Review of Cost Issues in
International Settlements
David N. Townsend
December 1997
DRAFT
I. Introduction: the question of costs
The center of the policy and economic debates surrounding accounting/settlement
rate reform is the question of costs. There is a broad consensus that settlement payments
– or any new mechanism for inter-country carrier compensation – must become more
“cost-based,” that present rates are well above cost, that market forces will increasingly
drive settlement charges toward cost, even where administrations may wish to forestall
such movement.
The challenge, of course, is to define which “costs” should, in fact, be recovered
through settlements or equivalent charges, and then to measure the appropriate cost
elements accurately.
This is fundamentally a regulatory, and a philosophical exercise. In the long fun,
if market forces truly come to dominate the international telephone service market, the
definition and measurement of costs would be moot, since the market would dictate both
costs and prices. But the prevailing belief is that any transition to a purely market-based
global industry will take a long time, and will come about piecemeal in different regions.
Thus, policymakers and negotiators should seek to agree on methods for determining
appropriate costs to use as a basis for revised settlement rates, as an interim step in the
industry‟s evolution.
“Appropriate,” in this context, ought to mean only those costs that would
ultimately be included in a theoretical market-based settlements regime. In practice,
without knowing precisely how the “market” will eventually play out, “appropriate” costs
are a matter of opinion – both the definition of what types of cost components should be
included, and the methods for measuring such costs.
DAVID N. TOWNSEND & ASSOCIATES 1 DRAFT, DECEMBER 1997
Cost Issues in International Settlements
II. Cost component definition: the ITU framework
The most widely accepted framework for defining settlements-related costs is the
ITU Standardization Sector‟s Recommendation D.140 on Accounting Rate Principles for
International Telephone Services (ITU-T Recommendation D.140). This model defines
three basic operational components (network elements) of international telephone
termination service:
1. International transmission facilities
2. International switching facilities
3. National extension
The ITU recommends identifying both direct costs of international service
associated with each of these elements, and some allocation of “indirect” or common
costs, to be included in the calculation of cost-based settlement rates.
ITU-T‟s TAS Group has also developed a spreadsheet model and a series of
recommendations and models for further assisting national administrations and carriers to
identify and measure the appropriate allocation of each of these cost components.
Following is a brief summary of those guidelines:
1. International transmission. This element consists of the following components:
Earth station
Submarine/terrestrial cable system
Cable landing station
International terrestrial radio links
National links between these facilities and the International Exchange
2. International switching. These facilities include the international switching center and
associated equipment:
International telecommunications maintenance and operations center
Telephone exchange
Associated transmission and signaling equipment
3. National extension. This component is defined by the Recommendation D.140 as
follows:
The national extension, used for international telephone traffic, consists of
national exchanges, national transmission facilities and, if appropriate and
identified under a bilateral or multilateral agreement, the local loop.
The questions surrounding this component are discussed in the next section.
DAVID N. TOWNSEND & ASSOCIATES 2 DRAFT, DECEMBER 1997
Cost Issues in International Settlements
The national extension cost issue
In the end, national extension costs represent the core philosophical question in
this entire debate. It is relatively clear that (properly measured) costs for international
transmission and switching should be compensated, in some manner, through charges for
international calls. Those components are dedicated to international services, more or
less exclusively. The more difficult question is: “What is the „cost‟ to deliver a call to its
end recipient after it crosses from the international gateway into the domestic exchange
network?”
(For additional discussion of this issue, see Section IV, below.)
Unfortunately, the ITU‟s guidelines do not offer much in the way of specifics to
answer this question. In the first place, the ITU hopes that negotiators will simply come
to some agreement on the national extension cost. According to the TAS cost model
guidelines, “it should be possible to identify and apportion costs for the national network
which relate to the extension of the international telephone service only.”
As further guidance, however, the TAS model‟s input instructions suggest two
sets of elements to include in the national extension cost:
A. For combined international and national administrations:
Trunk switches/national exchanges
National transmission facilities
Local loop, if appropriate and identified under bi-lateral/multilateral
agreement
B. For separate international and national administrations:
Payment by international administration to national administration on the
basis of:
Per minute
Annual lump sum
Revenue/Cost sharing (e.g. percentage of international collections), or
Combination of any above three
These clarifications, however, still leave open many fundamental questions. One
of the most controversial is that of local loop costs. For integrated carriers, the guidelines
insist that loop costs should only be included if “appropriate” and mutually agreed.
However, for independent international carriers, it is directly recommended that the full
amount of any interconnection payments to local carriers be part of the termination cost.
There is a risk of inconsistency in these two positions, because domestic interconnection
charges often include explicit or implicit allocations of local loop costs. The instructions
provide no further guidance in resolving this issue, or the means of calculating or
allocating the costs identified.
DAVID N. TOWNSEND & ASSOCIATES 3 DRAFT, DECEMBER 1997
Cost Issues in International Settlements
III. The FCC methodology: Tariffs as surrogate for costs
In its extensive and controversial Order on international settlement “benchmarks”
(Docket 97280), the U.S. Federal Communications Commission devised a method for
approximating international call termination costs for other countries, without the use of
actual cost data. The FCC effectively endorses ITU-T Recommendation D.140, claiming
that its approach is “based on” D.140. In particular, the FCC adopts the three part
classification of service cost components into international transmission, international
switching, and national extension.
