Professional Documents
Culture Documents
Abstract
q
An earlier version of this paper was presented at the IDEI / CIRANO / PURC International
Conference, Toulouse, October 1617, 1998, the ITS European Regional Conference, Turin, September
24, 1999, and at seminars at Caltech, University of Paris-Evry and University of Lima.
* Corresponding author. Tel.: 133-561-128-589; fax: 133-561-128-637.
E-mail address: gasmi@cict.fr (F. Gasmi).
0167-6245 / 00 / $ see front matter 2000 Elsevier Science B.V. All rights reserved.
PII: S0167-6245( 00 )00016-0
222 F. Gasmi et al. / Information Economics and Policy 12 (2000) 221 248
1. Introduction
1
Another option under consideration in the United States consists of simultaneously introducing
competition in low cost areas and organizing universal service auctions for the subsidies needed in high
cost areas to obtain affordable prices. Note, however, that the good performance of such auctions,
which might turn out to be quite complex, requires a great deal of regulatory expertise and depends
crucially on how genuinely competitive the participants behavior is.
F. Gasmi et al. / Information Economics and Policy 12 (2000) 221 248 223
2
In the case of duopoly, which will be the main focus of our analysis, qij will designate output of
firm j in area i (see below).
224 F. Gasmi et al. / Information Economics and Policy 12 (2000) 221 248
efficient) marginal-cost pricing (mc) and financing with public subsidies (ps) the
implied deficit in each case.3
The cost function of an integrated monopolistic firm serving the whole territory
is
C 5 C( b,e,q1 ,q2 ;N1 ,N2 ) (1)
where b is a technological efficiency parameter belonging to an interval [b, b ] and
] may
e represents an endogenous efficiency parameter (the firms effort) which
4
take any nonnegative value. The value of b is private information to the firm, but
the regulator knows the distribution function F( b ) and its density f( b ). An
increase in the effort variable, e, decreases observable cost, C, but also imposes a
disutility c (e) on the firms manager and workers. The level of effort carried out
will depend upon the regulatory mechanism implemented.
Under UT duopoly, the cost function of the incumbent (firm 1) that serves half
of the urban area and the whole rural area is
UT
S N1
C 1 ( b,e 1 ,q11 ,q2 ) ; C b,e 1 ,q11 ,q2 ;],N2
2
D (2)
where e 1 is the incumbents effort and q11 is the incumbents output in the urban
area. The cost function of the entrant (firm 2) under this scenario is
UT
S N1
C 2 ( b,e 2 ,q12 ) ; C b,e 2 ,q12 ,0;],0
2
D (3)
where q1j and q2j represent firm js output in the urban and rural area, respectively.
Fig. 1 illustrates the generic competitive market structures that will be the focus of
our analysis.6
3
Although the focus of this paper is to compare the performance of alternative competitive
scenarios, for clarity of exposition it is useful to start from a monopoly framework. At this point, the
reader might find it useful to take a glance at Fig. 1 which visualizes the generic market structures that
we consider in the empirical analysis.
4
Since, in our analysis, the urban and rural subscribers populations N1 and N2 are held constant, they
will be kept as arguments of the various cost functions considered only when necessary. We note that
this approach does not allow consideration of the issue of subscriber participation. Strictly speaking, we
must assume that any fixed fees for access to the network do not vary with our entry scenarios.
5
Unless indicated otherwise, we assume that the entrant has the same technological efficiency
parameter b as the incumbent.
6
For reference, the top of Fig. 1 shows a monopolistic situation.
F. Gasmi et al. / Information Economics and Policy 12 (2000) 221 248 225
7
Optimal effort equates marginal disutility and marginal cost saving.
226 F. Gasmi et al. / Information Economics and Policy 12 (2000) 221 248
C UT
1
c 9(e 1 ) 5 2 ]]( b,e 1 ,q11 ,q2 ) (5)
e 1
C 2UT
c 9(e 2 ) 5 2 ]]( b,e 2 ,q12 ) (6)
e 2
where C UT
1 and C 2UT are the cost functions of the incumbent and the entrant
defined in (2) and (3), respectively. In the urban sector, marginal cost of the
entrant imposes the price (and hence output). This price p1 solves
UT
C 2
p1 5 ]]( b,e *2 ,q12 ). (7)
q12
C UT
1
p2 5 ]]( b,e *1 ,q11 ,q2 ). (8)
q2
where we make use of the fact that q11 5 q12 ; q1 . Social welfare to be maximized
with respect to q1 and q2 is then given by
ps
SW UT mc 5 N1 S1 (q1 ) 1 N2 S2 (q2 ) 2 C UT UT
1 ( b,e *
1 ,q 1 ,q 2 ) 2 C 2 ( b,e *
2 ,q 1 )
ps
2 c (e 1* ) 2 c (e *2 ) 2 l[ $ UT mc 1 c (e 1* ) 1 c (e 2* )] (10)
where l . 0 is the social cost of public funds which ranges from 0.3 in the most
developed countries to 1 or even 2 in LDCs.
