November 30, 2001
HAND DELIVERED
Ms. Felecia L. Greer
Executive Secretary
Public Service Commission of Maryland
William Donald Schaefer Tower
6 St. Paul Street, 16th Floor
Baltimore, Maryland 21202-6806
Re: Case No 8879
Dear Ms. Greer:
Enclosed please find an original and fourteen (14) copies of “An Assessment of
the Competitive Local Exchange Carriers Five Years After the Passage of the
Telecommunications Act” by Dr. Robert W. Crandall. Excerpts of the article are
discussed in the surrebuttal testimony of Verizon Maryland Inc. (“Verizon”) witness John
R. Gilbert filed on October 15, 2001. This document was inadvertently omitted from
Verizon’s filing and in order to provide the Commission with as full and complete a
record as possible, Verizon would like to attach the full document to Mr. Gilbert’s
testimony as Exhibit JRG-3.
In response to a data request from Staff, Verizon has previously provided an
electronic copy of this document to Staff, AT&T, OPC, and Covad/Network Plus.
Verizon is making arrangements to get a hard copy to all parties to this proceeding.
If you have any questions, please do not hesitate to contact me
Very truly yours,
David A. Hill
DAH/mlw
Enclosure
cc:
All Parties of Record
-2-
AN ASSESSMENT OF THE COMPETITIVE LOCAL EXCHANGE CARRIERS
FIVE YEARS AFTER THE PASSAGE OF THE TELECOMMUNICATIONS ACT
ROBERT W. CRANDALL
JUNE 2001
C
R I T E R I O N
E
C O N O M I C S
, L.L.C.
-3TABLE OF CONTENTS
Executive Summary
I.
Introduction and Summary of Conclusions
II.
Local Competition Under the 1996 Telecommunications Act
A.
The Local Access/Exchange Sector
1.
The Incumbent Carriers
2.
The First Entrants--The Competitive Access Providers
3.
Entry Since 1996—The Competitive Local Exchange Carriers
(CLECs)
III.
Entry Strategies
A.
Customers
B.
Services
C.
Network Facilities--Building versus Resale or Leasing
1.
The Short-Run
2.
The Long-Run
D.
Tailoring Investment to Demand
E.
Product Quality
F.
Rate of Expansion
IV.
Industry Competition and Barriers to Entry
A.
Barriers Prior to the 1996 Act
B.
Existing Customer Contracts as Barriers to Entry
C.
The “Last Mile” Problem
D.
Interconnection
1.
Availability of UNEs
2.
Resale Discount Rates
3.
Agreements Between ILECs and CLECs
4.
Reciprocal Compensation
E.
The Cable TV Companies
V.
Empirical Analysis of CLEC Performance
A.
Empirical Analysis of Individual CLECs
B.
The Effect of Network Design and Customer Strategies
VI
Who Succeeded, and Why?
A.
The Most Successful CLECs
1.
McLeodUSA
2.
Time Warner Telecom
3.
Allegiance Telcom, Inc.
B.
The Second Tier of CLECs
1.
XO Communications (Formerly Nextlink Communications)
2.
Intermedia
C
R I T E R I O N
E
C O N O M I C S
, L.L.C.
-4-
VII.
Who Faltered, and Why
A.
ICG Communications, Inc.
B.
CTC Communications
C.
Teligent
D.
NorthPoint Communications
E.
Focal Communications
Conclusion
Appendix 1.
A.
B.
C.
D.
E.
Econometric Analysis of CLECs
Analysis of Individual, Publicly Traded CLECs
Analysis of CLEC Business Models
The Data
Estimation Technique
Regression Results
1.
Controlling for Individual Firms
2.
Analysis of Business Practices
Appendix 2. Articles Citing CLEC Bankruptcy Filings and Acquisitions
C
R I T E R I O N
E
C O N O M I C S
, L.L.C.
-5EXECUTIVE SUMMARY
An Assessment of the Competitive Local Exchange Carriers Five Years After the
Passage of the Telecommunications Act
Robert W. Crandall
The last 16 months have not been kind to most information technology
companies, including the new competitive local telephone carriers (CLECs) that have
formed since the passage of the 1996 Telecommunications Act. These new local
telephone companies’ equities rose sharply during the NASDAQ “bubble” in 1999 and
early 2000 and then declined just as rapidly. Many of the new entrants failed, but a large
number survived as vibrant new competitors in the local telephone business. A detailed
study of these survivors, as well as those that failed, shows that a company’s choice of
business strategy has been the most important determinant its of success or downfall.
Local Telephone Competition is Increasing
According to recent data from the Federal Communications Commission (FCC),
the new competitors controlled 8.5 percent of the local telephone lines in the United
States at the end of last year, double the total that they had in December 1999. Between
1998 and 2000 the revenues of the publicly traded CLECs increased four-fold. Clearly,
local competition is growing.
Three New Local Carriers Stand Out
The most successful of the new entrants are Time Warner Telecom, McLeodUSA,
and Allegiance. Each has contributed substantially to competition, employing different
business strategies. Time Warner has tripled the number of its customer lines since 1998,
and has increased its revenues six-fold during this time. McLeod has shown consistent
quarterly revenue growth of ten percent from 1998 to 2000, and it was one of the largest
of the new carriers with over $400 million in revenues during the fourth quarter of 2000.
In less than three years, Allegiance has grown from scratch to almost $285 million per
year in revenues, and its market capitalization of $1.7 billion is one of the largest in the
industry. These firms prove that a CLEC can succeed.
Business Strategies Determine Outcomes
The new local entrants with solid business strategies thrive, while those with poor
strategies are doomed to failure. Maybe the most important business decision for a CLEC
is its choice of network platform. I found very strong evidence that CLECs are best able
to produce revenue growth by building their own networks or significant parts of their
own networks. CLECs that only resold the established carriers’ services were generally
unable to convert investments into revenues, and these companies were likely to fail.
McLeod has been a stunning exception to the latter rule.
C
R I T E R I O N
E
C O N O M I C S
, L.L.C.
-6Leasing facilities from the established carriers or reselling their services can work
as part of an entrant’s business strategy, as McLeod and Allegiance have demonstrated.
Doing so allows an early jump-start over those building from scratch, but ultimately
revenues grow more rapidly if the entrants build their own networks.
Over-expansion has hurt many entrants, particularly in light of the sharp fall of
technology stocks in 2000-01. Building network components before a customer base has
been established, or providing service before the network is fully functional, places a
strain on capital resources and may eventually lead to failure.
Specific Examples
Time Warner is one of the most successful and thrifty CLECs. In January, it
expanded by purchasing GST, a failing entrant, funding the purchase during a brief
upturn in the market. McLeod and Allegiance are “smart builds.” McLeod takes
advantage of a unique type of resale—reselling US West’s bulk business services.
Allegiance leases the most costly network component—the line running up to a
building—from the incumbents in order to reduce costs. The latter two firms demonstrate
that it is possible to use incumbent companies’ facilities, under terms established by the
1996 Telecom Act, and succeed.
On the other hand, another entrant, ICG, expanded too quickly by adding markets
before its initial network operating problems were eliminated. Ultimately, it filed for
bankruptcy protection, citing service problems and revenue shortfalls. Another entrant,
NorthPoint, sold digital subscriber line (DSL) service to Internet Service Providers (ISPs)
rather than provide Internet access itself. With the recent financial crunch claiming many
Internet firms, many of its customers defaulted on their payments, resulting in
NorthPoint’s filing for bankruptcy protection. Yet another entrant, Focal, relied too
heavily on a gimmick -- collecting reciprocal compensation payments from established
carriers for simply placing itself between these established carriers and Internet service
providers. When this gambit was revealed and ultimately phased out by regulators,
Focal’s inefficient network design was exposed, placing it in substantial financial
difficulty.
A Common Deregulatory Pattern
Opening any market to competition after years of regulation creates enormous
uncertainty. We know from other industries that have been deregulated -- such as
trucking and airlines -- that the ultimate competitive structure of the industry takes years
to sort out and cannot be predicted in advance. When the market is first opened, new
competitors flood the marketplace with little history to guide them. Some succeed; many
fail. Bankruptcies ensue, and after an industry shakeout, strong entrants – such as
Southwest in the airline industry -- are left standing.
The local exchange market is no different. Time Warner, McLeod, and Allegiance
should be around in the long-run, increasing their reach as they add to their networks and
C
R I T E R I O N
E
C O N O M I C S
, L.L.C.
-7purchase companies such as GST, CapRock, Choice One, and Intermedia. In much the
same manner as Southwest has competed against the large airlines, the best CLECs have
flourished by offering a service that is different from that offered by the incumbents.
C
R I T E R I O N
E
C O N O M I C S
, L.L.C.
-8-
I. INTRODUCTION AND SUMMARY OF CONCLUSIONS
Given the recent shakeout of a number of competitive carriers in the local
exchange industry, it is apparent that an analysis of this industry and the competitive
local exchange carriers (CLECs) is required. I have performed a comprehensive analysis
and econometric study of these firms in order to determine (1) the relative performance of
the major CLECs, (2) the CLEC business strategies that have appeared to be successful
thus far, and (3) the causes of CLEC failures during the first five years under the 1996
Telecommunications Act. In particular, I have sought to determine if the recent economic
difficulties experienced by many CLECs have been related to problems in
interconnecting with the incumbent local exchange carriers (ILECs) or to other causes,
including their own business strategies.
An in-depth analysis of the numerous publicly traded CLECs leads me to
conclude that there is wide variation in their performance and that the CLECs’ own
strategies largely explain this variation. However, I can find no evidence linking CLEC
performance to alleged difficulties in interconnecting with the ILECs. Indeed, at least one
major CLEC is succeeding even though it primarily resells ILEC services, and my
general results suggest that a mixed strategy of building one’s own network, in addition
to using ILEC services or network elements, works rather well. Nevertheless, the best
CLEC strategy is to build out one’s own network in a careful, deliberate manner.
II. LOCAL COMPETITION UNDER THE 1996 TELECOMMUNICATIONS ACT
The 1996 Telecommunications Act opened all local telecommunications markets
to competition for the first time. Under the Act, the incumbent carriers are required to
C
R I T E R I O N
E
C O N O M I C S
, L.L.C.
-9negotiate interconnection agreements with the new entrants. They are also required to
offer components of their networks (“unbundled network elements” or UNEs) to the new
carriers at cost-based rates, and to provide their services to these carriers for resale at a
discount from the retail price that reflects the ILECs’ avoidable costs of retailing them.
These provisions allow an entrant to use ILEC facilities or services to access a customer’s
premises, thereby taking advantage of the large ILEC network already in place.