The Commission‟s approach to measuring costs in each category has been subject
to considerable debate, with the national extension component again the most
controversial. The main methods and issues in the FCC decision are summarized below:
1. International transmission:
The FCC uses a “Tariffed Component Price” (TCP) method to estimate costs for
the various cable, satellite, and radiowave segments of international transmissions. It
assumes that the tariff prices charged by a country‟s international carriers for international
leased circuits recover at least the cost of such physical connections to the meet-point
between countries. The FCC formula involves dividing the nation‟s highest monthly
price for a 2.048 Mbps circuit (1.544 Mbps in some cases) to the U.S. by an assumed
capacity of 120 voice-grade lines, and by an assumed usage level of 8,000 minutes per
line per month.
Needless to say, these simplifying assumptions have been disputed by countries
that claim their utilization of inter-country capacity does not match the levels used in the
FCC formula, or that the cost characteristics of dedicated lines otherwise do not mirror
those of international voice circuits. The FCC did not obtain any actual usage (or cost)
data from the countries studies, but based its figures on U.S. carrier experience, and
merely assumed the utilization relationships were conservative estimates for worldwide
average traffic patterns and network engineering.
The TCP results for the international transmission component using this
methodology range from 0.7¢ per minute (Mexico) to 25.5¢ per minute (Kenya), with
most countries‟ results falling below 10¢ per minute.
2. International switching:
While acknowledging that reliable data on international switching facility costs
are not available, and that, even if such data were available, an accurate cost estimation
methodology would be difficult to design, the FCC nonetheless devises a cost component
estimate for this network element. In essence, the FCC utilizes the existing settlement
charges among the 42 members countries of TEUREM (essentially the European
continent), as published by the ITU, which are broken down into transmission, switching,
DAVID N. TOWNSEND & ASSOCIATES 4 DRAFT, DECEMBER 1997
Cost Issues in International Settlements
and national extension components. The FCC has chosen to adopt the values of the
switching component of these settlements as a proxy for worldwide international
switching costs. The figures are divided among all countries according to three categories
of economic development, and range from US$0.019 to $0.048 per minute.
The FCC does not venture to explain why it accepts existing European settlement
charge levels for the switching cost element, while ignoring or rejecting the equivalent
TEUREM charges for international transmission and national extension costs. It also
does not address the likelihood that switching costs are lower in most European countries
than in developing countries, which face higher costs of capital and hard currency
requirements for the purchase of foreign manufactured equipment.
3. National extension:
For the national extension component, the FCC again uses a tariff-based proxy
formula for costs of call termination beyond the gateway. In principle, this method
attempts to apply the same mix and level of tariff charges that would be incurred for
domestic calls traveling equivalent routes over the national network. In practice, the
formula involves complex routing and timing distribution matrices, and weighted
averages of local and in-country long distance tariffs to arrive at a unique “tariffed
component price” for each country‟s national extension costs. The results range from a
high of $25.2¢ per minute to a low of zero, for three countries that don‟t charge for
domestic calls on a per-minute basis.
Problems with the benchmark approach
For the past two decades, the FCC has taken an assertive lead in the restructuring
of the U.S. telecommunications industry. Despite consistent challenges and resistance
from multiple vested interests on virtually every major issue that has come before it, the
FCC has steered a generally successful course toward widespread competition, lower
prices, technological revolution, and universal service in the U.S. market.
Unfortunately, in this, its most important and aggressive foray into international
policy reform, the FCC has not lived up to its own standards of analytical rigor,
thoroughness, and objectivity. On the contrary, the methodological and theoretical
shortcomings of the Tariffed Components Price approach to redefining international
settlements implicate a policymaking process that has been hasty at best, and
unnecessarily so.
To begin with, the FCC has chosen the concept of “benchmarks” to attempt to
force settlement rates down to drastically lower levels. The Commission has applied this
concept before, for example in its implementation of cable television rate regulation a few
years ago. The term “benchmark” is also in vogue to describe attempts to arrive at a
national model of costs of universal service. Benchmarks differ from “price caps”
DAVID N. TOWNSEND & ASSOCIATES 5 DRAFT, DECEMBER 1997
Cost Issues in International Settlements
principally in that they are meant to describe average or target price levels for an entire
industry, rather than specific price ceilings for a particular regulated carrier.
By their nature, benchmark prices will not necessarily reflect the actual cost
experience of any given operator. The goal, instead, is to establish an approximate
industry-average cost, and set this as an objective for all operators to move toward. The
Commission took this approach in the cable TV proceeding, calculating what it asserted
were appropriate “competitive” or “efficient” cost levels, which should become the basis
for cable service prices where effective competition did not yet exist. The FCC also
created a range of options for operators to seek exceptions to the benchmark rates, to
reflect unique cost conditions. Essentially, operators could file individual cost studies,
using the basic benchmark formula and their own company-specific costs, to seek
exceptions to the price levels mandated by the industry average costs.
In its settlements benchmarks decision, the FCC has gone even further, attempting
to determine the average costs of telephone services in the entire world, in a single,
accelerated proceeding. Rather than creating a formula for national administrations to use
as a guideline to determine their own costs (as the ITU has done), the FCC presumes to
define precise benchmark costs that should apply across all countries, loosely grouped
into three income levels.