In scenario TC ps
mc , entry occurs in both the urban and the rural zones. Again,
complete information allows the regulator to impose the (same) optimal level of
effort on each firm, e j** , which solves
TC
C j
c 9(e j ) 5 2 ]]( b,ej ,q1j ,q2j ) (11)
e j
where C TC
j is the incumbents (and the entrants) cost function defined in (4)
above. Marginal-cost pricing in both urban and rural areas amounts to
C TC
j
p1 5 ]] ( b,e j** ,q1j ,q2j ) (12)
q1j
F. Gasmi et al. / Information Economics and Policy 12 (2000) 221 248 227
C TC
j
p2 5 ]]( b,e j** ,q1j ,q2j ) (13)
q2j
Let us now turn to competitive alternatives that incorporate two factors. First,
firms must comply with the universal service obligation, i.e., the obligation to
provide service in (high cost) rural areas at an affordable price, which we take
here as meaning the same price as in urban areas.8 Second, we now impose a
budget balance condition on the regulated firm, and take account of the fact that
public funds are a scarce resource.9
First, we examine once again the framework in which entry occurs in the urban
sector only. Bertrand-like competition in the urban area is assumed to set the price
of urban service at the average cost of the entrant who serves half of that market.
If the incumbent matches this price in the urban area and serves the rural area at
average (remaining) cost, then cross-subsidies going from the urban to the rural
sector are, to a large extent, destroyed, and the incumbent cannot satisfy the
universal service obligation.10 One way to resolve this difficulty is to impose the
urban price in the rural area and finance the subsequent incumbents deficit
through public subsidies. This is scenario UT ps 11
ac which we derive next.
Optimal output of the entrant in the urban sector (which is also that of the
incumbent) q *1 maximizes
N
]1 S1 (q1 ) 2 C UT * ,q1 )
2 ( b,e 2 (16)
2
8
This interpretation of universal service may be oversimplifying to some extent the policy initiatives
in both developed and developing countries. For example, in the United States, universal service is
currently interpreted as the right to purchase a set of services (voice grade plus access to advanced
services) at a benchmark price which does not necessarily equal the price of the low-cost urban area.
However, in many countries universal service entails uniform pricing.
9
In this paper, we designate by balanced-budget regulation a regulation which saturates the firms
participation constraint without transfers from the regulator.
10
In fact, our empirical analysis shows that, in this case, the (residual) rural average cost function is
consistently above the inverse demand function.
11
The subscript ac indicates that average cost pricing is used.
228 F. Gasmi et al. / Information Economics and Policy 12 (2000) 221 248
UT
F UT
2 C 2 ( b,e 2* ,q 1* ) 2 c (e 2* ) 2 l C 1 ( b,e 1* ,q 1* ,q *2 ) 1 c (e *1 )
N1
2 ] p *1 q *1 2 N2 p *1 q *2
2
G (21)
where p *1 5 S 19 (q 1* ) 5 S 92 (q *2 ).
An alternative view of competitive entry with balance budget (as already seen in
scenario TC psmc above) would be to assume that the whole territory is divided in
half. In this scenario, labelled TC ac , social welfare is given by
SW TC ac 5maxhN1 S1 (q1 ) 1 N2 S2 (q2 ) 2 2[C TC ( b,e ** ,q1 ,q2 ) 1 c (e ** )]j
q1, q2
(22)
s.t.
P1 (q1 )N1 q1 1 P2 (q2 )N2 q2 5 2[C TC ( b,e ** ,q1 ,q2 ) 1 c (e ** )] (23)
and
P1 (q1 ) 5 P2 (q2 ) (24)
where the cost function C TC is the one that is defined in (4) and e ** satisfies (11).
Both of the above (competitive) scenarios satisfy the universal service obliga-
F. Gasmi et al. / Information Economics and Policy 12 (2000) 221 248 229
tion, i.e., uniform pricing across the urban and rural areas. However, while TC ac
ps
uses urban-to-rural cross-subsidies, UT ac relies on public subsidies to finance
universal service. Clearly then, the relative attractiveness of these scenarios will
depend upon the cost of public funds l applicable to UT ps ac , in relation to the
distortions created by the cross-subsidies in TC ac . This issue is addressed in our
empirical analysis.