Partly as a result of these provisions of the 1996 Act, a large number of CLECs
now offer local service to business and residential customers.1 CLECs now offer service
on 16.4 million lines, equal to 8.5 percent of the total number of end-user lines in the
United States (See Figure 1).2 Between December 1999 and December 2000, the CLECs
nearly doubled their lines in service.3 Furthermore, publicly traded CLECs reported total
revenues of $7.27 billion in the third quarter of 2000, a four-fold increase from $1.77
1. Trends in Telephone Service, U.S. Federal Communications Commission, December 21, 2000,
stated that 76 CLECs reported operations as of June 2000.
2. Federal Communications Commission, Local Competition Report, (May 21, 2001) (“Local
Competition Report”). See also Credit Suisse First Boston, Telecom Services—Clecs, April 11, 2001.
3. FCC, Local Competition Report, (May 21, 2001).
C
R I T E R I O N
E
C O N O M I C S
, L.L.C.
-10-
Figure 1
Competitors' Share of Access Lines
9
8
Own Net
7
Percent
6
UNEs
5
4
Resale
3
Total
2
1
0
1997-4
1998-4
1999-4
2000-4
Year-Quarter
Source: FCC, Local Competition Reports.
billion reported for the first quarter of 1998.4 Clearly, CLECs have been growing in
numbers and in size since the 1996 Act.
Despite this growth, many CLECs have recently experienced economic
difficulties. At least 10 of them suffered negative earnings growth in 2000,5 and the stock
values of most CLECs have fallen substantially since March 2000. 6
A number have filed for bankruptcy protection or have been acquired as they
faced the prospect of bankruptcy. ICG Communications’ stock fell precipitously between
March and September 14, 2000, when the firm filed for protection under chapter 11 of
4. Data compiled from 10-Q and 10-K filings with the U.S. Securities and Exchange Commission.
Note that total revenue includes money generated from long distance service and hardware sales in addition
to local service.
5. Ten publicly traded CLECs (Advanced Radio, US LEC, CoreComm, GST, ICG, IDT, RCN,
RMI.Net, RSL, and SpeedUS.Com) experienced negative revenue growth in this period.
6. To give specific examples, Adelphia Business Solutions, CTC Communications, Level 3
Communications, and Winstar Communications were trading at 66, 54, 116, and 116 respectively in late
March of 2000, but are trading at single digits one year hence.
C
R I T E R I O N
E
C O N O M I C S
, L.L.C.
-11the bankruptcy law. ICG had recently laid-off 10 percent of its workforce.7 Another
example is Net2000, a firm that has reduced its workforce by 8 percent in the last year,
and whose stock price has fallen from a 52 week high of 47.88 in May, 2000 to a current
price of approximately $1 per share.8 More recently, Northpoint, Covad, Rhythms,
Winstar, and Teligent have either declared bankruptcy or sharply scaled back their
activities in response to recurring losses and the lack of capital. Thus, despite rapid
industry growth since 1996, many CLECs currently find themselves in substantial
financial trouble.
Some observers have attributed the recent poor performance of the CLECs to the
actions of the incumbent local carriers without providing any evidence that these intercarrier relationships are the principal problem for the new entrants.9 What is often left
untold, however, is that not all CLECs are failing. Allegiance, McLeod, and Time Warner
Communications are generally considered by Wall Street analysts to be strong companies
with bright futures.10 Furthermore, the success of these firms relative to other CLECs is
no accident, as they have avoided the pitfalls that have plagued the troubled CLECs, and
engaged in adroit business practices that have eluded others.
In the sections that follow, I describe the local telecommunications sector, and
detail the critical business decisions that face a CLEC both initially and in the long term.
7. Jeff Smith “ICG Communications Files for Bankruptcy,” Denver Rocky Mountain News,
(November 15, 2000) at 1B (“ICG Files for Bankruptcy”).
8. Seth Sawyers, “‘Last mile’ for Many Companies: Linthicum Firm Net2000 Trying not to Become
Telecom Victim,” The Capital(Annapolis, MD), (March 4, 2001) at 8; and NET2000 (NTKK) current stock
price downloaded from Yahoo Finance at (http://finance.yahoo.com/q?s=NTKK&d=v1).
9. See, for example, “The Bell Monopolies Are Killing DSL, Broadband And Competition, And
America is Paying The Price—Part 1,” Broadwatch Magazine, 15(1), at 146.
10. Both Sandra Ward and Yuki Noguchi applaud these firms. See Sandra Ward, “An Interview with
Neil Druker,” Barron’s, (March 19, 2001), at 30-32 (“Neil Druker”); Yuki Noguchi, “Riding up to the
Challenge; 4 Upstart Telecom Companies are picking up where the Bells Left Off,” Washington Post,
(February 28, 2001), at G14 (“Riding up to the Challenge”).
C
R I T E R I O N
E
C O N O M I C S
, L.L.C.
-12I show that the choice of business strategy, not the behavior of the ILECs, is key to the
success of these new companies. It is largely their own business decisions that explain the
difficulties and even the demise of many of these CLECs. Very different business
decisions have made Allegiance, McLeod, and Time Warner successful.
A.
The Local Access/Exchange Sector
For decades, the local-exchange sector was comprised of the Bell Operating
Companies, a few large independent companies, such as GTE, Continental Telephone,
Rochester Telephone, Lincoln Telecommunications, and United Telephone, as well as a
large number of very small, subsidized rural carriers. The AT&T divestiture in 1984 and
the recent 1996 Act have dramatically altered this landscape.
1.
The Incumbent Carriers
The U.S. Department of Justice case against AT&T was settled by a 1982 consent
decree that resulted in the divestiture of the Bell Operating Companies into seven
Regional Bell Operating Companies (RBOCs) in 1984. Each RBOC was allowed to
provide local service, but each company was barred from providing long distance service
outside of its regional operating area, manufacturing equipment, or providing information
services.11 The 1996 Act was passed partly in response to the growing frustrations of the
RBOCs and their inability to obtain relief from the line-of-business restrictions of the
1982 decree. In return for opening their local markets to competition, the RBOCs were
promised access to the long-distance market. They would have to satisfy a complicated
“check-list” of market-opening requirements and pass three levels of regulatory review
for each state in order to gain such access. However, local and intrastate long-distance
11. A 1991 court decision allowed the RBOCs to provide information service, such as voice mail, but
their main operations were still local exchange.
C
R I T E R I O N
E
C O N O M I C S
, L.L.C.
-13rates remained regulated under the 1996 Act. In general, state regulators have kept local
rates for rural areas and many urban residences below cost, while business rates in urban
areas have been kept substantially above cost.
Since 1996, there has been substantial consolidation among the ILECs. SBC
acquired Pacific Telesis in 1996, Southern New England Telephone in 1997, and
Ameritech in 2000. Bell Atlantic acquired Nynex in 1997 and GTE in 1999. The result
has been a decline in the number of large ILECs (See Table 1).
Table 1.
Access Lines and Current Market Capitalization of
Incumbent Local Exchange Carriers (ILECs)
ILEC
Total Access Lines,
12/31/1999
(Thousands)
Market Capitalization, May
2001
($ Billion)
22,539
142
26,517
2,884
28,307
[Included in SBC]
[Included in SBC]
[Included in SBC]
50,464
149
23,354
31,444
[Included in Verizon]
75.9
25,647
63
10,696
18.4
1,162
5.7
1,045
13
822
18.2
Southwestern
Bell(SBC)
Pacific Telesis
Southern New England
Ameritech
Bell Atlantic
(Verizon)
GTE
Bell South
US West
(Qwest)
Sprint
Broadwing
(Cincinnati Bell)
Frontier
(Global Crossing)
ALLTEL
Source: Yahoo Finance, and “Statistics of Communications Common Carriers,” FCC, (December, 31,
1999), Tables 2.1 & 2.6.
None of these RBOC acquisitions affected local market concentration because
these firms were in separate geographic markets prior to the mergers. Nor does the larger
size of two of the three remaining RBOCs provide them with any particular advantage in
C
R I T E R I O N
E
C O N O M I C S
, L.L.C.
-14repelling CLEC entry. These companies are no larger, in terms of total assets, than many
other telecommunications companies, such as AT&T, British Telecom, or Nipon
Telephone and Telegraph. Moreover, mere size offers no particular advantage by itself,
as United Air Lines and American Airlines have discovered in competing with Southwest
Airlines, or as General Motors has discovered in competing with the rest of the auto
industry.
2.
The First Entrants—The Competitive Access Providers (CAPs)
Prior to the 1996 Act, some companies began to enter the local exchange sector in
urban areas. These competitive access providers (CAPs) took advantage of the regulated
rate structure imposed on the ILECs. These CAPs largely avoided the residential markets.
Instead, they entered the urban markets in business districts, providing local exchange
services to medium and large-scale businesses and special-access services to the longdistance carriers. As a result, the CAPs’ impact on the local exchange market as a whole
was limited, although many urban markets for mid-sized and large business customers
were quite competitive prior to the 1996 Act. WorldCom and AT&T, the two largest new
entrants into local telephone service markets, acquired the largest and most successful
CAPs.
3.
Entry Since 1996—The Competitive Local Exchange Carriers
(CLECs)
As a result of the 1996 Act, a large number of new competitive local exchange
carriers (CLECs) entered the local exchange market, using a variety of different
approaches. The proliferation of CLEC entry roughly corresponded with the sudden
emergence of the Internet; therefore, some CLECs focused their businesses around
offering Internet access to some combination of businesses and residences. Other CLECs
C
R I T E R I O N
E
C O N O M I C S
, L.L.C.
-15took advantage of the resale rules in the 1996 Act and provided an array of local
exchange services to both residences and businesses. In short, wide variation existed in
the services that CLECs decided to provide.
Table 2 provides a list of the CLECs that have been publicly traded and their
market capitalization in December 1999, before the sharp decline in the prices of
technology stocks, and on May 1 of this year. In addition, wherever possible, Table 2
provides estimates of the number of access lines captured by each CLEC by the end of
1999 and the second quarter of 2000.
C
R I T E R I O N
E
C O N O M I C S
, L.L.C.
-16Table 2. CLEC Access Lines and Market Capitalization, 1999 and 2000-01
CLEC
1999
Access Lines
(000s)
Adelphia Business Solutions
Advanced Radio
Allegiance Telecom Inc.
Allied Riser
Avista Corporation
CapRock
Choice One
Concentric
Convergent Communications Inc
CoreComm Ltd.
Covad Communications Group Inc.
CTC Communications Corp.
Cypress
e.spire Communications Inc.
Elec Communications
Electric Lightwave Inc.
Focal Communications Corp.
GST Telecommunications
ICG Telecommunications Inc.
Intermedia Communications
Inter-Tel Inc.
ITC DeltaCom Inc.
Ixnet Inc.
McLeod USA Inc.
Mpower Communications Corp.
