scenario, we demonstrate that callback is slowing down

the fall in retail and termination charges induced by all those factors, basically

competition, deregulation and new technologies, that currently characterize the

market.This is the effect of the reciprocal accounting rate regime, which is the current system

used for charging and settling inter-country telecommunications. It is now widely

recognized that the reciprocal accounting rate system is inefcient and prevents the

providers of international telephony services from setting adequate interconnection tariffs

and, therefore, retail prices.

Callback rms move telephone trafc from one route to different ones; this makes

trafc and tariffs on different routes more correlated than otherwise. As a consequence,

callback expands the distortions induced by the current accounting rate regime. Our results

are easily understood once the full effects of callback are considered; with callback, the

carriers in target countries receive an extra settlement revenue from the carrier where

callback rms are located. In order to maximize these extra revenues, the carriers agree on

higher reciprocal accounting rates which in turn pushes prices up and stimulates call

rerouting.The effects on prices charged by the carrier of the country where callback rms are

located depend upon the price discount offered to callback rms.

Callback proponents argue that callback helps to bring settlement rates, and therefore

retail prices, more closely in line with costs. We show that the opposite is true.

Paradoxically, the inadequacy of callback rerouting is the effect of what callback was

originally intended to correct: the distortions of the reciprocal accounting rate regime.

Rather then promoting trafc rerouting to bypass the accounting rate regime, it would be

preferable to speed up the current process of reform of the accounting rate regime towards

the introduction of new interconnection arrangements.

Our approach is highly simplied; further research should be devoted to analyzing less

restrictive models in order to provide a more general support to our results and to test

empirically the impact of callback on retail prices. An interesting area for future researchwould be to study the other forms of call rerouting, such as hubbing and reling. These are

additional alternative calling procedures that are frequently considered as instruments to

reduce the inefciencies of the international interconnection system. Other issues that

deserve to be studied concern the impact on traditional telephony ( pricing and

interconnection) of the new forms of communications based on the use of the Internet

and characterized by partially or completely circumventing the current interconnection

regime.



18/21

Appendix

Proof of Proposition 1: Let us rst consider the prices for standard calls on the route

AB. From the denition ofpi the rst order condition with respect to pji is

qpi

qpji

Qji pji c a

qQji

qpji

ti coqY

ji

qpji

0:

Due to the symmetry of the game,25 we can rewrite this expression as follows

p c aqQ

qp

Q t coqY

qp

recalling the denition of the cross-elasticity r and rearranging

p c a

p

1

Z

t cop

r

Z

Y

Q

and expression (6) follows immediately.

The collection price for calls form A (or B) to C solves the following f.o.c.

qpiqpci

Qci pci c ti

qQciqpci

0:

This is a standard monopoly's prot maximizing condition. Given the symmetry it is easy

to derive expression (7).

Finally let us consider the price for an outgoing call from country C; using the fact that

pcb ypic p

jc and rearranging, (4) becomes:

pc ypic p

jc 2cc ti tjY

ji ? Y

ij ?

i;j A;B

pic cc tiQic ? ti coQ

ci ?

differentiating, the rst order condition is:

qpcqpc

ypic pjc 2cc ti tj

qYji

qvji

qvjiqpic

qYij

qvij

qvij

qpic

2 3

yYji Yi

j Qic p

ic cc ti

qQicqpic

0

25 At the symmetric equilibrium pji p

ij p;p

ic p

jc pc and v

ji v

ij v. These conditions imply that

Qji Q

ij Q; Q

ic Q

jc Qc.

38 FABIO M. MANENTI


19/21

since qvji =qp

ic qv

ij=qp

ic y, and given that, at the symmetric equilibrium,

qYji

qvji

qvji

qpicqYij

qvij

qvij

qpic 2y

qY

qv

and that v 2ypc, this expression reduces to

qpcqpc

Qc pc cc tqQcqpc

2yY 4y2pcqY

qv 4ypc

qY

qv

pc cc t4yqY

qv 0

rearranging:

pc cc t

pc1 2

e

Z

Y

Qc

!

1

Z 2

y

Z

Y

Qc 2

e

Z1 y

Y

Qc

and expression (8) is obtained.

Second order conditions are assumed to hold. &

Proof of Proposition 2: Using second stage prices provided in Proposition 1, carriers'

prots can be expressed in terms of the access carges only. Therefore, according to

expression (1) the accounting rate between carriers A and B is given by

a argmax log pa pbaa;p

aba; a

logpb p

baa;p

aba; a

:

Differentiating the argument of the above expression and using the fact that, by the

envelope theorem, qpi=qpi 0, the bargained accounting rate solves the equation

qpaqpab

dpabda

qpaqa

!pb

qpbqpba

dpbada

qpbqa

!pa 0:

In our symmetric environment, this expression reduces to

a coqQ

qp

t coqY

qp

0 11

rearranging

a co t cor

Z

Y

Q: 12

Where t is the interconnection charge between host and target countries.26 By

construction, the trafc ow between the host country C and target country i is



20/21

unbalanced. Then, as we saw before, the negotiated accounting rate between the host

carrierC and the target carrier i is set above the cost of interconnection: t4co; given that

CB calls and standard calls are substitutes r40, it follows from expression (12) thata4co.