Still another alternative way to finance universal service is by using explicit
taxes or surcharges.12 In fact, since it imposes average-cost pricing, TC ac may be
interpreted as a scenario that imposes a particular (implicit) tax on the low-cost
urban subscribers through price-averaging. Moreover, one can also rely on explicit
taxes applied to the urban sector in a UT-type entry scenario to finance universal
service, i.e., to cover the incumbents deficit from using uniform pricing over the
whole territory. Let us succinctly describe the main features of such a scenario
which we label UT tac .
Let t designate the tax rate applied in the urban sector. Note that since the
incumbent has to match the entrants price in the urban area and apply that same
price in the rural area, we have q11 5 q12 ; q1 and P1 (q1 ) 5 P2 (q2 ). Furthermore,
assuming that P1 represents the after-tax price (which is then applied across the
whole territory), the firms per unit revenue in the urban and rural markets is given
by P1 /(1 1 t ). This yields the following budget balance constraint for the entrant:
where e 1* is the incumbents optimal effort level which satisfies (5) and q2 is its
output in the rural area. Comparing the levels of social welfare achieved under this
12
Excise taxes on telecommunications services have been imposed in the United States at the federal
level and by various state and local authorities. In the implementation of the 1996 Telecommunications
Act, a universal service fund was funded through surcharges on revenues of all telecommunications
carriers as defined in the Act.
230 F. Gasmi et al. / Information Economics and Policy 12 (2000) 221 248
t
scenario SW UT ac with SW TC ac will give us some indication of the relative
t
importance of the distortions associated with implicit (TC ac ) and explicit (UT ac )
taxation, accounting for the fact that, in the latter case, tax money is vulnerable to
corruption.
The final step of the analysis consists of taking account of the asymmetric
information with respect to the firms technologies. There are various ways to
introduce competition in the context of asymmetric information. In order to make
the most favorable case for competitive entry into the urban sector only, we
assume that the entrant has the most efficient technology, corresponding to b, and
]
that the incumbent is regulated as in scenario UT ps ac , i.e., still with complete
ps
information. This scenario will be designated by ] UT ac and social welfare achieved
UT ps
by SW ] ac
. We then compare this scenario with the optimal price cap (PC) with
asymmetric information obtained when the whole territory is divided in half
(scenario TCPC yielding social welfare SW TCPC ).13 This again allows us to assess
the value of cross-subsidies as a means of financing universal service when
asymmetric information is taken into account.
3. Empirical procedure
3.1. The cost structure of the network: an engineering process model approach
13
This price cap regulatory mechanism is discussed in great detail in Gasmi et al. (1999).
14
Gasmi et al. (1997, 1999) have used this approach to analyze the efficiency and welfare properties
of traditional and incentive regulation, and Gasmi et al. (1998) to suggest an extension of the standard
natural monopoly test.
F. Gasmi et al. / Information Economics and Policy 12 (2000) 221 248 231
duopoly by multiplying switching costs by 1.01 and transport costs by 1.12 for
each simulation. Then we used these data to obtain a translog representation of a
duopolistic cost function that accounts for interconnection costs. A similar exercise
was performed to adjust for interconnection costs under a UT-type entry scenario.
We found a multiplier of 1.01 for switching costs and 1.13 for transport costs.
We use LECOM to simulate each of the cost functions which are implied by one
of our scenarios. As a measure of output, we use traffic volume, expressed in units
F. Gasmi et al. / Information Economics and Policy 12 (2000) 221 248 233
of hundreds of call seconds, CCS.16 LECOM allows for the independent spe-
cification of demands for access, switched local and toll usage, and local and toll
private line services. In order to keep the analysis and the number of simulations
within tractable limits, however, we hold the number of subscribers fixed, and we
constrain the other outputs to vary proportionally with our measure of traffic
volume which is measured on a per-line basis.
As a measure of technological efficiency, b, we use a multiplicative shift term
for the price of capital, which is one of the user supplied inputs for LECOM. The
effect of this multiplier is to scale the cost of those inputs that are sensitive to the
cost of capital in a linear manner. In the case where the regulator has some
uncertainty about the firms technology, we use this scale parameter to proxy it.
We believe that this multiplier represents a plausible one-dimensional measure of
technological uncertainty which has a direct impact on all of the technological
variables. Equivalently, if we had detailed access to the internal structure of
LECOM, we could have modeled the technological uncertainty by the variation of
a large number of technological parameters all of which would be constrained to
vary proportionally.
In a similar manner, we use a multiplier for the price of labor as a measure of
the effect of managerial effort on total cost. In this interpretation, an increase in
effort leads to a reduction in the price of labor, which we interpret as an increase
in the efficiency units of labor associated with a given size of workforce. The
underlying assumption is that effort is primarily directed toward efficiently
utilizing labor inputs. This assumption seems plausible given the dramatic
reductions in workforce that have accompanied changes in regulatory regimes
(e.g., when government-owned firms were privatized). Below, we see that this
specification is useful for the calibration of the firms disutility of effort function.