Net 2000
Net2Phone
Network Access
Network Plus CP
North Pittsburgh
Northern Optic
NorthPoint
Pac West
Primus
RCN Corp.
Rhythms
RMI.Net
RSL
SpeedUS.Com
Teligent Inc.
Telocity
Time Warner TLC
Universal Access Inc
US LEC Corp
USOL Holdings
Winstar Communications Inc.
World Access
XO Comm. (Nextlink)
ZTEL
331
NA
241.7
NA
NA
40.4
NA
NA
NA
NA
57
NA
NA
NA
NA
NA
311.5
NA
731.0
501.1
NA
101.5
NA
579
142.7
55.9
NA
NA
66.7
NA
NA
23.5
NA
NA
223.0
12.5
NA
NA
NA
280
406.2
192.4
NA
409.2
NA
618
NA
428.0
NA
Market
Capitalization
(12/31/99)
6.45
($billions)
2000/2001
Access Lines
Market
(000s), 2000-2nd Capitalization
(05/01/01)
Quarter
496
0.60
($billions)
0.94
6.97
NA
0.71
1.25
NA
0.00
0.47
2.80
6.64
0.69
0.00
0.32
0.05
0.95
1.48
0.00
0.98
2.12
0.79
1.70
0.00
11.79
1.90
0.00
NA
1.55
1.30
0.21
1.17
3.21
NA
2.00
3.97
2.46
0.23
1.04
0.10
3.93
0.00
5.28
NA
0.89
0.06
4.63
1.42
15.19
1.36
NA
394.3
NA
NA
82.6
NA
NA
NA
NA
136
NA
NA
NA
NA
196.3
502.3
NA
1113.0
636.1
Na
148.4
NA
722.9
200.7
85.7
NA
NA
112.9
Na
Na
63.2
NA
NA
352.2
67.0
NA
NA
NA
358.6
461.3
NA
NA
461.3
NA
804
NA
627.2
NA
NA
2.04
0.04
0.95
Na
0.32
NA
NA
0.015
0.17
0.16
0.03
0.007
0.01
0.14
0.40
NA
NA
0.92
0.24
0.34
NA
5.37
0.13
0.08
0.47
0.04
0.20
0.17
0.16
Na
0.11
0.09
0.38
0.02
NA
NA
0.02
0.04
0.18
5.79
0.43
0.15
0.14
0.01
0.02
1.46
0.15
Notes: NA = not available; Sources: Yahoo Finance, Credit Suisse First Boston, Individual
Company Websites.
Note that these companies had a total market capitalization of approximately $95
billion as of December 31, 1999. This was considerably more than the market
capitalization of the Big Three U.S. auto producers, and about three times the
C
R I T E R I O N
E
C O N O M I C S
, L.L.C.
-17capitalization of the entire airline industry. Surely, this was an astounding valuation for a
set of firms that had less than 5 percent of the local exchange telecommunications market
in 1999—a market that generates about $110 billion in revenues per year. To sustain such
a valuation, these companies would have had to displace a large share of ILEC access
lines and to achieve positive (and growing) cash flows in a very short period of time. For
instance, assuming that an access line is worth $3,000 in the current marketplace, this
valuation is equivalent to that of an established ILEC with nearly 32 million lines -- or
roughly four times the number of lines held by all CLECs as of December 31, 1999.
By May 2001, the value of these firms had fallen to $28 billion -- still
comparable to the market capitalization of the entire airline industry -- but 70 percent less
than 17 months earlier. Moreover, as Table 3 shows, a number of these new entrants had
become insolvent, were consolidating and reorganizing, or had filed for bankruptcy
protection. Nevertheless, the new entrants are still increasing their market share, as
Figure 1 above shows.
C
R I T E R I O N
E
C O N O M I C S
, L.L.C.
-18Table 3.
Bankruptcy Filings and Consolidation of CLECs
CLEC
Occurrence
Date
Allegiance Telecom Inc.
Considering Purchase of Choice
April, 2001
One
CapRock
Purchased by McLeod
December 7, 2000
Concentric
Purchased by Nextlink
January, 2000
Convergent Communications Inc
Lays off 500 employees, sale of
April, 2001
assets
CoreComm Ltd.
Lost $318 Million, plans cut in
April 14,2001
operations
Covad Communications Group Inc.
Expects Nasdaq delisting
May 7, 2001
e.spire Communications Inc.
Filed for Bankruptcy
Protection
March 22, 2001
GST Telecommunications
Filed for Bankruptcy Protection,
January 10, 2001
Acquired by Time Warner
ICG Telecommunications Inc.
Filed for Bankruptcy Protection;
May 2001
IDT increasing holdings of ICG
Intermedia Communications
Being acquired by World Com.
March, 2001
Inter-Tel Inc.
Purchased Convergent’s
integration operations
January, 2001
NorthPoint
Filed for Bankruptcy Protection,
March 22
Purchase by AT&T approved
Teligent Inc.
4 of the 7 board seats owned by
May 4, 2001
IDT
Teligent Inc.
Filed for Chapter 11 Bankruptcy
Telocity
Acquired by Hughes for $178
April 3, 2001
million, 82% less than IPO.
Winstar Communications
Filed for Bankruptcy Protection
voice
May 21, 2001
April 18, 2001
Sources: See Appendix 2.
Moreover, the more successful CLECs continue to have a substantial market
value. Six of the largest of these companies had a market value per line – that is, market
value of common equity plus book value of long-term debt divided by estimated access
C
R I T E R I O N
E
C O N O M I C S
, L.L.C.
-19lines in service -- that ranged from $3,600 to $12,700 on May 1 of this year (See Table
4). By contrast, the three largest ILECs had market values of $3,100 to $4,200 despite
their large investments in wireless services, foreign operations, and other activities. From
this comparison, it is quite clear that the financial markets still place a growth premium
on these more successful CLECs.
Table 4. Market Value Per Access Line
Company
Mkt Value per
Switched Access
Line
RBOCs
SBC Communications
Verizon
BellSouth
CLECs
Allegiance
Intermedia
McLeod USA
RCN
Time Warner
XO Communications
$3,100
$3,600
$4,200
$3,600
$5,100
$9,700
$12,500
$12,700
$6,900
Source: Yahoo Finance; “Statistics of Communications Common Carriers,” FCC, (December, 31,
1999), Table 2.6; “Telecom Services—CLECs,” Credit Suisse First Boston Corporation, (June 5,
2001).
A large share of the recent growth has come from the deployment of UNEs, in
addition to the continued construction of the CLECs’ own networks. Much of the
increase in the use of UNEs, however, was likely due to the use of the so-called “UNE
Platform” in New York -- essentially the rental of all the unbundled elements required to
deliver local service. Verizon agreed to make the entire platform available in New York
at the low UNE rates when it (as Bell Atlantic) acquired Nynex in 1999. Of the 8.1
million new CLEC lines in 2000, 1.6 million were obtained in New York. CLECs now
C
R I T E R I O N
E
C O N O M I C S
, L.L.C.
-20have fully 20 percent of all access lines in New York, relying heavily on Verizon’s
existing network to reach their customers.
III. ENTRY STRATEGIES
Entry into the local telecommunications market requires a number of key
decisions. The entrant must target its customers, decide on a bundle of services to offer,
plan its network deployment, and raise sufficient capital to survive the start-up phase.
A.
Customers
Given the regulated rate structure in most states, the obvious targets for the
CLECs -- like the CAPs before them -- are the mid-size and large business customers. A
very large number of the CLEC access lines are urban business lines, particularly in the
larger metropolitan areas. According to the FCC, in December of 2000, 40 percent of
CLEC end-user lines were extended to residential and small business customers.12 As of
this same date, CLECs served only 4.6 percent of the country’s small business and
residential customers, but controlled more than 20 percent of all medium and large
business lines.13
B.
Services
To attract customers from their existing local carriers, CLECs likely must offer
more than the same services and the promise of comparable service quality to that
provided by the existing local carriers. The typical local residential bill is only about $34
per month. Even if an entrant could offer service at a cost below the current regulated
residential rates, the potential savings to residential subscribers would be so small that
12. FCC Local Competition Report, (May 21, 2001).
13. Id.
C
R I T E R I O N
E
C O N O M I C S
, L.L.C.
-21few customers would be likely to change providers unless the new service offered
something different from the incumbent’s traditional services.
Fortunately, the CLECs had the advantage of commencing service during the
dawn of the “Internet age.” As a result, a new CLEC can offer a variety of services to
customers in addition to simple voice telephony. It can bundle local telephone service,
long distance service, high-speed data service, Internet connections,14 and other services,
including standard video services.15
Typically, a CLEC will offer some combination of the above services, but in a
few cases, a CLEC may choose to specialize in providing data services. For example,
Allied Riser derives 100 percent of its revenue from the provision of high-speed data
access to businesses. On the other hand, in the first quarter of 2001, McLeodUSA
reported total revenues of $433 million,16 of which approximately 28 percent was from
local calls, 20 percent was long distance, 16 percent was data service, and 36 percent was
“other,” such as voice mail, equipment sales, and telemarketing.17 Furthermore, CLECs
vary in the extent to which they provide service to residential customers and businesses.
Using the above two examples, McLeod serves both residences and businesses in the
mid-west and Rocky Mountain regions, whereas Allied Riser’s only clients are
businesses in heavily populated cities.
14. Data service usually refers to access to the Internet. Data service could simply be an additional
local phone line, with which the customer accesses the Internet via dial-up modem, or it could refer to the
installation of high speed digital subscriber lines (DSL) that allow simultaneous voice and data transfer so
that the customer is able to use their phone and Internet connection with one line. In this situation, phone
services could be classified under a data heading, because the services are being offered together through
one line. Data service could also be listed under local phone service if the user is accessing the Internet
over her local phone line via dial-up modem. Finally, the heading of “data service” can also encompass
video conferencing.
15. “Other services” could be cable television, voice mail, or web design.
16. “Telecom Services—CLECs,” Credit Suisse First Boston Corporation, (June 5, 2001) at 62.
17. Id at 11.
C
R I T E R I O N
E
C O N O M I C S
, L.L.C.
-22The choice of service to supply should depend on the expected profits to be
garnered from that service. Ideally, a CLEC will wish to offer a service that the ILECs
have either not been providing at all, or have been providing at an unsatisfactory level,
because higher profit margins are available for other services. If a CLEC is deciding
between offering high-speed data service to business or offering local phone service to
residents, it must assess the levels of competition and quality at which those services are
currently offered.
High-speed data service to small and medium sized businesses is a relatively new
and rapidly growing service that has been an attractive vehicle for CLECs as use of the
Internet expanded. These services provided high margins at the time the Act was passed
because CAPs had not aggressively built out their networks to serve such customers. Not
surprisingly, numerous CLECs decided to provide this service to businesses and, as a
result, have been competing aggressively against one another. This competition, while
good for consumers, places downward pressure on rates, and the entrants’ profits.
