T
HE
N
ATURE OF
LEC C
OMPETITION IN
L
ONG
D
ISTANCE
20
1. E
CONOMIES OF
S
CALE AND
S
COPE
20
2. T
HE
U
NIQUE
C
OMPETITIVE
R
OLE OF
LEC E
NTRY
22
B. E
XAMPLES OF
LEC E
NTRY
23
V. E
VOLVING
C
OMPETITION IN
L
OCAL
M
ARKETS
25
A. T
HE
N
EED
F
OR
C
OMPETITIVE
P
ARITY
25
B. C
OMPETITIVE
D
ISPARITY
F
ROM
D
ELAY IN
RBOC E
NTRY
26
C. E
MERGING
C
OMPETITION FROM
W
IRELESS AND
C
ABLE
28
1. L
OCAL
C
OMPETITION
F
ROM
W
IRELESS
28
2. L
OCAL
C
OMPETITION
F
ROM
C
ABLE
30
D. E
FFECTIVENESS OF
S
AFEGUARDS
33
-
1
-
I.
S
UMMARY
It is ironic that MCI presents itself as the champion of competition while arguing for continuing
regulatory restriction of LECs resulting in protection of the long distance market. If the long
distance market were as competitive as MCI claims, then it is surprising that MCI is so
concerned about the entry of additional competitors. Competition cannot be threatened by
increasing competition particularly given the framework that the 1996 Act provides to ensure
competitive parity in long distance through separate LEC long distance subsidiary and other
accounting and non-accounting safeguards.
This paper provides a discussion of the points raised by MCI as well as a careful evaluation of
the arguments.
1
Generally we find that MCI has misrepresented facts where it has decided to
comment and omitted material facts from its discussion. Specifically, we find that:
a)
MCI credits long distance competition for large reductions in average long distance
prices which were actually, principally brought about LEC access reductions.
b)
MCI cites estimates of industry aggregate average revenue per minute as evidence of
price reductions, however, price indices show that long distance prices to millions of
customers have increased substantially in the last few years.
c)
MCI claims that currently long distance markets are fiercely competitive, however,
the recent history of IXC pricing behavior demonstrates that the current interexchange
carriers have settled into a comfy oligopoly which denies small customers of the full
benefits of true competition.
d)
MCI claims that RBOC entry into long distance markets would distort and not
enhance competition, however, MCI has failed to recognize the numerous competitive
safeguards created under the Telecommunications Act. In addition, recent events show
that customers have benefited substantially when LECs have participated in long distance
markets.
I. O
N THE
C
OMPETITIVENESS OF THE
L
ONG
D
ISTANCE
I
NDUSTRY


1
We refer to MCIs document, True Competition In The Long-Distance Market, January 27, 1997 at
HTTP://www.competition.mci.com/abo...icpolicy/nfr/press/012797wpd.shtml as the MCI White Paper.
In its White Paper, MCI presents a series of statements about the degree of competition in long
distance markets. MCI claims that competition has taken the place of monopoly control in the
long distance market in the period since divestiture. It claims that competition has brought
substantial benefits to consumers in the form of lower prices, greater choices, and higher service
quality. Upon closer examination of the statements MCI presents, we find that MCI has 2
2
substantially misrepresented facts in its recitation and ignored important features of long
distance markets in its analysis of those markets. In the sections below, we present a summary
of the MCI position and then a critique of that position with a more thorough examination of the
particular issue MCI is addressing.
A.
Analysis of Prices
MCI begins with an analysis of prices, and they purport to look at long distance prices in
several different ways. They claim to examine general price levels, prices of long distance
relative to carrier access and average best prices. We take each of these subjects in turn.
1.
General Price Levels
MCI first begins with a discussion of the level of prices in the long
distance industry since 1985. They claim that the price of a call has dropped by
over 70% and that thanks to competition long distance calls are far cheaper
today than they were in 1984.
In its presentation, MCI has mistakenly characterized reductions in average revenue per
minute as reductions in average prices. Aggregate average revenue per minute are not measures
of price, any more than average revenue per bag of groceries from the local supermarket is a
measure of the price of groceries. In the current telecommunications environment, there are
many reasons to expect average revenue per minute to overstate true price declines. We present
several examples that clearly demonstrate how declines in ARPM during the last several years
almost certainly overstate the actual declines in prices:
a)
Suppose AT&T customers demand ten minutes of message toll service (MTS) for
each minute of wide area toll service (WATS) (and no other services) and that the price
of MTS (per minute) is twice that of WATS. If MTS and WATS prices increase slightly
but demand for WATS grows at 50 percent per year while MTS demand grows at 10
percent per year, then the ARPM of usage decreases by slightly less than two percent. In
other words, ARPM declines despite the fact that both of the component usage prices
have increased.
2