Our (generic) LECOM cost function is therefore defined in terms of three
independent variables: outputs, technology and effort. Holding all other LECOM
inputs fixed we are able to simulate the cost functions by repeatedly running
LECOM for a range of plausible values of the arguments of interest. The end
result of a typical simulation exercise, which may involve as many as 1300
simulation runs, is a detailed map of the cost function. For all of the simulations,
we take as a base case a local exchange market consisting of 50,000 urban
subscribers and 2000 rural subscribers in a total territory covering about 207
square miles.
Since our measures of technological efficiency (or uncertainty when appropriate)
and effort are multipliers for the price of capital and labor, respectively, where the
16
The first C is the Roman numeral for 100. This measure gives, in hundreds of seconds, the time
during which an average line is used per hour. Its relation to the standard engineering measure of
traffic, the Erlang, is given by 1 Erlang536 CCS.
234 F. Gasmi et al. / Information Economics and Policy 12 (2000) 221 248
average price for each input is equal to one, we have chosen to simulate local
exchange costs in the range [0.5;1.5] for both parameters.17 As mentioned above,
output levels in LECOM are measured in units of hundred call seconds (CCS) per
line, per hour, which is the standard North American measure of traffic intensity.
We have simulated costs for values of outputs in the range [2;10].
By appropriately specifying internal LECOM parameters to reflect each of our
assumed market structures, we are able to generate different cost data sets that we
use to estimate a cost function appropriate to each entry scenario.18 More
specifically, we have generated data to estimate the following cost functions: a cost
function for an incumbent who serves the rural area while facing entry in the urban
area, a cost function for an entrant who serves half of the urban area only and a
cost function for an entrant who serves half of both the urban and the rural areas.
In order to evaluate welfare, we must specify both the inverse demand functions
and the cost of public funds l. We use the exponential form which has been widely
used in empirical studies of local telecommunications markets (see Taylor, 1994)
for the demand functions. The pairs of parameters of these demand functions are
determined by relying on two assumptions. First, we assume, in our base case, that
the elasticities of demand in the urban and rural areas are equal to 20.2 which is
standard in the literature on local telecommunications.19 Second, we assume that
revenue collected from the representative customer corresponds to the annual cost
of serving this customer. As to the cost of public funds, we consider values that
range from below 0.3, which is taken to reflect the relative efficiency of taxation
systems in the most advanced economies, to high values above 1.0, which
correspond to the much less efficient taxation systems in developing countries.
Clearly, much less is known about the disutility of effort function c than about
consumer demand, since this function is, by definition, unobservable. Neverthe-
17
This range of values corresponds to cost variations (which, under incomplete information, can be
interpreted as uncertainty on costs due to adverse selection and moral hazard) that could be as large as
30%.
18
We thank M. Kennet for his help in identifying the appropriate LECOM parameters which need to
be modified. For entry scenario UT we simulated the cost functions of both the incumbent and the
(urban) entrant by adjusting the size of the city and its population to represent the territory served by
each. The territory served by the entrant represented one-half of the territory formerly served by the
monopolist with one-half of the monopolists urban customers (and hence the same urban population
density). The territory served by the incumbent represented the entire urban area and rural territory of
the former monopolist, with one-half of the urban area and one-half of the urban population removed.
For the TC entry scenario, we assumed that each duopolist served a territory having one-half of the area
(urban and rural) and one-half of the population (urban and rural) of the former monopolist. All of
these simulations were implemented by appropriately modifying the input files populatn.dat and
rectangl.dat as explained in the LECOM documentation (Gabel and Kennet, 1991).
19
In Section 4.5, we analyze the sensitivity of our results to the demand elasticities.
F. Gasmi et al. / Information Economics and Policy 12 (2000) 221 248 235
less, it is possible to assume a functional form which is consistent with theory, i.e.,
increasing and convex. We calibrate a disutility of effort function of a specific
quadratic functional form. The pair of parameters associated with this quadratic
disutility function is found by relying on two assumptions. First, under cost-plus
regulation, the marginal disutility of effort is equal to zero. Second, under
deregulation, marginal disutility of effort is equal to marginal cost-saving.20
Since we do not have precise information about the regulatory environment, we
assume that the technological efficiency parameter b belongs to the interval
[0.5;1.5] which includes the set of values for which we define our cost functions.
When appropriate, we also assume that the regulators uncertainty about this
technological parameter is modeled by a uniform distribution over this range.
4. Empirical results
Recall that both the urban-targeted entry scenario UT ps mc and the territory-
ps
constrained entry scenario TC mc impose marginal-cost pricing in each of the
(low-cost) urban and (high-cost) rural areas. Clearly then the universal service
obligation to serve both areas at the same price is violated. Nonetheless, a
comparison of these two alternatives will provide us with useful information on
their relative merits on pure technological efficiency grounds.