C.
Network Facilities—Building versus Resale or Leasing
Among the most important decisions facing any CLEC is the choice of network
design. Networks may be a combination of fiber optics, coaxial cables, and copper wires,
or they may be based on a fixed wireless technology. In addition, given the provisions of
the 1996 Act, an entrant may build its own facilities, resell the incumbent’s services, or
use a combination of its own facilities and incumbent’s unbundled elements to deliver its
services.
C
R I T E R I O N
E
C O N O M I C S
, L.L.C.
-231.
The Short-Run
Initially, a CLEC’s choice between on-net facilities, unbundled network elements
(UNEs), or total service resale (TSR) facilities will depend on its ability to construct a
network and the relative prices of UNEs and TSR. On-net facilities are network elements
that a CLEC constructs itself, and are thus wholly owned by the CLEC. An on-net facility
component could consist of a few lines that the CLEC runs from the ILEC network to the
consumer, or it could consist of a whole series of lines and switches that the CLEC
develops in order to provide a service completely independent of the ILEC network. Such
facilities require substantial investments in capital and time to build.
UNE facilities have emerged as a result of the FCC’s interpretation of the 1996
Act. They are network pieces or elements, such as local loops, switching facilities, and
transport facilities that a CLEC leases from an ILEC at negotiated or arbitrated wholesale
prices. As a result of the FCC’s rulings under the 1996 Act, virtually all ILEC network
components have been made available to the CLECs through UNE leasing. 18 The FCC
has ruled that these wholesale UNE prices are to be based on forward-looking estimates
of the cost of building the network component in question: the Total Element Long Run
Incremental Cost (TELRIC). Thus, UNE prices can be thought of as the prices required to
recover the cost of the element, over the total life of the asset.19
18. See “Interconnection” within Part II, Section 251, Subsection C, Paragraph 3 of the 1996
Telecommunications Act, which specifies that ILECs have “the duty to provide, to any requesting
telecommunications carrier for the provision of a telecommunications service, nondiscriminatory access to
network elements on an unbundled basis at any technically feasible point on rates, terms, and conditions
that are just, reasonable, and nondiscriminatory in accordance with the terms and conditions of the
agreement and the requirements of this section.” Cite to regulatory history here
19. This price may be substantially less than the embedded cost of a facility because of technical
change. For this and other reasons, the ILECs have appealed the FCC’s decision to require forward-looking
costs in the Federal courts. The U.S. Supreme Court has agreed to hear ILEC arguments on UNE pricing
during the year 2001. See Verizon Communications, et al. v. Federal Communications Commission, et al.,
121 S. Ct. 877, on January 22, 2001.
C
R I T E R I O N
E
C O N O M I C S
, L.L.C.
-24TSR is simply the sale of ILEC services to consumers under a different brand
name. The CLEC leases the services at a discount below the ILEC’s retail price and
offers them to its customers under its own brand name. These wholesale discounts are
generally negotiated or settled through arbitration at approximately 15 to 25 percent
below retail price.20 Thus, if a CLEC wishes to provide an array of services similar to the
ILECs, it can use two leasing schemes to solidify its network. It could use TSR, and
simply resell all ILEC services, taking over the billing and promotion functions of the
ILEC. The disadvantage of TSR is that the CLEC product is identical to the ILEC
product, and therefore a CLEC using TSR cannot offer a superior service to the ILEC.
Alternatively, it can lease some facilities at UNE rates and build a hybrid network to
deliver its own services.
The extent to which a CLEC uses any of the above methods to build a system at
the time of entry will depend on current wholesale rates relative to retail costs and the
existing ILEC infrastructure. Ideally, the CLEC wishes to capture the largest margin
between price and cost as possible, thus maximizing profits. Given that the ILEC network
is already in place, a thrifty CLEC will use the extant network to its advantage, leasing or
reselling when UNE or TSR rates are sufficiently below retail prices. In addition, the
CLEC must know that its entry into the industry will place downward pressure on retail
prices because entry increases competition.21 Thus, the CLEC must anticipate the retail
price decline its entry causes and be certain that this adjusted price is sufficiently greater
than leasing rates.
20. Robert Crandall, and Jerry Hausman, “Competition in U.S. Telecommunications Services Four
Years After the 1996 Act,” in Sam Peltzman and Clifford Winston (eds.), Deregulation of Network
Industries, Brookings, 2000.
C
R I T E R I O N
E
C O N O M I C S
, L.L.C.
-25Any CLEC that wishes to provide a new service, such as DSL, to its customers
would be better served to enter the industry as a small, facilities-based, regional provider
with few competitors, than as one of several competitive CLECs leasing facilities from an
ILEC. At the very least, leasing from the ILEC suggests competition from that ILEC, and
quite possibly from other CLECs as well, with the attendant downward pressure on
market prices or requirement of substantial promotional expenditures. A better strategy
would be to build a new DSL network in an area that is not currently served by a carrier
offering such services, if the costs of that network can be defrayed from prospective
customer receipts. In following such a strategy, the CLEC can presumably obtain a
higher retail price for DSL because it is not competing against the ILEC.
To illustrate these points, the following two quotations were taken from a CLEC
web site. The quotations are from DLEC owners (CLEC owners who provide DSL), and
reflect the problems faced in leasing DSL from ILECs.
Owner “AN”: “Hmm: buy for $44 [from an ILEC], sell for $42 to $50,
but then pay for a staff… it’s just not possible. More margin has to be
maintained by the ISP.”22
Owner “VB”: “The ones that are making it [as a DLEC] aren’t reselling
someone else’s DSL. With just fourteen customers here in this rural area,
I’m making money from DSL as a DLEC.”23
“AN” must also compete against the ILEC for customers, thus reducing his
margin. Owner “VB” is much more clever, providing DSL in an area where there are few
(if any) competitors and higher margins.
21. Robert S. Pindyck & Daniel L. Rubinfeld, Microeconomics, Fourth Edition (Prentice Hall, 1998),
272.
22. “Why are the DSL Resellers
planet.com/technology/dsl_down_bol.html.
23. Id.
C
R I T E R I O N
in
E
Trouble?”
ISP-Planet
C O N O M I C S
web
, L.L.C.
site,
http://www.isp-
-262.
The Long-Run
The long-run decision as to whether or not a CLEC invests in on-net components
in its network is obvious. Simply put, if a CLEC uses only resale, it will suffer in the
long-term, as it is offering a service identical to the ILECs at a prospective margin that is
capped by the resale discount. Moreover, because there are fewer barriers to entry in
resale, other CLECs can enter as resellers, thus reducing retail prices until all CLECs
providing service via resale are making no more than a competitive return on investment.
The CLEC that survives in the long run is the firm that builds on-net components in order
to maintain cost and service advantages over its rivals, including the ILECs. Such a
CLEC can attract customers, because it offers quality service, maintained through stateof-the-art network components, but it can also offer this service at lower prices if it can
provide service through newer or somewhat more efficient facilities than those in the
ILEC network. A clever CLEC may be able to utilize UNE leasing, while also improving
the ILEC network to generate cost savings and service advantages; eventually, an
efficient and cost-conscious CLEC will be able to leverage these investments and move
to an on-net platform.
As of the end of 2000, approximately 35 percent of CLEC lines were on-net, as
compared with about 36 percent resale and 29 percent UNE.24 Thus, CLECs have been
adding to the ILEC network in addition to leasing and reselling. However, there is
substantial variance in these proportions across CLECs. First, the above percentages
reflect all CLECs that responded to the FCC’s questionnaire, and therefore many small,
private firms are included in those figures. The publicly traded CLECs, however, favor
on-net lines somewhat more heavily. To be specific, the major public CLECs have 37
C
R I T E R I O N
E
C O N O M I C S
, L.L.C.
-27percent on-net, 32 percent resale, and 31 percent UNE lines.25 Thus, the average, public
CLEC network apparently has more on-net lines than the average private CLEC,
although a number of public CLECs still rely heavily on UNE and resale. For example,
Time Warner Telecom, one of the most successful CLECs, has 81 percent of its lines onnet, 19 percent through resale, and no UNE lines, whereas CTC Communications has 6
percent on-net lines and 94 percent resale, and Focal Communication’s network is
comprised of 100 percent UNE lines.26
D.
Tailoring Investment to Demand
In addition to service and network decisions, a CLEC must assess the demand for
its services before attempting to supply that product. First, if the CLEC will be competing
against an ILEC in a given market, the CLEC must predict the likelihood that existing
ILEC customers will switch to CLEC service. Once an attractive market is found, the
CLEC must carefully assess the rate at which customers can be enrolled to begin to
defray the costs of the network. Such an assessment is necessary in order to procure the
necessary funding to pay for the network.
If the CLEC fails to arrange for the network capacity to handle its traffic, poor
service quality will occur and customers, revenue, and reputation will be lost. If the
CLEC overbuilds the facilities necessary to support network traffic, scarce investment
capital will have been wasted, and financiers will become wary of lending to the CLEC.
While these excessive capital expenditures may not hurt the CLEC in the short term, the
future capital squeeze will have negative effects on the CLEC. Thus, predicting consumer
24. FCC, Local Competition Report, May 21, 2001.
25. Credit Suisse First Boston, Telecom Services—CLECs, (April 11, 2001), 18-19.
26. Id.
C
R I T E R I O N
E
C O N O M I C S
, L.L.C.
-28demand, and tailoring their facilities to meet that demand, are crucial to these new
entrants.
The CLECs have shown sub-par performance in balancing investment with
consumer demand. Industry analysts have estimated that revenue from newly added
customers has increased only 11 percent, while capital expenditures rose by 39 percent in
the last two years. Furthermore, it is estimated that as much as 97.5 percent of existing
high-capacity fiber-optic lines is unused.27
Even though these statistics are rather
general, they indicate that the industry as a whole was not effective in tailoring capacity
to demand. Therefore, it is likely that newly constructed capital assets were not
effectively generating revenues for the CLECs. I explore this topic further in the
empirical analysis in section V.
E.
Product Quality
A CLEC may attempt to compete by claiming to offer better service than the
ILEC in its region. However, many CLECs have lost customers as a result of poor
service. Customers have complained that service is either not installed by the agreed date,
or that their calls are disconnected and poorly routed.28 CLECs have blamed the ILECs
for these problems, arguing that faulty ILEC network connections reduce CLECs’ quality
of service. ILECs have in turn responded that CLEC problems are the result of CLEC
errors in routing calls.29 However, as I shall show, these complaints are relatively rare
and could not be the principal cause of poor CLEC quality. Integrating CLEC lines to the
27. Tom Fredrickson, “Too Many Lines, Too Few Callers; Telecom Upstarts Shelving Expansion
Plans,” Crain’s New York Business, (April 23, 2001), at 20 (“Too Many Lines”).