Even if one assumed that the FCC could somehow accurately calculate worldwide
average costs using its methodology, this approach would be clearly inappropriate for
deciding settlement rates that should apply to bilateral relations. By definition,
benchmarks are averages, which means about half of the countries entering into the
calculation would exhibit actual costs somewhere below the benchmark, and the rest
would be somewhere above. To attempt to impose settlement rates equal to the average
upon carriers whose costs are above average is to demand that those carriers, in effect,
incur financial losses on all calls received from the U.S.
The benchmark theory, as applied to U.S. cable operators, is that the industry‟s
cost structure should be largely similar across comparable systems throughout the
country, with allowances for specific extenuating circumstances. If some cable
companies are operating above the benchmark cost, the pricing requirement is intended to
compel them to achieve greater efficiencies, as a competitive market would demand.
This theory is not explicitly repeated in the settlements Order, but it is clearly
assumed in the blanket application of identical benchmark rates across the planet. And
yet, it is the most elementary fact of international economics that cost structures of given
industries do vary, often substantially, from country to country and region to region. This
is not simply a question of management efficiencies in different corporations, but of the
interweavings of the entire fabric of a nation‟s economy: labor costs, capital costs, trade
balances, inflation, geography, demographics, climate, and a hundred other factors.
DAVID N. TOWNSEND & ASSOCIATES 6 DRAFT, DECEMBER 1997
Cost Issues in International Settlements
Take as just one example the difference between Thailand and Russia, using the
FCC‟s own figures. If we accept, for sake of discussion, the tariff-based component costs
determined by the FCC, then the “cost” of completing international calls is 17¢ per
minute for Thailand and 35.4¢ per minute for Russia. The overall average benchmark
level for all countries in the Lower Middle income group is 19¢ per minute. Under the
FCC‟s ruling, then, Thailand would receive settlement payments of 19¢ for every minute
of incoming international traffic, and hence would retain a “profit” from those calls of
some 2¢ per minute. Russia, on the other hand, would receive the same 19¢ per minute
in settlements, but, according to the FCC’s own calculations, its cost is 35.4¢ per minute,
and thus Russia would be forced to absorb a 16.4¢ loss for each minute of incoming U.S.
traffic.
This effect is a self-evident result of the use of broad worldwide averages to
determine country-specific bilateral prices, and is an inherent flaw in the FCC‟s method,
even assuming all cost calculations were accurate. The FCC‟s only response to this
objection is that the tariff rates used to develop proxy costs are likely to be greater than
actual underlying network costs, because retail tariffs recover costs for marketing,
uncollectibles, and similar non-network functions, and thus achieve a margin of comfort
above true incremental costs (para. 48). This reasoning, which the FCC repeats numerous
times in its Order, will be addressed separately below. But it is worth pointing out that
the wide variation in individual country results in the FCC‟s own calculations makes it
highly unlikely that even this “retail cost margin” would make up the roughly 45% loss to
Russia in the above example, or similar disparities for many other countries.
The FCC further notes that its decision to classify countries according to income
levels, and to allow higher benchmarks (and delayed implementation) for poorer
countries, is intended to help mitigate the impact of drastically reduced settlements
revenues on those countries. The use of averages, however, as opposed to country-
specific benchmarks, is explicitly intended to overcome the inherent “inefficiencies” in
those countries telecommunications operations. Again, the Commission‟s method
ignores inherent, real cost differences between countries, and presumes to compel
efficiency improvements around the world by ordering below-cost payments for
terminating traffic.
Problems with the “Tariffed Component Price” methodology for national
extension costs
The criticisms put forth by various commenters of the FCC's assumptions
concerning leased line tariffs as a proxy for transmission costs, and using the ITU's data
for international switching costs, merit much more thoughtful consideration than the
Commission provides in its Order. Even if costs were developed on a per-country basis,
rather than broad averages, these proxy approaches would be suspect at best.
DAVID N. TOWNSEND & ASSOCIATES 7 DRAFT, DECEMBER 1997
Cost Issues in International Settlements
Still, the most troubling aspect of the TCP methodology is clearly its approach to
national extension costs. In this case, it is not immediately clear what standard of policy
the FCC is attempting to apply. On the one hand, the Commission claims to be pursuing
“cost-based” rates, and devising a method for estimating legitimate costs. On the other
hand, it justifies the use of tariffs as the basis for settlement payments by saying it would
be “discriminatory” to charge international carriers rates different from those paid by
domestic customers for the same service.
Apparently, this “non-discrimination” standard dominates the “cost-based”
standard, although the FCC nowhere articulates the distinction. If this is the case,
however, there is no reasonable basis for claiming that international telephone calls
should necessarily be charged precisely the same rates for the national extension segment
as domestic local and long distance calls, in the absence of any cost standard for either
type of usage. Nor is such a practice realistically possible, because each incoming call
would then have to be charged according to its terminating location. This reasoning
makes no more sense than suggesting that a 100 km domestic call in any given country
should be charged precisely 2 times the price of a 50 km call, or 100 times the price of a 1
km call, or else there is “discrimination” among such calls.
In truth, discrimination only enters into the issue if a carrier charges different rates
to different international carriers for the identical termination service. No one outside of
the FCC has seriously suggested that a revised international settlements policy should
include dictating domestic tariff setting policy, and specifically strict parity between
domestic and international charges.