Because they both rely on costly public funds to finance the deficit, entry
scenarios UT ps ps
mc and TC mc see their social welfare deteriorate with increasing
values of the cost of public funds l. A closer examination of the data, however,
shows that, for low values of l, TC ps ps
mc performs better than UT mc , whereas for high
values of l the UT-type scenario achieves higher levels of social welfare.21 This
improvement of the performance of the urban-targeted scenario relative to the
territory-constrained scenario as l increases may be explained by the relative size
of the deficit, which is found to be somewhat larger for TC ps ps
mc than for UT mc . Fig.
3 visualizes the situation and shows the critical value of the cost of public funds,
l * , above which social welfare achieved under the UT-type scenario is greater
than that achieved under the TC-type scenario. Such a value is found to satisfy
0.4 , l * , 0.5. The following empirical proposition can then be stated:
20
This calibration of the disutility of effort function has been used in our previous work (see Gasmi
et al., 1997, 1999).
21
See Appendix A for the raw data obtained in this empirical analysis.
236 F. Gasmi et al. / Information Economics and Policy 12 (2000) 221 248
Fig. 3. UT vs. TC under marginal cost pricing critical value of the cost of public funds.
targeted entry scenario UT ps mc is strictly higher than that achieved under the
territory-constrained entry scenario TC psmc .
Let us now examine the results for the competitive solutions that satisfy both the
universal service obligation and the balanced-budget constraint.22 Scenario UT ac ps
22
In fact, from now on we will consider only scenarios that satisfy the universal service obligation,
i.e., the obligation to serve the urban and rural areas at a uniform price.
F. Gasmi et al. / Information Economics and Policy 12 (2000) 221 248 237
assumes that the entrant captures half of the urban area and that the incumbent
serves the other half as well as the whole rural area. Under this scenario, the
incumbent has to match the entrants (average-cost) price in the urban area and
apply that same price in the rural area because of the obligation to offer service at
affordable prices. This implies a deficit which is financed from external funds
provided by public subsidies.23 In contrast, under scenario TC ac , which assumes
that the whole territory is divided in half, the universal service obligation is
financed internally through urban-to-rural cross-subsidies imposed by budget
balance in each half of the territory. Consequently, while the performance of TC ac
is independent of l, socially costly public subsidies make social welfare achieved
under scenario UT ps
ac decrease with increasing values of l. Fig. 4 shows the critical
ps
value of the cost of public funds l ** beyond which TC ac outperforms UT ac . This
parameter is found to satisfy 0.1 , l ** , 0.2.
We can then state the following proposition:
Fig. 4. UT vs. TC under average cost pricing: critical value of the cost of public funds.
23
As mentioned in Section 2, we have found that a UT scenario in which the incumbent matches the
entrants price in the urban market and sets price equal to average residual cost in the rural market is
not viable, i.e., there is no rural price that allows the incumbent to break even.
238 F. Gasmi et al. / Information Economics and Policy 12 (2000) 221 248
achieves a strictly higher level of social welfare than the urban-targeted entry
scenario UT psac , in which public subsidies are used to finance the incumbents
deficit due to uniform pricing.
Proposition 2 says that, for countries with values of the cost of public funds well
above the critical value l ** , the territory-constrained entry scenario TC ac is more
attractive than the urban-targeted entry scenario UT psac . In view of the finding stated
in Proposition 1, Proposition 2 has some interesting policy implications. Since our
default value of l for highly developed countries is equal to 0.3, Proposition 2
suggests that the territory-constrained solution can be expected to yield higher
expected social welfare in both developed and developing countries when a
uniform price is imposed in both urban and rural regions. Our analysis suggests
that even highly efficient public funds transfer systems may impose a social cost
that is greater than the deadweight losses associated with uniform pricing regimes,
at least for the demand elasticities that we assume.
To conclude this subsection we note that Proposition 2 conveys an additional
piece of information. Indeed, because scenario TC ac relies on transfers internal to
ps
the firms, whereas UT ac uses public funds to finance universal service, the critical
value l ** provides us with an evaluation of the social cost associated with the
distortions created by the cross-subsidies in scenario TC ac . We note that this
implicit cost of public funds associated with the territory-constrained solution is
quite low.