Without CB, Y 0 and therefore a co. This proves the proposition.27

&

Proof of Corollary 1: Without CB, the price for a standard call between A and B is the

monopolistic price pM. Instead, from expression (6), the price for the same call when CB

rms are in the market is

p c co a co t corY

Q !Z

Z 1

using expression (12), then p4pM if

t corY

Q

Z 1

Z 140

which is always satised for t co40. &

Proof of Corollary 2. When y 0, (8) becomes:

pc co t

pc

1

Z

1 2e YQc

1 2 eZ

YQc

: 13

Without CB, the monopolist sets the standard mark-up 1=Z, which is clearly smaller than(13).

Fory 1, (8) becomes:

pc cc t

pc

1

Z

1 2 YQc

1 2 eZ

YQc

14

which, fore4Z, is always lower than the mark-up without CB. &

26 Recall that, given the symmetry of the game, at the solution ta tb.27 To guarantee the existence of an internal equilibrium Y40, we implicitly assume that CB rms enjoy a

price discount y, such that there is always room for CB services. In reality the presence of competition in

the host country, which we do not model here, justies this assumption. Competition has two effects: on

the one hand it implies lower prices for standard calls in the host country thus providing bigger arbitrage

opportunities; on the other hand, it induces carriers in country C to compete for rerouted trafc (namely,

by offering greater volume discounts). This additional form of competition drives down further the price

charged to CB companies, hence reducing their marginal cost.

40 FABIO M. MANENTI


21/21

References

Alleman, J. 1998. `Ìnternational Settlements: An Analysis of Rates.'' University of Colorado & Hagler Bailly

Consulting. Paper presented at the 1998 EARIE Conference in Copenhagen.

Armstrong, M. 1998. ``Network Interconnection in Telecommunications.'' Economic Journal 108: 545564.

Carter, M., and J. Wright. 1994. ``Symbiotic Production: The Case of Telecommunication Pricing.'' Review of

Industrial Organization 9: 365378.

Carter, M. and J. Wright. 1999. `Ìnterconnection in Network Industries.'' Review of Industrial Organization 14:

125.

Choi, H. W., K. L. Yun, I. J. Kim, and B. H. Ahn. 1999. `Òn the Economics of Call-Back Services.'' Journal of

Regulatory Economics 15: 165181.

De Fraja, G., and P. Valbonesi. 2001. ``Revenue Sharing Rules for International Telephony.'' Journal of

Regulatory Economics 20: 120.

Frieden, R. M. 1997. ``The Impact of Call-Back and Arbitrage on the Accounting Rate Regime.''Telecommunications Policy 21: 819827.

ITU. 1996. Direction of TrafcTrends in International Telephone Tariffs. Geneva: ITU Press.

ITU. 1999. Direction of TrafcTrends in International Telephone Tariffs. Geneva: ITU Press.

Kelly, T. 1996. ``Call-Back, Cash-Back?'' Global Telecoms Business. August/September 1996 edition.

Laffont, J. J., P. Rey, and J. Tirole. 1998a. ``Network Competition: I. Overview and Nondiscriminatory Pricing.''

Rand Journal of Economics 29: 137.

Laffont, J. J., P. Rey, and J. Tirole. 1998b. ``Network Competition: II. Price Discrimination.'' Rand Journal of

Economics 29: 3856.

Lande, J., and L. Blake. 1997. ``Trends in the U.S. International Telecommunications Industry.'' Report of the

Federal Communications Commission.

Manenti, F. M. 1997. ``The Reciprocal Accounting Rate Regime and the Impact of Call-Back on International

Telephony.'' University of Brescia, Discussion Paper n. 9713, December 1997.

Manenti, F. M. 2000. The Demand for International Telecommunications in Italy. Mimeo: University of Padua.

Nye, S. 1999. ``Call-Back Patents Yields Slim Pickings.'' Communications Week 25 October 1999.

OECD. 1995. ``Rele and Alternate Calling Procedures: Their Impact on Accounting Rates and CollectionCharges.'' OECDCommittee for Information, Computer and Communications Policy.

OECD. 1997. ``New Technologies and Their Impact on the Accounting Rate System.'' OECDCommittee for

Information, Computer and Communications Policy.

Propp, K. R. 1996. ``The Eroding Structure of International Telecommunications Regulation: The Challenge of

Call-Back Services.'' Harvard International Law Journal 37: 493520.

Wright, J. 1999. `Ìnternational Telecommunications, Settlement Rates and the FCC.'' Journal of Regulatory

Economics 15: 267291.


On the Impact of Call-Back Competition on International Telephony

Documents

Transcript of On the Impact of Call-Back Competition on International Telephony