If, on the other hand, we are talking about apply the actual costs of domestic long
distance and local call delivery to international settlements (and doing so on a non-
discriminatory basis), then we are at least still in the territory of the ITU‟s mandates on
the issue. The problem is, the FCC‟s methodology fails to provide a reasonable
measurement of such costs. This is true even if we assume that the Commission‟s
International Bureau has calculated as precisely as possible the distribution of incoming
calls to each country across distances and times of day, and has properly applied the
relevant domestic tariffs in coming up with its weighted average TCPs.
The fundamental flaw is that the FCC assumes that domestic usage tariffs for
most countries tend to reflect closely the pertinent costs of national switching and
transport. The Commission asserts such a relationship, without any supporting evidence,
and in spite of the fact that it is widely accepted throughout international
telecommunications policy circles that most countries‟ domestic tariffs are far from cost-
based. In fact, one of the principal goals of the ITU and other global policy fora in
relation to telecommunications policy reform is to help promote the “rebalancing” of
tariffs, and general movement toward costs. To start with an assumption that such tariffs
inherently provide a reasonable estimate of underlying service costs is to completely
ignore this reality.
DAVID N. TOWNSEND & ASSOCIATES 8 DRAFT, DECEMBER 1997
Cost Issues in International Settlements
Moreover, there is no disagreement about the direction of the mismatch between
most countries‟ domestic tariffs and costs: in the aggregate, the tariffs are below-cost.
This is especially true of local service tariffs: the systematic tendency in nearly all
countries has historically been to cross-subsidize local telephone service with revenues
from other services (especially international services). Domestic long distance prices
follow a less consistent pattern; in many countries they are above cost, and contribute to
the local service subsidy, in others they are lower, and may be subsidized themselves.
The FCC dismisses this vital criticism of its methodology by assuming, without
evidence, that above- and below-cost domestic tariffs will balance each other out:
to the extent domestic local tariff rates are priced below cost, any
underrecovery of costs should be offset by traffic that is distributed to
service classifications such as long distance or international whose tariff
rates are above-cost. (para. 70)
This assumption, however, has no basis in fact. It appears that the FCC is
drawing solely upon the experience of the United States, in which the very great distances
and preponderance of interexchange calling (and high prices) indeed formed the principal
source of local service subsidies for many years. However, anyone familiar with the
tariff, traffic, and revenue patterns of developing countries in particular can verify that
domestic inter-city long distance revenues invariably contribute but a small percentage of
total revenues, not nearly enough to overcome basic local network service deficits. Those
shortfalls are made up principally from international service revenues, including both
settlements and domestic user charges, along with any specialized business services, data
services, cellular concessions, paging, Internet access, and the like. Notwithstanding the
misleading reference in the above quote, the FCC‟s formula takes none of these other
services into account, including international service tariffs. Instead, it produces national
average tariffs solely for local and inter-city charges, which are likely anywhere from
50% to 90% or more below the aggregate unit costs of the services to which the tariffs
apply.
In fact, in some cases, they‟re 100% below actual costs. The most clearly
fallacious element of the FCC‟s national extension formula is that it applies zero cost to
three countries (Kuwait, Barbados, and Hong Kong) which do not have local or national
per-minute charges at all. Rather than remove these obvious outliers from the average, or
apply some substitute, they are incorporated in the calculation as if it truly costs nothing
to transit telephone calls from the international gateway to end users in those countries.
A similar assumption is somewhat hidden under the national extension tariff
calculations for several other countries, which charge domestic long distance rates, but
not local per-minute tariffs. This effect can be seen in the results for Guyana (0.6¢ per
minute), El Salvador (1.1¢), Singapore (0.7¢), Jamaica (1.0¢), and others. No U.S.
carrier regulated by the FCC would tolerate an Order claiming that its systemwide cost
for interconnection and call termination was this low.
DAVID N. TOWNSEND & ASSOCIATES 9 DRAFT, DECEMBER 1997
Cost Issues in International Settlements
The simple fact is that, in addition to cross-subsidies from other services, many
countries choose the same perfectly legitimate policy for local service tariffs as exists in a
large portion of the U.S. market: they include local usage charges in a “flat rate” monthly
basic service subscription tariff. The fact that such tariffs form an integral part of local
service cost recovery in a number of countries is thoroughly ignored by the FCC, which
makes no attempt to incorporate any portion of local subscription charges in its national
extension TCP formula.
The above problems are exacerbated by the fact that the FCC‟s traffic data reveal
that the majority of incoming international traffic in most countries terminates in major
metropolitan areas. Where the international gateway is located in the same city, the
formula only applies local usage tariffs in developing its weighted average costs. To the
extent relevant per-minute local tariffs either don‟t exist, are incorporated in basic
subscription charges, or are heavily subsidized by international and other services, the
formula systematically and substantially underestimates the costs of domestic call
completion for nearly every country in the sample.
This irony of this analysis is that, the most likely, and generally advocated, effect
of reducing settlement payments, and of market restructuring in general, will be a
rebalancing of tariffs that will lead to increases in domestic prices, especially for local
service. To the extent this occurs, the FCC‟s own formula will rapidly become outdated,
as the tariff proxy inputs will be well below the new cost-based tariffs. The FCC has not,
however, proposed revising its formula in the future to take into account changes in the
data inputs. Nor, for that matter, has it considered updating the formula even within the 2
to 3 year time frame by which its benchmarks are to be imposed, to reflect any
rebalancing, or inflation-related or other changes in the very tariffs it presumes to use to
estimate other countries‟ costs.