While scenario UT psac analyzed in the previous subsection uses funds from the
general budget to finance universal service, one might think of a more restricted
fiscal base, such as the urban telecommunications sector, to generate the needed
funds. In fact, as discussed in the Introduction, scenario TC ac (also examined in
the previous subsection) can be interpreted as a scenario which imposes an implicit
tax on the urban customers as average-cost pricing over the territory served by
their operator makes them support some of the high cost of serving rural
customers. An alternative way to achieve uniform pricing across the urban and
rural areas, and within an urban-targeted entry framework, is to impose an explicit
tax on the urban sector and use these tax revenues to cover the incumbents deficit
due to the universal service obligation. This scheme is labelled UT tac .24
In the previous subsection, we have assessed the social cost of the distortions
created by the cross-subsidies within the territory-constrained entry framework
TC ac . Such a cost was found to be estimated by a shadow price of public funds of
24
In fact, one might argue that any difference in the performance of these two alternative ways of
t
financing universal service, through implicit (TC ac ) and explicit (UT ac ) taxation, might be attributed to
the differences in technological efficiency highlighted in Section 4.1.
F. Gasmi et al. / Information Economics and Policy 12 (2000) 221 248 239
the order of 0.1 to 0.2. The nature of the economic distortion associated with an
urban-targeted entry framework using explicit taxation UT tac is two-fold. First,
taxation distorts consumption and one might assess the extent of this distortion by
empirically evaluating the magnitude of the deadweight loss it creates.25 Second,
and the issue is particularly relevant for developing countries, revenues generated
by taxes may or may not be totally used for the purpose of financing universal
service. In particular, institutions that leave substantial discretionary decision
power to executives may open the door to corruption which most probably will
cause a leakage of a nonnegligible part of the tax revenues from the economy.26 In
the empirical analysis, we attempt to take account of this phenomenon by
assuming that a fraction d of the tax revenues, which we allow to vary from 0.5 to
1 in increments of 0.1, is actually used to finance universal service and, therefore,
a fraction 1 2 d of these public funds is embezzled by private interests.
We observe that, in the no corruption state, i.e., when the totality of tax revenues
is indeed used to finance universal service (d 5 1), the performance of scenario
UT tac should be as good as that of scenario TC ac in the absence of a technological
advantage to the TC. Under these circumstances, both of these scenarios can be
expected to create distortions of the same magnitude, i.e., corresponding to a
shadow price of public funds within the range [0.1;0.2]. We have argued above that
the cost of public funds is generally considered to be greater than l * . l ** and
therefore we would expect that UT tac is dominated by TC ac in such countries.
Clearly, decreasing values of d would increase the social cost of the distortions
associated with explicit taxation and hence would favor the territory-constrained
entry scenario TC ac . We have evaluated scenario UT tac for each of the six values of
d and calculated the corresponding values of the implicit cost of public funds l .
Table 1 displays the results which, as expected, show that, as the portion of tax
Table 1
Funding of universal service through explicit taxation (implicit cost of public funds)
d l
0.5 1.358
0.6 0.939
0.7 0.646
0.8 0.431
0.9 0.265
1.0 0.134
25
We approximate this distortion in the usual fashion, namely as the area of the triangle formed by
the demand and supply curves, which measures the excess burden of the tax, i.e., the amount that
consumers and producers are willing to pay to avoid this tax.
26
While the phenomenon of corruption is not peculiar to developing countries, its quantitative
importance (e.g., in terms of the percentage of GNP concerned by the phenomenon), its generalization
to most of the sectors of the economy and the fact that it is chronic, makes it a serious impediment to
their economic and social development.
240 F. Gasmi et al. / Information Economics and Policy 12 (2000) 221 248
So far, even though competition was motivated for informational reasons, the
analysis has been conducted under complete information and we have presumed
that all types of competition eliminate information rents. However, given the
importance of asymmetric information in both developed and developing
countries,28 it is interesting to examine further how it affects our comparisons. We
now investigate further the desirability of cross-subsidies as a means of financing
universal service when the (social) cost of incomplete information is taken into
account. A UT-type entry scenario is considered once more, but we now assume
that the entrant in the urban area has the most efficient technology and, as in
scenario UT psac , we impose affordability through complete information regulation.
This scenario, labelled ] UT ps
ac , is compared with a TC-type scenario in which the
incumbent and the entrant serve half of the whole territory and are regulated under
incomplete information with price cap (TCPC).29 Because the performance of
ps
] ac decreases with l, we find the critical values of this parameter l
UT *** , which is
between 0.3 and 0.4, beyond which the TC-type scenario TCPC dominates despite
incomplete information (Fig. 5).30 This result is stated as the following proposi-
tion:
27
Clearly though, the value of the elasticity of demand, which is quite low in our base case (20.2),
plays an important role in the determination of this deadweight loss. In fact, we have recalculated Table
1 with a higher demand elasticity of 20.6 and found that the corresponding values of the implicit cost
of public funds were much higher even for high values of d.
28
In particular, the fact that there are no satisfactory accounting and auditing procedures in
developing countries appropriately raises the issue of asymmetric information.