28. Yuki Noguchi, “Local Phone Frustration; CLECs Blame Bells, Bells Blame Hookups, Some
Blame Agencies,” Washington Post, (December 14, 2000), E1.
29. Id.
C
R I T E R I O N
E
C O N O M I C S
, L.L.C.
-29ILEC network has resulted in numerous problems, and the CLECs have suffered as a
result of poor service, but such problems cannot simply be attributed to ILEC behavior.
The processes involved in rolling out a new network, or integrating new lines into
an existing network, are likely to produce a number of problems with service quality.
Royce Holland, chairman of Allegiance Telecom (one of the more successful CLECs),
states that “[CLECs] are going to be blamed [for poor product quality] regardless of
where the problem is, and rightly so,”30 because part of their pledge to customers is
quality service.
F.
The Rate of Expansion
Once a CLEC has established its service within a market, it may consider
expanding into another market. Expansion will depend upon the CLEC’s assessment of
potential profitability, current market conditions and prices, the potential for entry from
other firms, and the time and cost of establishing a new network and client base.
Furthermore, the CLEC must consider the effect that expansion will have on service in its
existing markets. It is possible that expansion can divert resources from existing markets;
thus, expansion can divert the CLEC’s attention away from its customer base and lead to
reductions in service quality that will ultimately hurt the CLEC.
Furthermore, the rate at which new markets are added can have detrimental
effects on the CLEC’s network. If new markets are added before existing market
networks are fully operational, existing network capital will not generate sufficient
revenues or cash flows to cover the costs of the CLEC’s market extension. Such rapid
expansion can cause concern over the strength of the CLEC, leading investors to
withhold capital from the CLEC in the future. Because a CLEC’s extension into new
C
R I T E R I O N
E
C O N O M I C S
, L.L.C.
-30markets has effects on the company’s customers and investors, the CLEC must fully
assess the consequences that its rate of expansion will have on the profitability of current
and future operations.
IV. INDUSTRY COMPETITION AND BARRIERS TO ENTRY
To understand the current position of the CLECs, one must understand that local
competition for small business and residential customers is a very recent phenomenon,
both in the United States and the rest of the world. There are no well-defined models for
entrants to follow.
A.
Barriers Prior to the 1996 Act
As mentioned earlier, the CAPs were competing with the ILECs to
provide local service to medium and large sized businesses prior to 1996. This
market was at least partly competitive prior to CLEC entry; therefore, profit
margins for urban business services may have already been under some pressure
even prior to 1996. As a result, some of the benefits to CLECs from entering the
medium to large-scale business market had already been whittled away even
before the CLECs entered.
Notwithstanding exceptions for some CAPs, state regulators had blocked
entry into most of the local exchange/access markets in all but a few states prior
to 1996. When the Telecom Act was passed, the new entrants faced the prospects
of competing for most local customers for the first time.
30. Id.
C
R I T E R I O N
E
C O N O M I C S
, L.L.C.
-31B.
Existing Customer Contracts as Barriers to Entry
Some CLECs attempted to enter the local-exchange market after the initial wave
of entrants, but experienced difficulties due to unforeseen barriers to entry in the form of
long-term contracts between business customers and the CAPs or early CLECs. In
addition, early CLECs also focused attention on small to medium sized businesses
because they did not have to compete against the CAPs as vigorously as in the medium to
large business market, and margins for these customers were kept relatively high by state
regulation.
C.
The “Last Mile” Problem
The “last mile access line” is the line that runs into a customer’s building from the
service provider’s switch or remote terminal. Late CLEC entrants have encountered some
difficulty in providing last mile access partly because of the cost of constructing lines, but
more often because of the difficulty in extending into some office buildings. In some
buildings, a new installation can require up to 100 square feet of floor space within a
building, and because of building design, many landlords have been reluctant to allow
more than one or two companies to provide service to their buildings.31 In other cases,
substantial time and capital expenditures are required to trench through city streets or
string cables along existing utility pole lines. Nevertheless, as Figure 1 shows, many
CLECs are building their own facilities.
D.
Interconnection
Successful entry requires interconnection with the local ILEC in the form of
UNEs, resale, or collocation.
31. Paul Rogers, “Last Mile is Longest in Battle to Get Wired; New Carriers Face Many Challenges,”
Crain’s Chicago Business, (July 24, 2000), at SR16 (“Last Mile is Longest”).
C
R I T E R I O N
E
C O N O M I C S
, L.L.C.
-32-
1.
Availability of UNEs
With the 1996 Telecommunications Act, the FCC required each ILEC to allow a
CLEC to connect to its network at any feasible point by making the incumbent’s UNEs
available to the CLEC. Initially, virtually all elements of the ILEC network were
available to the CLEC. According to the FCC, UNEs can consist of “loops, network
interface devices, local circuit switching, dedicated and shared transport, signaling and
call-related data bases, and [the ILECs’] operations support systems.”32 While this set of
network components is smaller than the total feasible set of components, it still represents
almost all of the ILEC network, and must be made available to a CLEC upon request.
This affords a CLEC great flexibility to piece together an unbundled element network
platform (UNE-P) through the existing ILEC network.
A problem with developing a UNE-P, however, is that ILEC networks were not
designed with the knowledge that CLECs would be allowed to immediately interconnect
to the network wherever and whenever they desire. As a result, negotiations between a
CLEC and an ILEC over network connections can take large amounts of time and
resources, frustrating both parties. For example, Qwest and AT&T spent six months
negotiating over interconnection issues in Minnesota.33
Another problem with developing a network via UNE-P is that the CLEC strives
to compete against an ILEC through superior service. The CLEC cannot, however, offer
superior technical service through a UNE-P, as the UNE-P facility relies on the ILEC for
its service. The only way that a CLEC can truly offer superior service is by developing its
32. “FCC Makes Decision on Unbundled Network Elements,” Utilities IT, 5(1), at 52 (“FCC Makes
Decision”).
C
R I T E R I O N
E
C O N O M I C S
, L.L.C.
-33own network. Some CLECs have attempted to accomplish this task gradually, by first
designing a network using the UNE-P and then methodically replacing single elements of
the platform with their own facilities.34 While this may appear a compromise between
the need to develop a customer base and to provide superior service, it is a risky solution
because superior service cannot be provided immediately. In the short term, the CLEC
relies totally on the ILEC for its service, whose network was not designed around the
business strategy of the CLEC. Thus, product quality in the short term may even decline,
resulting in the loss of customers.
2.
Resale Discounts
Resale is an almost perfect short-run substitute to UNE leasing for the CLEC who
wishes to provide many services to customers quickly -- albeit at higher cost. Resale will
be used to provide service if the prospective margin from resale is greater than from
offering a service based on UNEs. However, even the largest CLECs, with the most
potential power and influence to obtain favorably low resale rates, have found resale
unprofitable. Shortly after the 1996 Act, AT&T began offering local service through
resale to California, Illinois, Michigan, Texas, Georgia, and Connecticut. AT&T received
approximately 17 percent resale discounts; nevertheless, the company found that reselling
at this rate was unprofitable. Analysts agreed, observing that AT&T’s decision to provide
local service through resale was unwise because of the small margins. They added that a
better decision would have been to acquire an existing network and to provide service
through that network, gaining a more attractive margin in the process. Not surprisingly,
33. “AT&T Files Formal Complaint Against Qwest in Minnesota,” PR Newswire, (March 22, 2001),
Financial News Section.
34. Last Mile is Longest, supra note 31, at SR16.
C
R I T E R I O N
E
C O N O M I C S
, L.L.C.
-34AT&T began to pull back from this resale strategy by November of 1997 when it decided
to stop enlisting new customers.35
3.
Agreements Between ILECs and CLECs
Some failing CLECs have attributed their problems to the ILECs, arguing that
ILECs have resisted their attempts at interconnection via UNE or resale. To assess the
validity of such complaints, I conducted a survey of state public utility commissions
(PUCs). Through this survey, I determined that approximately 10,342 interconnection
agreements have been completed between an ILEC and a CLEC since the 1996 Act.
However, I found only 126 instances in which a CLEC filed a complaint with these PUCs
over compliance with the interconnection rules promulgated under the 1996 Act. 36 Thus,
formal complaints occur in only about 1 percent of the sample of agreements. In addition,
I was able to determine that AT&T, WorldCom, or Sprint filed at least 11 of these
complaints.37 These firms have an obvious incentive to state that the ILECs are not
complying with the Act, in order to deter ILEC entry into the long distance market,
meaning that the true rate of CLEC complaints is likely even lower. The paucity of CLEC
complaints surely suggests that difficulties in interconnecting with the ILECs cannot have
played a very important role in CLEC failures.
35. Steve Rosenbush, “AT&T Hanging it up in the Local Phone Market,” USA Today, (November 14,
1997), 3B.
36. The number of formal complaints is even less. Some PUCs do not formally track complaints in
dockets. When conducting the survey, we asked for numbers on “formal complaints”, but when this was
not available, we asked for an estimate of serious complaints. Quite often the person gave a range, say 5 to
10, in which case I took the larger number to make our calculation. Furthermore, I included, whenever
possible, arbitration agreements that were specifically filed as CLEC “complaints or disputes” with the
PUC, even though the case was, or will be, settled in arbitration. Thus, I can say that no more than 1.2
percent of potential interconnection agreements involve formal dispute, while the actual number is lower
than 1.2 percent.
37. This is an underestimate of long distance carrier complaints because I can identify the originator of
the complaints in only a few states.
C
R I T E R I O N
E
C O N O M I C S
, L.L.C.
-354.
Reciprocal Compensation
Reciprocal compensation refers to the fee that one phone company pays to
another for completing a local call, as required by the 1996 Act. For example, if a
customer using local carrier A places a local call to a customer using local carrier B, then
company A must pay company B for completion of the call. The regulation was designed
with the assumption that, over the long term, call minutes initiated on A’s lines and
completed on B’s lines would equal call minutes placed on B’s lines and completed on
A’s lines. The advent of the Internet and dial-up modems has led to enormous debates
over reciprocal compensation policy. An Internet subscriber will frequently dial his or her
ISP, but the ISP does not call the subscriber. Thus, if the Internet provider receives its
connections through a CLEC, the ILEC must pay the CLEC for all of the minutes of
Internet connections, but the ILEC will receive no offsetting compensation because there
are no calls in the other direction. The CLEC will not have to establish itself as a full
local-exchange provider -- it can simply sit between the ILEC and the ISP, and passively
hand off calls to the ISP, collecting a windfall of revenues.38
The rates for reciprocal compensation are determined through negotiation or
arbitrations supervised by state regulators with the implicit understanding that calls will
move in both directions between a CLEC and the ILEC. 39 These rates are generally
between two-tenths and nine-tenths of a cent per-minute for a call, but the costs of
terminating Internet calls are substantially lower.40 With such a large margin of revenue
38. See: Dave Nichols, “Keeping Internet and Phone Rates Low,” The San Diego Union-Tribune,
(November 29, 2000), B7; And Sandra Jones, “A disconnect for phone upstart; Focal sees threat to key
sales engine,” Crain’s Chicago Business, (January 8, 2001), at 1.