The nature of costs in developing networks: infrastructure development
As mentioned above, the FCC‟s main response to various criticisms that its
formula understates true costs of international call termination is to point out that retail
tariffs typically cover costs for marketing, billing, uncollectibles, and other activities that
are not relevant to inter-carrier call termination services. Thus, the FCC is not only
confident that any underestimate of costs is compensated, but in fact believes that its
result exceed the true costs of the service (see, e.g., para. 44).
However, the above discussion demonstrates that, at a minimum for the national
extension component, retail tariffs are typically well below costs. Therefore, the theory
that these tariffs recover non-network function costs is incorrect. In any event, the FCC‟s
assumptions about the magnitude of commercial retail costs again reflects a toward a U.S.
style market. In developing countries with monopoly public operators, “marketing” costs
are typically a fairly insignificant aspect of operations (although there may be other
DAVID N. TOWNSEND & ASSOCIATES 10 DRAFT, DECEMBER 1997
Cost Issues in International Settlements
inefficient administrative costs). At the same time, basic infrastructure costs, including
technological upgrades and network expansion, make up a far larger portion of ongoing
costs than they do in mature networks.
These infrastructure costs do not solely consist of so-called “universal service”
obligations, i.e., the extension the access network to unserved customers and regions,
although that is also an important policy question. In many cases, network investments
by national carriers involve replacement of outdated, poorly operating physical plant in
the core network; introduction of digital switching and advanced signaling systems; and
increasing central office and trunk capacity to reduce blockage and improve transmission
quality. All of these activities directly benefit international calling services as well as
domestic services, by making it more likely that calls will go through, enhancing quality
of service, and increasing overall efficiency in the system.
The same can be said, to a large extent, of operations costs in developing
countries. Where services and technology are rapidly evolving, operation and
maintenance support of domestic network functions is especially crucial to assure that
incoming calls are properly processed, that faults are repaired, that problems can be
resolved effectively, and that users can access directory and operator services when
necessary. These are legitimate costs to apportion to international termination service,
not extraneous commercial costs of no consequence to foreign callers.
Again, the bulk of these infrastructure and operations costs have traditionally been
recovered from sources other than local usage tariffs, and disproportionately from
international services. A tariff proxy method that only takes into account domestic usage
tariffs as a measurement of national extension costs cannot adequately compensate
operators, especially in developing countries, for the bona fide services that are provided
to incoming international telephone callers.
Universal service contributions
Finally, the FCC objects to the prospect that U.S. carriers should contribute to
subsidizing universal service obligations in other countries, even though it acknowledges
that such subsidies are “legitimate telecommunications policies” (para. 86). The thrust of
the objection is again the issue of “discrimination,” as the FCC feels that “foreign
termination services from certain countries should [not] be required to finance a
disproportionate share of network costs . . .” (id.).
The policy question of using international settlements revenues to subsidize basic
network expansion in developing countries is largely separate from the previous
discussion about legitimate costs of international call termination, but not entirely so. It
is important to recognize that, on the whole, expansion of telephone access worldwide is
clearly to the economic benefit of international carriers, essentially in proportion to their
share of the global market.
DAVID N. TOWNSEND & ASSOCIATES 11 DRAFT, DECEMBER 1997
Cost Issues in International Settlements
In the case of the U.S., for example, the volume of outgoing calls to, say, Mexico
or the Philippines from expatriates and immigrants is already vast, and, all things equal,
should increase as the penetration of telephones in those countries increases. Settlement
debates aside, those calls are certainly profitable to the U.S. carriers that connect them.
One could argue, therefore, that it is in the economic interest of the U.S. international
telephone industry to continue to contribute, to some degree, to universal service
investments in developing countries. (This is to say nothing, of course, of the U.S. and
other developed country operators that are now part owners of numerous developing
country carriers.)
It is difficult to accept, therefore, that attempts to include universal service
elements in international carrier payments constitute “discriminatory” treatment. Again,
discrimination would occur principally if some foreign carriers were charged more than
others for the same services.
As competition is introduced in many countries for non-basic services, often
including international gateway service, this issue becomes more complicated. As some
commenters to the FCC have pointed out, competing and independent international
gateway operators in many carriers operate under interconnection rules that require
payments to the national network operator of charges that can include universal service
contributions, or otherwise are considerably greater than the FCC benchmark levels. If
the benchmarks are imposed as planned, these carriers could be forced out of business, as
they would lose money on every incoming international call, with little opportunity to
raise outgoing rates to make up the difference. Such an anti-competitive effect of the
FCC policy would certainly not be consistent with its stated objectives.
The FCC has stated that it is prepared to consider exceptions to its benchmarks
where carriers can demonstrate “they do not permit the carrier to recover the incremental
costs of providing international termination service” (para. 74), although this offer is
specifically presented in the context of currency exchange rate fluctuations. It is unclear
what standards the Commission would apply in reviewing applications for such
exceptions, but if the benchmark policy remains unchanged, it should be prepared for a
flood of requests.