29
We perform this comparison to take account of the fact that the UT solution is likely to be more
competitive. It should make the most favorable case for the UT-type scenario in the relevant range of l
and hence would reinforce the case for cross-subsidies if the TC solution still turns out to be the best.
Nevertheless, note that we require a rather efficient regulation in the TC solution (optimal price cap)
which might not be feasible in developing countries.
30 ps ps
Note that because UT ac dominates UT ac in the relevant range of l and TC ac dominates TCPC, we
]
find, as expected, l *** . l ** . Furthermore, note that a comparison of TCPC with UT ps ac rather than
ps
UT ac yields a smaller critical value of l (between 0.2 and 0.3).
]
F. Gasmi et al. / Information Economics and Policy 12 (2000) 221 248 241
Fig. 5. UT (under ac pricing) vs. TC (under pc regulation with asymmetric information): critical value
of the cost of public funds.
This subsection summarizes the main empirical findings of our study and
explores the issue of their robustness to demand conditions. Recall that basically
we have considered two ways of financing the USO (universal service obligation)
in a context where entry into the sector is envisioned, the USO being viewed as the
242 F. Gasmi et al. / Information Economics and Policy 12 (2000) 221 248
efficiency. This was the purpose of the preliminary exercise in which we compare
the marginal cost pricing scenarios UT ps ps
mc and TC mc . .From the standpoint of the
sole technological efficiency, we find that, for low values of the cost of public
funds (values lower than 0.5), the territory-constrained configuration is preferred to
the urban targeted one. Then, we impose the USO and budget balance. A
comparison of the performance of scenario UT ps ac (funding of USO through the
general budget) with scenario TC ac (funding of USO through urban-to-rural
cross-subsidies or equivalently through implicit taxation of the urban sector) turns
out to show that the distortions associated with the TC solution are rather small,
corresponding to a shadow cost of public funds in the 0.10.2 range. This implicit
cost of public funds corresponds (by the equivalence argument discussed above) to
the deadweight loss created by explicit taxation of the urban sector in an
urban-targeted framework (UT tac ) which uses (all of) the tax revenues to finance
the USO.
UT ps ps
A comparison of scenario ] ac (an improved version of UT ac in which the
entrant is the most efficient firm in our grid) with scenario TCPC (which allows
for cross-subsidies, or implicit taxation of the urban sector, under price cap
regulation) yields a shadow cost of the distortions created by the TC solution in the
0.30.4 range. What can we then conclude from this comparison? While under
complete information the distortions associated with the cross-subsidies in the TC
solution turned out to be small, (socially costly) informational rents increase them
quite a bit (from the 0.10.2 range to the 0.30.4 range) and hence make this
solution noticeably less attractive than explicit taxation of the urban sector.
However, we find that, when a fraction of the tax revenues close to 20% is taken
out of the USO financing system, the TC solution again becomes desirable. More
specifically, under a regime with this level of corruption, the global (net) effect of
complete information on the technology of the entrant which is assumed to be the
most efficient (in favor of the UT-type entry scenario), incomplete information (in
favor of the UT-type entry scenario) and pure technological efficiency (in favor of
the TC-type scenario), appears to favor the TC solution as the latter creates
distortions corresponding to a smaller shadow cost of public funds. Fig. 6
illustrates our discussion.31
Let us now conclude this section with a few comments on the robustness of the
results. Abstracting away from revenue effects, one could think of these markets in
terms of the relative availability of communications means which can be good
substitutes for the phone. We consider the case where those substitutes are
practically non-existent (very low elasticity of demand for telephone usage of
20.1) and the case where there exist good substitutes to the telephone service
31
In Fig. 6 we designate by tax-revenues drain factor the proportion of tax revenues which is taken
away from the universal service funding system.
244 F. Gasmi et al. / Information Economics and Policy 12 (2000) 221 248
Table 2
Sensitivity analysis a
Elasticity in l* l ** l ***
rural (h2 )
20.1 0.45 0.11 0.37
20.2 0.45 0.11 0.37
20.6 0.75 0.12 0.36
a
Elasticity of demand in urban h1 5 2 0.2.
Table 3
Sensitivity analysis a
Elasticity in l* l ** l ***
urban (h1 )
20.1 0.35 0.05 0.25
20.2 0.45 0.11 0.37
20.6 1.75 0.37 0.93
a
Elasticity of demand in rural h2 5 2 0.2.
F. Gasmi et al. / Information Economics and Policy 12 (2000) 221 248 245
(very high demand elasticity of 20.6). Tables 2 and 3 give some estimates
obtained through linear interpolation of the critical values of the relevant
parameters (see Propositions 13) for various combinations of the demand
elasticities. Examination of these tables shows that the estimated parameters
remain low for the alternative configurations of demand elasticities.32 This says
that the main policy implication of our empirical propositions still holds, namely
cross-subsidies are more attractive than explicit taxation of the urban sector for the
financing of universal service in developing countries.