39. U.S. Federal Communications Commission, “FCC Adopts Order Addressing Dial-up Internet
Traffic,” Report No. CC 99-2, (February 25, 1999).
40. Sandra Jones, “A Disconnect for Phone Upstart; Focal Sees Threat to Key Sales Engine,” Crain’s
Chicago Business, (January 8, 2001), at 1 (“Disconnect for Upstart”).
C
R I T E R I O N
E
C O N O M I C S
, L.L.C.
-36over cost, it is no wonder that CLECs availed themselves of this opportunity. In late
2000, the average CLEC generated about 10 percent of its revenues from reciprocal
compensation, yielding a total of approximately $2 billion per year in CLEC revenue.41
On February 26, 1999, the FCC ruled that calls to an ISP are not subject to
reciprocal compensation charges because once the ISP is contacted the call is generally
routed to Internet sites in different states, and interstate calls are not subject to reciprocal
compensation fees. On March 24, 2000, the United States Court of Appeals for the D.C.
Circuit overturned various elements of the FCC ruling.42 Subsequently, the FCC revised
its policy, capping reciprocal compensation from ISP-bound traffic and reducing
reciprocal compensation rates gradually to 0.07 cents per minute,43 thus sharply reducing
a large revenue source for the CLECs.44 For this reason, a CLEC that initially built its
business around extracting reciprocal compensation from ILECs is likely to encounter
severe difficulties. Furthermore, this firm could lose a competitive edge relative to other
CLECs who wisely avoided reciprocal compensation revenues and instead focused on
investing in new network components to build a long-lasting customer base and reduce
their operating costs.
E.
The Cable TV Companies
I have omitted from the empirical analysis in Section V an entire group of new
entrants into local telephony -- the cable television companies -- because I do not have
separate revenue data for their cable telephony services. Nevertheless, the experience of
41. Credit Suisse First Boston, Telecom Services—CLECs, (April 11, 2001), 18-19.
42. U.S. Federal Communications Commission, Public Notice FCC 00-227 “Comment Sought On
Remand Of The Commission’s Reciprocal Compensation Declaratory Ruling By The U.S. Court Of
Appeals For The D.C. Circuit,” (June 22, 2000).
43. U.S. Federal Communications Commission, Order on Remand and Report and Order, Docket
Nos. CC 96-98 and CC 99-68, April 19, 2001.
44. FCC Makes Decision, supra note 32, at 52.
C
R I T E R I O N
E
C O N O M I C S
, L.L.C.
-37these companies in offering local telephone service is relevant to any analysis of the first
few years of local competition. Cable television companies have a unique opportunity to
deliver telephony services to their subscribers over their own facilities. Many of these
companies have upgraded their systems for the delivery of two-way services. They
simply need to invest in local switches and use regular communications with their
subscribers to sell them telephony.
As of June 2000, the nation’s cable systems had 67.7 million subscribers.45 Cable
service passed approximately 95 percent of all U.S. households. Yet, by the end of 2000
there were no more than 1.1 million of these subscribers who had chosen to purchase
telephone service from a cable company—or about 1.6 percent of cable television
subscribers.46 The cable companies’ lack of success in offering a competitive local
telephone service cannot be attributed to any unfair activities of the ILECs. It must be
attributed to the difficulty in competing for residential telephone customers at the low,
monthly regulated rates, or to the inability of cable companies to assure subscribers that
they are capable of providing a service as reliable as that provided by the incumbent
telephone companies.
V. EMPIRICAL ANALYSIS OF CLEC PERFORMANCE
In order to improve our understanding of the CLECs, I use statistical techniques
to gauge the performance of these firms relative to one another. In this manner I can
provide a meaningful test of the determinants of CLEC success and failure based on the
45. U.S. Federal Communications Commission, Seventh Annual Report, In the Matter of Annual
Assessment of the Status of Competition in the Market for the Delivery of Video Programming, CS Docket
00-132, (January 2001) at ¶ 7.
C
R I T E R I O N
E
C O N O M I C S
, L.L.C.
-38CLEC’s choice of network platform or customer base. Descriptions of the model,
estimation technique, data, and variables used in this analysis are contained in the
Analytical Appendix at the end of the document. In this section I highlight the main
results.
I estimate two basic models, one gauged at analyzing the performance of specific
firms, and one tailored toward specific business practices of the CLECs. A listing of the
CLECs used in my analysis, their choices of network type, their principal customer
specialization, and heavy reliance on reciprocal compensation is displayed in Table 5.
46. Local Competition Report, supra note 2. The Commission reports that only 1.125 million CLEC
lines are coaxial cable lines. Some of these lines have been installed by a few of the new CLECs, not cable
companies.
C
R I T E R I O N
E
C O N O M I C S
, L.L.C.
-39Table 5
Data for Listed CLECs in Sample
Firm
Network Type
Customer Type
Reciprocal
Adelphia
Resale and On-net
Business
Yes
Compensation
Allegiance
UNE
Business
Allied Riser
Business
Advanced Radio
Business
Avista Corporation
Combination
US LEC Corp
UNE
Business
CoreComm Ltd.
Resale
Combination
Convergent
Business
Covad
UNE
Combination
CapRock
Resale
Business
CTC Communications
Resale
Business
Electric Lightwave Inc.
On-net and UNE
Business
Focal
UNE
Business
Yes
GST
On-net and UNE
Business
ICG Telecommunications
On-net and UNE
Business
Yes
Intermedia
UNE
Business
Yes
Inter-Tel Inc.
Business
ITC DeltaCom Inc.
UNE
Business
McLeod USA Inc.
UNE and Resale
Combination
Metromedia
Business
Mpower
UNE
Business
Network Access
UNE
Business
Network Plus CP
UNE
Business
NorthPoint
Combination
North Pittsburgh
Combination
Net 2000
UNE and Resale
Business
Net2Phone
Business
Primus
Combination
RCN Corp.
On-net and Resale
Residential
RMI.Net
Resident
RSL
Business
Rhythms
UNE
Business
SpeedUS.Com
Combination
Teligent Inc.
On-net
Business
Telocity
Residential
Time Warner TLC
On-net
Business
World Access
Business
Winstar Communications
On-net/UNE
Combination
XO Communications
On-net/UNE
Combination
ZTEL
Resale
Residential
Sources: “Telecom Services—CLECs,” Credit Suisse First Boston, (June 5, 2001) at 19;
“Telecom Services—CLECs,” Credit Suisse First Boston, (April 11, 2001) at 18; “Analysis of
Local Exchange Service Competition in New York State,” New York State Public Service
Commission, (December 31, 1999); “Broadband Barometer,” Merrill Lynch, (July 3, 2000), at 4;
Company Websites, and SEC Filings.
A.
Empirical Analysis of Individual CLECs
In Appendix 1, I provide a detailed empirical “regression” analysis of CLEC
performance. Given that most of these companies are in an early stage of development, it
would be pointless to focus on their profitability. Moreover, the market’s assessment of
C
R I T E R I O N
E
C O N O M I C S
, L.L.C.
-40their likely future as reflected in their stock prices has shown wild swings in the past two
years as Table 2 showed. Therefore, to gauge the initial success of each CLEC, I
examined how it translated investment in fixed assets into revenues. Specifically, I
estimated the relationship between revenues in each quarter and fixed assets in the
previous quarter. The successful firms should be enrolling customers and realizing
revenues as they deploy their networks. Those that fail to attract customers as rapidly are
obviously more likely to fail to satisfy investors that they should continue to fund
negative cash flows.
The results of this initial analysis may be gauged by the relative size of the
coefficients in Table 6.47 For example, McLeod, Time Warner, RCN, and Intermedia
have positive values, suggesting that they are successful in generating increases in
revenues through the addition of fixed assets. However, Rhythms, Covad, NorthPoint,
and Teligent have very large negative values, which means that their ability to generate
revenues from asset expansion is less than the trend rate in the sector. The latter firms are
either in bankruptcy or very close to bankruptcy, while the former are in a much more
solid condition. Even though these more successful firms have suffered a decline in
market capitalization, they continue to grow and to invest in facilities.
C
R I T E R I O N
E
C O N O M I C S
, L.L.C.
-41Table 6
The Relative Success of Individual CLEC’s in Deploying Capital
Adelphia Business Solutions
Allegiance Telecom Inc.
Allied Riser
Advanced Radio
US LEC Corp
CoreComm Ltd.
Convergent
Covad
CapRock
CTC Communications Corp.
Electric Lightwave Inc.
Focal Communications.
GST Telecommunications
ICG Telecommunications
Intermedia Communications
Inter-Tel Inc.
ITC DeltaCom Inc.
McLeod USA Inc.
Metromedia
Mpower
Network Access
Network Plus CP
NorthPoint
North Pittsburgh
Net 2000
Primus
RCN Corp.
RMI.Net
RSL
Rhythms
SpeedUS.Com
Teligent Inc.
Telocity
Time Warner TLC
World Access
Winstar Communications Inc.
XO Comm. (Nextlink)
0.372
-0.007
-0.132
-0.219
0.027
-0.003
0.057
-0.023
0.051
0.046
0.004
0.018
0.021
0.041
0.065
0.105
0.036
0.088
-0.025
-0.021
-0.083
0.036
-0.063
-0.005
-0.047
0.103
0.034
-0.011
0.124
-0.090
-0.351
-0.091
-0.151
0.026
0.054
-0.051
0.027
Note: Avista is not included in this analysis because it is impossible to separate telecom revenues
from revenues of other operations such as electric and natural gas.
47. The full results are shown in the appendix in Table A-3.
C
R I T E R I O N
E
C O N O M I C S
, L.L.C.
-42B.
The Effect of Network Design and Customer Strategies
The first empirical analysis focused only on the identity of the CLEC. In this
section, I report on the results that were obtained from a statistical analysis of the effect
of network design and type of customer on the ability of an entrant to translate fixed
capital assets into revenues. The results of this analysis, reported in the Appendix,
provide strong evidence that building one’s own network is the best entry strategy. Using
UNEs to leverage fixed assets into revenues is much less successful in building revenues,
and the use of resale -- on average -- produces very poor results.