DAVID N. TOWNSEND & ASSOCIATES 12 DRAFT, DECEMBER 1997
Cost Issues in International Settlements
IV. Cost allocation options: Toward a theoretical basis for
calculating national extension costs of terminating
international traffic.
If, in some theoretical future, market forces or enlightened regulation drive
telephone tariffs worldwide to the true level of costs, then some tariff-based proxy for
costs might be a reasonable basis for setting international call termination charges. In the
meantime, if there is to be an economically rational basis for such charges, some form of
cost-oriented calculation methodology will need to be established.
Conflicting cost standards for different market structures
The initial question in developing such a methodology is, what should be the
theoretical basis for defining “appropriate” costs of terminating international traffic? The
theories of regulated markets and emerging competition that can be readily applied within
national boundaries are considerably more problematic in a multi-lateral international
environment. A national regulator, for example, can decide how much cross-subsidy to
incorporate in tariff rates, and how fast to open markets to competition, and these
decisions will partly influence which cost methodology will be most appropriate for
interconnection charges.
In a world where many nations are actively moving into a competitive structure,
while others are likely to remain monopolized for the foreseeable future, it is difficult to
prescribe a single standard that will be both fair and efficient.
The preference of the U.S. FCC, and presumably of most policymakers
advocating a market-oriented direction, is to base inter-carrier compensation on
incremental costs: the added cost per minute to provide the service. If there were, in fact,
a competitive market for terminating international calls, the charges for such terminations
would tend toward incremental cost. Were a carrier to attempt unilaterally to impose
charges well above incremental cost, competing carriers could offer lower prices while
still earning a profit, and force the market price back toward the incremental cost level.
The FCC preference, in fact, is for use of the concept of “Total Service Long Run
Incremental Cost,” or TSLRIC, plus a “reasonable” contribution to “joint and common
costs”. TSLRIC is defined as the added cost of providing the entire service, measured
over a period long enough for all input costs to be variable. TSLRIC costs are a purely
theoretical standard used to promote efficiencies in competitive markets, for example in
interconnection charges between carriers. These costs may differ from short-run
incremental costs to the extent network design, technology deployment, and operational
efficiencies might change in the future relative to current experience. In practice,
however, since future costs cannot be known in advance, current unit costs are used to
estimate future unit costs, only with adjustments for known future changes.
DAVID N. TOWNSEND & ASSOCIATES 13 DRAFT, DECEMBER 1997
Cost Issues in International Settlements
In a traditional regulated monopoly market, especially a developing market,
incremental cost is not necessarily the preferred standard for pricing all services. Support
for fixed investment costs of infrastructure expansion and technological upgrades, which
may not be immediately recoverable through service-specific prices1, typically comes
from other services exhibiting the highest value to the customer: so-called Ramsey
pricing. Where regulation supports this type of policy, some form of “Fully Distributed
Cost” (FDC) allocation is often employed, implicitly or explicitly. The monopoly
environment permits this approach, as no competitors can underprice the incumbent, even
where marginal costs are below the regulated price.
It is a fallacy to assume that the FDC monopoly approach is inherently less
economically efficient than a regulated market that requires incremental cost pricing. In
macroeconomic terms, the cross-subsidization of network development and basic service
subscriptions is justified on efficiency grounds if the network externalities resulting from
universal service exceed the “dead weight losses” of suppressed usage demand or supply
resulting from to cost-price distortions. In a fully developed network, such subsidies can
likely be minimal, as the externality effects are already fully realized. But where
penetration rates are below 50%, or even 25%, there is room for huge economic gains
from expansion, and thus FDC and other subsidy-oriented cost methods may be perfectly
legitimate elements of national telecommunications policy.
The problem of international communication is unique in this respect, particularly
for traffic between fully developed and developing economies and networks. This is
because, where competition and incremental cost-based pricing prevail, market forces
will tend to draw demand toward the lower prices, and away from higher priced,
monopoly subsidizing tariffs. This is especially possible with modern technology (hence,
the call-back phenomenon).
Thus, the legitimacy of alternative costing methods for different envionments
comes into direct conflict, not merely because foreign carriers, and callers, are compelled
through settlements to contribute to other countries‟ network development, but because
competitive markets will inevitably be “penalized” for their lower prices, with an
increasing proportion of traffic, and thus subsidy payments, shifting in their direction.
On the other hand, a move to purely incremental cost-based prices for terminating
international traffic in developing countries would potentially undermine broad economic
goals, to the detriment of long-term growth and efficiency objectives, even for foreign
users.
1
The extent to which such subsidies may, in reality, be necessary to permit infrastructure investment, is a
separate analytical question. Monopoly markets have been shown to be inefficient in other ways, including
limiting the degree to which economic forces might promote self-financed network expansion. This is a
domestic policy issue, however. For purposes of international negotiations, it is necessary to accept a
country‟s actual cost and market structure, albeit while adjusting the results as those costs change over time.
DAVID N. TOWNSEND & ASSOCIATES 14 DRAFT, DECEMBER 1997
Cost Issues in International Settlements
The “economically correct” answer, for any particular bilateral pathway, lies in an
impossibly complex calculus of network costs, efficiency gains, externalities, demand
elasticities, and the like. There is some theoretical formula which could capture all of
these factors, including the economic benefit to calling countries of wider access to
called-country populations, and generate a unique (and ever shifting) termination
cost/price for that communication partnership.