5. Conclusion
This paper has contributed to the debate on the financing of universal service in
the context of the deregulation of telecommunications by arguing that cross-
subsidies may still be regarded as a valuable means, in particular for developing
countries. The analysis presented here suggests that this and other important issues
facing the telecommunications industry today can be dealt with in a framework
that combines engineering process models and the tools of modern industrial
organization.
Flexible state-of-the-art technology engineering process models allow us to
circumvent the formidable problem of availability of data in a sector that evolves
at an impressively fast pace. We make use of one of these models, LECOM, to
generate pseudo-data which we use to fit flexible form cost functions. Further-
more, by appropriately specifying some internal parameters of this engineering
process model, we are able to customize the cost data sets to various market
structures and to proxy (asymmetric) information on the technology that regulators
might not possess.
The methodology has been employed to compare the (social welfare) per-
formance of various competitive solutions to the problem of the provision of
telecommunications service in the rural side of a territory composed of an urban
and a rural area. We have considered various types of complete information entry
scenarios; targeted entry towards the (profitable) urban center only (UT-type
scenarios) and entry in both the urban and rural areas (a TC-type scenario). While
the UT-type scenarios analyzed rely on either funding from the national budget or
taxation of the urban telecommunications sector to finance universal service, the
TC-type scenario allows for urban-to-rural cross-subsidies. We have also investi-
gated the impact of incomplete information. The main empirical output of the
analysis is to provide thresholds for the cost of public funds beyond which
32
Note that the case of a very high elasticity in the urban market leads to a somewhat higher value
of l *** , i.e., a higher social cost of asymmetric information.
246 F. Gasmi et al. / Information Economics and Policy 12 (2000) 221 248
cross-subsidies constitute a powerful tool for financing universal service. The main
policy implication of the results is that developing countries often satisfy those
thresholds.
Acknowledgements
Appendix A
This appendix provides the raw data used for the analysis of the competitive
scenarios in our base case. Data used for the sensitivity analysis contained in
Section 4.5 may be obtained from the authors. To help the reader in the
examination of these figures, we provide in the first table the label and a brief
description of each of the scenarios considered in the study. The subsequent tables
give, respectively, the competitive solutions relying on public subsidies, the
competitive solutions relying on cross-subsidies, and the competitive solution
relying on taxation (all with expected social welfare in $100M).
Table A1
Scenario Description
UT ps
mc Urban-targeted entry, mc pricing, deficit financed with ps
UT ps
ac Urban-targeted entry, ac pricing, USO a , deficit financed with ps
UT tac Urban-targeted entry, taxation of urban to finance USO
ps ps
UT Scenario UT ac but with most efficient technology entrant
]ps ac
TC mc Territory-constrained entry, mc pricing, deficit financed with ps
TC ac Territory-constrained entry, ac pricing
TCPC Territory-constrained entry, (incomplete information) price-cap
regulation
a
Universal service obligation.
F. Gasmi et al. / Information Economics and Policy 12 (2000) 221 248 247
Table A2
ps ps ps ps
l SW UT mc SW TC mc SW UT ac SW UT
] ac
Table A3
SW TC ac SW TCPC
1.388 1.380
Table A4
t
d SW UT ac
0.5 1.370
0.6 1.377
0.7 1.381
0.8 1.383
0.9 1.385
1.0 1.387
248 F. Gasmi et al. / Information Economics and Policy 12 (2000) 221 248
References
Frisch, R., 1935. The principle of substitution. An example of its application in the chocolate industry.
Nordisk Tidskrift for Teknisk Konomi, 1227.
Gabel, D., Kennet, M., 1991. Estimating the cost structure of the local telephone exchange network.
National Regulatory Research Institute Report N$ 7810, Institute for Policy Analysis, University of
Toronto.
Gabel, D., Kennet, M., 1994. Economies of scope in the local exchange market. Journal of Regulatory
Economics 6, 381398.
Gasmi, F., Laffont, J.J., Sharkey, W.W., 1997. Incentive regulation and the cost structure of the local
telephone exchange network. Journal of Regulatory Economics 12, 525.
Gasmi, F., Laffont, J.J., Sharkey, W.W., 1998. The natural monopoly test reconsidered: an engineering
process-based approach to empirical analysis of telecommunications. Mimeo, IDEI, Toulouse I,
France.
Gasmi, F., Laffont, J.J., Sharkey, W.W., 1999. Empirical evaluation of regulatory regimes in local
telecommunications markets. Journal of Economics and Management Strategy 8 (1), 6193.
Taylor, L.D., 1994. Telecommunications Demand in Theory and Practice. Kluwer Academic,
Dordrecht.