2
This is not merely a theoretical possibility. According to AT&Ts 1994 Annual Report, Although we raised
prices on basic services over the past two years, the shift in the mix of services that customers selected reduced
average per-minute revenues in 1994 and 1993 (at 24). -
3
-
b)
A similar problem arises in the context of volume discount plans. Suppose
the prices in the plan remain fixed, but customers are able to receive lower effective
marginal prices when their demand expands (e.g., because they have installed fax
machines). In that case, ARPM would decline not because the price of usage declined,
but because customer demand increased.
c)
ARPM will also overstate the effect of a price change if the own-price
elasticities for different services are different, even when the percentage price change for
each is identical. For example, suppose (1) the price of service A is one dollar per
minute, ten minutes are sold, and the A own-price elasticity is - 0.2, and (2) service B has
a price of fifty cents per minute, a demand of ten minutes and an own-price elasticity of -
5.0. If the price of each of the services decreases by 10 percent, ARPM will decrease by
17 percent. Observe that the anomalous result is not caused by substitution of lower-
priced servicetheir demands are assumed to be independent in this examplebut
reflects the inadequacies of the index itself.
d)
ARPM (as measured by the IXCs) goes down when facilities by-pass is
initiated by the end-user and will overstate the cost savings enjoyed by the customer. For
example, when a large customer builds a private network perhaps bypassing LEC access
facilities, AT&Ts ARPM from that customer could go down (relative to MTS) but the
cost to the customer of any minutes reflect both AT&T charges (ARPM from AT&Ts
perspective) and its own network costs.
The examples above illustrate two general circumstances producing a tendency for
ARPM to exaggerate recent price reductions. First, when different prices are charged to
different customer groups or for different services, then differences in underlying rates of growth
of sales (whether or not caused by the change in prices) can cause aggregate ARPM to overstate
price reductions. Second, ARPM from any one IXC will misstate end user costs when end users
assemble services through a variety of vendors.
3
Aside from the fact that average revenue per minute does not measure prices, MCI has
miscalculated average revenue per minute. Figure 1 in the MCI White Paper is based on a study
done by Robert Hall in 1993 and revised in 1995, with a 1996 number appended by MCI. The
Hall Study, the source for most of the numbers in Figure 1 has been criticized elsewhere.
4

A major area of criticism of the Hall study revolves around data issues: it is not clear what the numerator,
total long distance revenues, in Halls average revenue per minute is, nor what he has used to estimate the
denominator, the total number of minutes in the industry. One does not need to go past the four corners of the
MCI White Paper to see that the Hall numbers are inconsistent with other estimates of average revenue per
minute. For example, in Figure 1 average revenue per minute in 1992 was estimated by the FCC to be 16.63
cents. This number estimated by the FCC (and the FCC has expressed some concern about the veracity and
accuracy of its calculations in this regard as we discuss below) is very different from the 19.1 cents thats
displayed in Figure 1 which is the estimate produced by Dr. Hall. In the next three years, the two average
revenue per minute series do not even trend together consistently. The discrepancy between the two figures in
1994 is larger in percentage terms than the discrepancy in 1992.

3
Changes in average access cost per minute (AAPM)reflecting changes in payments to LECs will likewise
exaggerate the reductions in access costs that IXCs or their customers have actually realized, when they bypass
LEC facilities. The tendency that we have described for ARPM to overstate price reductions is therefore offset
to some extent by the similar tendency for AAPM to overstate reductions in access charges. At most, the errors
cancel out one another. What is far more likely is that average revenue per minute net of access overstates t