Outline of a cost model
In the absence if perfect information, however, we can construct a “second best”
model that at least sets the boundaries for the cost to be associated with international
calling. The elements of such a model should include the following:
1. For the international transmission and international switching
components, the average cost per minute for the use of these facilities to
transmit or receive switched telephone traffic.
2. For the national extension component, the floor cost should be (1) the
incremental cost of domestic network usage. Added to this should be (2) a
contribution to residual joint and common costs that is proportional to the
service‟s share of total carrier operations, plus (3) a contribution
representing the proportional estimated value of network externality
benefits accruing to foreign users due to infrastructure expansion. The
attached model framework illustrates the elements of this calculation, as
discussed below.
How are these elements to be measured in practice? The following sections
elaborate on the proposed method, describing both the theoretical and practical basis for
calculating each of the elements, and sub-elements.
For the international transmission and switching components, the ITU-T
Recommendation D.140, and the associated TAS cost model form an excellent
foundation for this cost analysis. The cost calculations must be country-specific,
however, and not based upon averaging or estimates from outside sources, if at all
possible. Any given country‟s operating costs for its international facilities are real costs,
and foreign carriers using those facilities to bring in traffic should expect to pay their
proportionate share. Just because apples cost 50¢ apiece in one country doesn‟t mean
travelers from that country should expect to buy apples for 50¢ in another country, where
the grocers wholesale cost is 75¢. As long as the average direct costs of international
facilities are accurately measured, even where they may include “inefficiencies,” those
costs are the only valid basis for component prices.
Incremental cost of national extension in developing networks
As for the national extension component, there are theoretical and methodological
issues surrounding each of the elements of that cost definition. To begin, the definition of
DAVID N. TOWNSEND & ASSOCIATES 15 DRAFT, DECEMBER 1997
Cost Issues in International Settlements
the incremental cost of national network usage by incoming international calls itself
raises important theoretical questions. Whether the specific concept of TSLRIC is
relevant here or not is uncertain.
The notion of an “incremental” cost implies a cost that is added to something else,
to a pre-existing network or service. In the U.S. market, where most of these cost
standards originate, the generally accepted framework is of a basic exchange telephone
network, to which national and international long distance have been added. A local
switch, therefore, is part of local service, and only the extra capacity required to
accommodate long distance calls (a small, variable cost) is considered part of the
incremental cost of long distance. The loop is almost entirely a local service cost, or at
most a joint cost.
In many developing countries, it is worth noting that this image of the telephone
network, and its costs, evolving from the center outward is historically inaccurate. A
more common evolutionary construct would see international connections as often the
first network component to be built, especially in countries that remained colonized until
well into the era of telegraph and telephone technology. The installation of international
cables between the colonial power and a landing point in the colony often initiated the
construction of domestic communication facilities, invariably linked back to the original
gateway. Over time the internal network expanded, and other international links were
added, but the legacy of underdevelopment is evidence that the emphasis of
telecommunications investment in many countries was not, historically, domestic
services.
This raises the question: in developing country networks, what is incremental to
what? If the national network has been engineered, designed, and configured initially to
support international communication, then isn‟t its “total service” incremental cost
equivalent to the cost to replicate such a national extension, including tandem switches,
trunks, and even local exchanges?
The issue seems, at the very least, to be open to reasonable debate, and the actual
answer should again differ from country to country. In a forward-looking, long-run study,
one might put aside the historical legacy and anticipate the evolution of the future
network as more inwardly-focused. But again, for many countries, the balance of
international versus domestic demand is likely to remain very different from the U.S.
model, and hence the burden of cost allocation should differ as well.
The only workable answer, therefore, to the question of national extension
incremental costs, is to assume that all minutes of use, throughout the network, depend
equally upon the same infrastructure, and no one usage is incremental or additional to
another. In practice, this requires measuring the capital and operating costs of all usage-
sensitive network facilities (exchange switches, tandem switches, trunks, signaling
equipment, etc.), and dividing by total usage to produce a per-minute national incremental
DAVID N. TOWNSEND & ASSOCIATES 16 DRAFT, DECEMBER 1997
Cost Issues in International Settlements
usage cost. This is shown in the attached chart under Section 1, Incremental national
usage cost.
Joint and common cost recovery
The costs of the local loop, and of basic telephone service operation in general,
are not, in this construct, treated as part of incremental usage costs, even if we consider
that loops are an extension of a network that began at the international gateway. Loops
are, more accurately, a “joint” facility, commonly used by all telephony services,
including incoming international calls, but whose cost is not usage-sensitive. Operating
costs such as commercial services, operator assistance, and so forth (but not marketing
costs), should similarly be considered “common” costs, not incremental costs of usage.
Economic theory would suggest that fixed loop and operating costs should be borne
principally by fixed monthly rental charges. However, since some of the benefit of these
joint facilities and common activities accrue to those who call into a given line, including
international callers, it is also appropriate to recover a portion of their recurring costs
from usage-based charges.
To see this, we should consider the marginal basic service subscriber, whose
presence on the network provides a marginal externality benefit to all other subscribers,
including international users. If a price set equal to incremental cost were just above the
perceived benefit to the subscriber of the service, that subscriber would disconnect, and
the externality to others would be lost. Thus, the collective body of worldwide
subscribers should be willing to pay, through usage charges, the marginal difference
between incremental basic service subscription costs and customer prices, where those
prices are below cost for some customers who otherwise couldn‟t afford the service.