Specifically, the statistical regression analysis shows that CLECs with their own
networks are typically able to increase revenues 2.6 percentage point above the average
rate of increase for every 1 percent increase in capital assets.48 The use of a combination
of their own networks with a substantial share of UNEs or resale generates a 1.4
percentage point increase above the average growth rate in revenues for every 1 percent
increase in capital assets. However, using either UNEs or resale or a combination of the
two to build its network results in much lower revenue growth. A principal reliance on
UNEs generates only a 0.7 percentage point above-average revenue increase for each 1
percent increase in capital assets while resale and a combination of UNEs and resale
provides almost no incremental boost in revenues. In short, a mixed strategy of using
UNEs or resale, in addition to investment in a CLEC’s own facilities, is far superior to
relying on UNEs or resale by themselves.
Surprisingly, I find that there is no difference in performance between CLECs that
target business customers and those that primarily serve residential customers.
C
R I T E R I O N
E
C O N O M I C S
, L.L.C.
-43Apparently, the few CLECs that address the residential market, such as RCN, do not
systematically under-perform the vast majority of CLECs that target the business market,
all other factors equal. Finally, reliance upon reciprocal compensation does not
contribute significantly to revenue growth -- a surprising result given the limited effort
required to obtain such revenues when terminating calls directed toward an ISP. The
FCC’s recent decision to revise and reduce reciprocal compensation rates has severely
limited the success of this strategy.
These results provide strong support for the conclusion advanced above -- namely
that the entrant’s best strategy for growth is to build its own facilities. A few, such as
McLeod, have succeeded with a resale and UNE strategy, and Intermedia has been
relatively successful with a UNE-only strategy, but the statistical results suggest that
building one’s own network is likely to be the best way to build revenues. Of course, this
does not guarantee that an entrant will ultimately become profitable and survive. Only
time will provide the proof of long-term profitability.
These results provide no support for the notion that the inability to gain
interconnection through UNEs or the transfer of resale customers has impeded CLEC
growth. The results simply point out that building one’s own network is likely the best
platform strategy for long term revenue growth. Indeed, a mixed strategy of using UNEs
or resale with one’s own network appears to work relatively well, but simply relying on
the ILEC’s network appears to be a strategy that limits an entrant’s growth. Just changing
the nameplate on the service is not typically a very good strategy for attracting customers.
48. These results are shown in Table A-5 in the Appendix. Note that the principal reliance on on-net
facilities leads to very low initial revenue growth because of the time required to build one’s own network.
Thereafter, the revenue growth of on-net CLECs far outstrips that of their rivals.
C
R I T E R I O N
E
C O N O M I C S
, L.L.C.
-44VI. WHO SUCCEEDED, AND WHY?
A.
The Most Successful CLECs
During the winter of 2001, many public CLECs’ stocks were trading at prices in
the single digits, and some firms were even filing for bankruptcy. Despite the problems of
some firms, analysts have continued to view Allegiance Telecom, Inc., McLeod USA,
Time Warner Telecom, and XO Communications as strong companies49 and have stated
that these companies are proof that CLECs can thrive and contribute to a competitive
telecommunications marketplace. Furthermore, analysts have attributed these CLECs’
performance to a “…deep knowledge of how to coordinate the physical and
administrative change from former Bell company to new carrier.”50 These opinions seem
well grounded, as certain CLECs have separated themselves from the rest of the pack
over the past few years. Nevertheless, I believe that modifications to the above list are
necessary to categorize the firms properly.
For reasons I describe below, it is clear that McLeod and Time Warner are the
most successful CLECs, with Allegiance closely behind. XO is somewhat below the top
three CLECs, along with Intermedia, a firm with a very successful Internet service.
The variation in CLEC revenue over time is the first indication of the differences
in CLEC performance. Publicly traded CLECs reported average revenues of
approximately $45 million in the first quarter of 1998. This figure increased to over $167
million by the third quarter of 2000. Clearly, some CLECs have been growing. Over this
same period, however, the variation in CLEC revenues has increased substantially. In the
49. Neil Druker, supra note 10 at 30-32; and Riding up to the Challenge, supra note 10, at G14.
50. Riding up to the Challege, supra note 10, at G14.
C
R I T E R I O N
E
C O N O M I C S
, L.L.C.
-45first quarter of 1998, the standard deviation51 of CLEC revenues was approximately $100
million, but the standard deviation increased to over $440 million by the third quarter of
2000. These numbers reflect the fact that not all CLEC revenues were growing at the
same rate. Thus, some CLECs have established themselves as larger companies, with
higher rates of expansion, while others remain relatively small.
Allegiance, McLeod, Time Warner, and XO have all performed well in this
respect. They are all relatively large CLECs, the smallest being Allegiance with revenues
of slightly more than $94 million in the fourth quarter of 2000, and all have had
consistent revenue growth. Allegiance’s revenues have increased by more than 400
percent since 1998 and by more than 70 percent since the fourth quarter of 1999. McLeod
has grown over 100 percent since 1998, and over 32 percent since the end of 1999,
boasting total revenues of over $410 million by the end of 2000 and consistent revenue
growth of about 10 percent per quarter since 1998. Time Warner has seen its revenues
increase by more than 170 percent since 1998, and over 28 percent since the end of 1999,
while XO Communications has grown over 213 percent in revenue since 1998 and 91
percent since the fourth quarter of 1999.
Of the four firms listed above, it is clear that McLeod and Time Warner are the
strongest firms. Both of these firms are fully funded, and have positive earnings before
interest, taxes, and depreciation (EBITDA), although their business models are different.
McLeod relies heavily on the resale of ILEC services, while Time Warner relies mostly
on its own network. I classify Allegiance as a solid firm, but below the level of the above
two, because it is not as mature as McLeod and Time Warner. For example, Allegiance
51. Standard deviation is a common statistical tool that measures the spread of the data. Higher
numbers correspond to greater spread.
C
R I T E R I O N
E
C O N O M I C S
, L.L.C.
-46reported only $1.2 million in quarterly revenues early in 1998, while McLeod and Time
Warner reported $155.7 million and $27.05 million respectively in the same period. Thus,
Allegiance was growing quickly in 1998 and 1999, at least in part, because it was a small
company in new markets. Allegiance stands out from many other CLECs in its sustained
revenue growth through the year 2000, when some CLECs were having difficulties.
XO Communications has not been as stable as the other three firms. It has
invested in some alternative network platforms, such as fixed wireless, which could prove
extremely profitable in the long term, but these investments in unproven technologies
place the company at greater risk in the short term. Further, XO has had, until recently,
more trouble securing funding than the other three firms. For these reasons, I rate XO
below McLeod, Time Warner, and Allegiance.
In addition to the above firms, Intermedia is a large firm with $1 billion in annual
revenues derived from a mix of Internet, web hosting, local access and voice, and
integration services. Intermedia experienced difficulties in December 2000, and its stock
price fell to less than $3.7 per share after two quarters of stagnant revenues.52 Since then,
the company has rebounded, and its stock has rallied to over $17 per share at a time when
most CLEC shares were under severe pressure.
Simply put, these CLECs appear to have understood the industry prior to entry,
had well devised business models, and developed their networks with the intention of
making themselves valuable to their customers. Below, I highlight the specific business
strategies that have allowed these firms to succeed.
52. This figure was downloaded from Yahoo Finance.
C
R I T E R I O N
E
C O N O M I C S
, L.L.C.
-471.
McLeodUSA
McLeodUSA provides local service, long distance service, data service, and voice
mail to residents and businesses. McLeod uses a combination of resale, UNE leasing, and
new construction in order to serve its customers. While resale has not been a particularly
rewarding strategy for most CLECs, McLeod has been able to take advantage of the
resale of Centrex services (a bundled service to businesses that predates the 1996 Act in
US West and Ameritech States) in order to expand service. McLeod has also been adding
CLEC lines and installing its own switches. From the fourth quarter of 1998 to the first
quarter of 2001, McLeod reported a 279 percent increase in its total access lines. Since
the second quarter of 2000, resale lines as a percent of total lines fell from 70 percent to
67 percent53, and since the fourth quarter of 1998, the number of McLeod owned
switches increased from 7 to 50. Thus, the on-net portion of McLeod’s network has
increased along with the size of its network.54 In addition, McLeod has expanded by
purchasing CapRock Communications.55
2.
Time Warner Telecom
Time Warner Telecom is a subsidiary of AOL Time Warner. During the first
quarter of 1998, Time Warner Telecom reported revenues of approximately $22
million.56
Revenues increased to over $173 million in the first quarter of 2001.57
Individuals at Time Warner Telecom attribute the company’s success to its ability to
53. See “Telecom Services—CLECs,” Credit Suisse First Boston Corporation, (June 5, 2001), 19 and
“Telecom Services—CLECs,” Credit Suisse First Boston Corporation, (September 12, 2000), 19.
54. Downloaded
from
McLeodUSA’s
website
at
(http://www.mcleodusa.com/html/ir/quarterlyreleases.php3)
55. George C. Ford, “McLeodUSA Buys Dallas, Texas-Based Fiber Optic Company to Increase
Empire,” The Gazette(Cedar Rapids), (December 8, 2000).
56. Time Warner Telecom was not public in 1998. The 1998 revenue figure is taken from the
company’s U.S. Securities and Exchange Commission Form 10-Q, (May, 1999), which lists the first
quarter 1998 figure.
C
R I T E R I O N
E
C O N O M I C S
, L.L.C.
-48maintain its strategy, adding that adopting new technology because it is “in vogue” can
hurt a CLEC.58 Time Warner has also engaged in prudent business and financial actions
in order to improve the firm. First, Time Warner bought a bankrupt CLEC, GST
Communications, for $690 million on January 10, 2001. The acquisition of GST was a
sensible decision because it allowed Time Warner to grow in a calculated manner. Prior
to January of 2001, Time Warner had been growing very methodically through the
construction of its own facilities. In the first quarter of 2001, its total access lines were
comprised of 81 percent of on-net lines and 19 percent resale lines,59 and it offered
service in 23 markets by December 2000.60 Thus, the firm’s strategy was to build its own
network in major markets, taking advantage of large, regulated margins in those markets,
while offering lower cost service with new technology.
When GST began to experience financial difficulties, Time Warner saw the
acquisition of GST as an opportunity to expand at an accelerated rate and discounted
cost. Prior to the acquisition, GST’s network consisted of approximately 50 percent onnet lines, and 50 percent UNE lines, well above the industry average of 36 percent on-net
lines.61 Thus, GST had already taken the time and energy to build a large portion of their
network. Furthermore, GST’s operations covered 49 cities by the fourth quarter of
1999.62 Time Warner was able to acquire a bankrupt company whose current network
largely reflected what Time Warner would have built on its own years hence.
57. Time Warner Telecom, U.S. Securities and Exchange Commission, Form 10-Q, (May, 2001).
58. Riding up to the Challenge, supra note 10, at G14.
59. “Telecom Services—CLECs,” Credit Suisse First Boston Corporation, (June 5, 2001), at 19.