Not knowing what the optimal price for basic telephone service should be, we
must accept actual tariff prices as the best measure of average willingness to pay for basic
service subscriptions. The formula for including joint and common cost contributions in
national extension costs, therefore, should be to subtract gross subscription revenues from
total basic network operating costs (excluding marketing costs), and to divide the
remainder by all minutes of use for all services. This is shown in the attached chart, as
Section 2.2
In any event, it is clear that such an allocation of ongoing basic network costs
represents a “reasonable” portion of joint and common costs, which even the FCC‟s
theory advocates. Pure incremental cost pricing for all services would result in a carrier
being insolvent, and thus all services, even in a purely competitive environment, should
be expected to support such joint and common costs to some degree. The allocation
2
Note: There might be a more sophisticated calculation of this element: If costs, revenues, and traffic could
be differentiated between, for example, business and residential usage, then the relative proportions of
international versus domestic traffic in each category could determine an appropriate weighted component
charge.
DAVID N. TOWNSEND & ASSOCIATES 17 DRAFT, DECEMBER 1997
Cost Issues in International Settlements
suggested here is equitable, and recognizes the inherent benefit in maintaining maximum
subscribership levels.
Support for infrastructure development
An important distinction must be made here between ongoing operational costs
associated with the existing loop network, and the investment costs of new facilities and
expanded infrastructure. When international calls are received, they go to present
network subscribers, and the cost, today, to reach those subscribers is reflected in current
–operating costs, and recurring costs of capital for past investments. The correct long-run
incremental cost of using today‟s network does not include the cost of adding new lines.
(This is not a philosophical distinction, it is merely a precise application of the theory.) In
the future, when the network has, in fact, expanded, its long run incremental operating
costs may or may not increase in reflection of new investments: that will depend upon
factors of technology and usage, among other things, related to the new subscribers.
The final component of the cost model, therefore, must take account of the future
benefits of expanded infrastructure, by allocating a portion of the investment cost to
present users. (The theory unavoidably assumes that present and future users are the
same population.) The net amount to be allocated should be the difference between
average annual investment in loop network growth, and average annual connection charge
revenues from new subscribers. This investment cost does not, in this instance, include
future operation and maintenance costs, as these are already captured in the loop
operation cost calculation above.
Again, the simplest allocation method would be to divide the resulting deficit by
total usage, to achieve an average per-minute capital cost, as shown in the attached chart,
under Section 3.. This would in effect assume that every category of caller, both
domestic and international, receives an equal average externality benefit from access to
every existing and new subscriber line. This is probably not accurate, as the marginal
benefit likely diminishes as the network grows, and the relative value of expansion to
international callers in particular is probably somewhat smaller than to domestic callers,
although there is no empirical evidence to prove this.
It may be most appropriate to leave to bilateral negotiations the precise allocation
of net loop investment costs in international termination charges. The average cost for all
network usage should perhaps be treated as a ceiling, with the possibility of recognizing
lower incremental benefits to specific countries, based upon demonstrated usage patterns,
demand elasticities, and other factors.
Evolution to open markets
In conclusion, we should recognize that this theory is meant to apply to
international telephone service inter-carrier settlements, according to the traditional
regime of separate operators in separate jurisdictions. As that system itself is evolving,
with competitive entry by foreign carriers into domestic markets, joint ventures, and
DAVID N. TOWNSEND & ASSOCIATES 18 DRAFT, DECEMBER 1997
Cost Issues in International Settlements
mergers and acquisitions, the cost elements attributable to connections between telephone
carriers of different countries, or at least different borders of origin and termination, are
likely to evolve as well.
In particular, the point of interconnection between networks may increasingly
vary, from the traditional international half-circuits and gateway switches, all the way
down to the local exchange itself in some cases. Obviously in these situations,
“unbundled” costs for the only facilities used would be the appropriate interconnection
charge standard. The cost methods and data required to fulfill the methodology described
here should help provide a foundation for calculating facility-specific unbundled costs as
well.
However, the theory‟s treatment of joint network costs and network investments
to achieve future externality benefits should still be incorporated in such unbundled
interconnection charges as well. Indeed, even in a fully competitive market, mechanisms
such as Universal Service Funds and other contribution or construction (“play or pay”)
options are certain to remain a central element of telecommunications development
policy, and international correspondent carriers should expect to continue to participate in
those practices, as they will continue to benefit from them.
DAVID N. TOWNSEND & ASSOCIATES 19 DRAFT, DECEMBER 1997
Cost Issues in International Settlements
National Extension Cost Model
Annual
Annual Operation +
Capital costs Maint. costs Total
1. Incremental national usage cost
National trunks
Tandem switches
Local switches
Total
/ minutes of use
= Cost/minute
2. Proportionate share of joint, common costs
Local loop
Administration + commercial
Total
- subscription revenues
= Net cost
/ minutes of use
= Net cost/minute
3. Contribution to infrastructure development
(ceiling)
Annual investment in netwk expansion
- connection chg revenues
= Net cost
/ minutes of use
= Net cost/minute
Total National Extension Cost
DAVID N. TOWNSEND & ASSOCIATES 20 DRAFT, DECEMBER 1997