60. “Time Warner Telecom Expands Network to Columbus Suburbs,” (December 11, 2000),
Downloaded from Time Warner’s website at (http://www.twtelecom.com/jsp/allnews.jsp).
61. “Telecom Services—CLECs,” Credit Suisse First Boston Corporation, (September 12, 2000), at
19.
62. GST Telecommunications, U.S. Securities and Exchange Commission, Form 10-K, (March, 2000),
at 3.
C
R I T E R I O N
E
C O N O M I C S
, L.L.C.
-49Another successful Time Warner strategy was its financing of the GST
acquisition. Amid an increase of the value of Time Warner’s stock in January of 2001, it
sold shares in the firm to raise over $480 million. Furthermore, the demand for junk
bonds was rising simultaneously with the value of Time Warner’s stock, and the firm
used this opportunity to sell $400 million of their junk bond holdings at attractive prices.
Through these deals, the company paid for its purchase of GST and reduced the riskiness
of its balance sheet.
3.
Allegiance Telecom, Inc.
Allegiance
Telecom,
Inc.
is
a
CLEC
that
offers
“state-of-the-art
telecommunications products - voice, data and Internet - all from a single source on one
affordable bill.”63 Allegiance began operations out of Dallas, TX in 1998, and filed its
first form 10-Q with the U.S. Securities and Exchange Commission after the first quarter
of 1998, reporting revenues of approximately $1.2 million. Since that time, its revenues
have grown to over $94 million in the fourth quarter of 2000,64 and the company has
expanded its operations to the top 28 markets in the United States.65 Furthermore,
analysts predict that Allegiance has obtained sufficient investment funding to sustain it
until it begins to report positive earnings.66 As a result, analysts view Allegiance as one
of the top CLECs in the industry.67
One of the keys to Allegiance’s success has been its strategic use of the existing
ILEC network in building its own network. Allegiance leases last mile access lines from
63. Downloaded from Allegiance’s website at (http://www.algx.com/about_main.php).
64. Allegiance Telecom, Inc., U.S. Securities and Exchange Commission Form 10-Q, (November 14,
2000), 4.
65. Downloaded from Allegiance’s website at (http://www.algx.com/news/san_antonio.php)
66. Neil Druker, supra note 10, at 31.
67. Riding up to the Challenge, supra note 10, at G14
C
R I T E R I O N
E
C O N O M I C S
, L.L.C.
-50ILECs, and then builds its own equipment on either side of the last mile line.68 Leasing
last mile lines (UNEs) can accelerate growth in service deployment, and quicken the
development of a customer base when the CLEC first starts operation. In addition,
building equipment on either side of the last mile line can significantly improve service
and lower cost because some ILEC components can be outdated and unreliable. Thus,
when Allegiance leases a network component from the ILEC, the company also installs
new cost-effective components in order to improve product quality and lower costs.69
Thus, Allegiance has succeeded not by repackaging and reselling ILEC services; rather,
Allegiance has solidified its presence in the telecommunications industry by upgrading
and improving the ILEC network in order to offer customers cheaper service with
superior quality.
B.
The Second Tier of CLECs
McLeod and Time Warner are clearly leading the CLECs, and Allegiance is not
far behind. In addition to these three firms, there are other CLECs that are successful, but
at least presently, to a lesser degree. Two of these firms are XO Communications and
Intermedia. Both firms have been successful in portions of their operations, but less
successful in others, accounting for their rating in the second tier of CLECs.
1.
XO Communications (Formerly Nextlink Communications)
Nextlink Communications was a CLEC providing Internet access to small and
medium sized businesses through a fixed wireless network. From the first quarter of 1998
to the fourth quarter of 1999, the firm’s revenue increased from approximately $26.5
68. Neil Druker, supra note 10, at 31.
69. Id.
C
R I T E R I O N
E
C O N O M I C S
, L.L.C.
-51million to approximately $90 million.70 Furthermore, Nextlink’s network contained 26
percent on-net lines and 74 percent UNE lines in the second quarter of 2000.71 Thus, the
firm was not relying wholly on ILEC elements in order to provide service. The
aggressive nature of its expansion set Nextlink apart from other CLECs.
First, Nextlink invested in LMDS spectrum licenses so it could supply customers
with service via a fixed wireless technology. Through this technology, antennas are
placed on the customer’s roof, and signals are then sent to a hub station. The advantage of
wireless is that the last mile access problem can be avoided, and installation is quicker
(approximately 5 days as opposed to 30 days with wireline installation). 72 In January
2000, Nextlink announced the purchase of Concentric Network Corporation. This
acquisition allowed Nextlink to expand its local and long distance telephone service to
provide high-speed Internet connections for business.73
Nextlink, which became XO Communications shortly after acquiring Concentric,
was able to expand from 49 markets in early 2000 to 60 by February of 2001, and it even
expanded telephone service to Canada and Europe. Furthermore, XO Communications
states that it has always procured the requisite funding 12 to 18 months prior to any
expansion in order to maintain the strength of the company.74
This allows XO to
continue its aggressive approach to expansion, which is one reason why it is viewed as a
solid competitor.
70. Nextlink Communication, U.S. Securities and Exchange Commission, Form 10-Q and 10-K,
(May, 1998 and March 2000).
71. “Telecom Services—CLECs,” Credit Suisse First Boston Corporation, (September 12, 2000), at
19.
72. Last Mile is Longest, supra note 31, at SR16.
73. “Nextlink Pays $2.9 Billion for Concentic Network,” The Buffalo News, City Edition (January 10,
2000), 1C.
74. Riding up to the Challenge, supra note 10, at G14.
C
R I T E R I O N
E
C O N O M I C S
, L.L.C.
-52XO’s rapid expansion, however, caused it eventually to fall behind in required
funding. The company did have problems in gaining new sources of finance early in
2001, but it recently procured sufficient funding to continue into the year 2003.75 XO’s
long-term value will depend in part on whether its wireless network proves as effective in
transferring data as the more conventional fiber networks. If so, XO’s assets will add
substantial value to the industry well into the future. If not, then the firm’s value will be
downgraded. XO’s strength is shown by the fact that it can still obtain funding from a
financial market that has shown considerable skepticism toward telecommunications
firms. Its somewhat unconventional network choice (which could more than pay off in
the long term) and its somewhat overzealous expansion plans make it a less vibrant firm
than the three described above.
2.
Intermedia
Intermedia is being purchased by World Com. 76 Intermedia has struggled even
though it has consistently been able to effectively deploy new capital assets to produce
revenue (see Table 6). Its success derives largely from its 54 percent ownership of the
valuable web hosting company, Digex, which has contributed substantially to its recent
growth in revenues. Intermedia’s results for the fourth quarter of 2000 show that revenue
from data transfer and web hosting grew at approximately 14 percent per quarter during
the year 2000. At the same time, Intermedia’s revenues in the area of voice and local
access actually declined.77 Intermedia attributes this decline in voice and local revenue to
its earlier reliance on reciprocal compensation. Specifically, total revenues fell from
75. “XO Gets Financing into 2003, Shares Surge,” Reuters, (April 26, 2001).
76. See Table 3 and Appendix 2.
C
R I T E R I O N
E
C O N O M I C S
, L.L.C.
-53$261.7 million to $247.4 million between the first and second quarters of 2000. Revenues
net of reciprocal compensation increased during this time period, rising from $229.8
million to $239.4.78 Clearly, Intermedia made a strategic error in relying too heavily on
reciprocal compensation revenues, but it made a wise decision in targeting data exchange
and web hosting as a large portion of its business. The incremental value of Intermedia to
the market—and to MCI WorldCom—lies largely in its web hosting business.
VI. WHO FALTERED, AND WHY
Time Warner acquired GST after GST filed for chapter 11 bankruptcy protection.
ICG Communications also filed for chapter 11 protection shortly thereafter. 79 Recently,
several other CLECs have filed for bankruptcy protection (see Table 3), and 10 publicly
traded CLECs have experienced negative revenue growth since the fourth quarter of
1999. The most common problems that have plagued these unsuccessful CLECs have
been over expansion -- leading to poor quality, reliance on resale, and reliance on
reciprocal compensation.
A.
ICG Communications, Inc.
ICG Communications, Inc. filed for chapter 11 bankruptcy in September 2000.80
Shortly before this event, ICG’s stock value had declined 60 percent, and the company
reduced its expectations of revenue for the year 2000, citing customer service problems
77. Downloaded
from
Intermedia
(http://www.intermedia.com/company/press/release.cfm?releaseid=400).
78. “Telecom Services—CLECs,” Credit Suisse First Boston, (April 11, 2001), at 11.
79. ICG Files for Bankruptcy, supra note 7, at 1B.
80. Id.
C
R I T E R I O N
E
C O N O M I C S
, L.L.C.
at
-54as a reason for the revenue shortfall.81
When asked to comment on ICG’s recent
performance, Andrew Morley of Level 3 Communications stated, “you need to know
who your customers are, know why you serve them and remember they are your No. 1
priority. That’s where I think ICG took its eye off the ball.”82 In explaining why ICG had
problems with customer service, analyst Dave Heger of A.G. Edwards said that “the
company put in all [those] lines and a lot of them must not have been working right. Now
you have major customers saying they may pull their business.” 83 Thus, industry sources
believe that over expansion was a major problem in the case of ICG, leading to poor
product quality, and eventually lost business.
These views of ICG’s problems are supported by data on its revenue and accessline growth from 1998 to the third quarter of 2000. During this time period, ICG’s
average revenue growth was approximately 9.1 percent per quarter, while average line
access lines growth was approximately 19 percent per quarter. ICG was extracting less
money for each access line in its network over this time period. 84 This was typical of the
CLECs in general, as revenue per line for even the highest performing CLECs decreased
approximately 3 to 4 percent per quarter from 1999 to 2001.85 ICG suffered a 56 percent
decline in revenue per line over this time period, confirming that over expansion was the
principal cause of ICG’s problems. The more successful CLECs suffered much smaller
81. KW Meyers, “ICG Troubles Offer Lesson for the Industry,” Denver Rocky Mountain News,
(September 25, 2000), 11B.
82. Id.
83. Heather Draper, “ICG’s Tumble a Wake-up Call to Telecom Firms,” Denver Rocky Mountain
News, (September 24, 2000), 1G.
84. Revenue figures are obtained from U.S. Securities and Exchange Commission, Forms 10-Q and
10-K.
85. See “Telecom Services—CLECs,” Credit Suisse First Boston, (June 5, 2000), at 15 and “Telecom
Services—CLECs,” Credit Suisse First Boston, (June 5, 2001), at 15.
C
R I T E R I O N
E
C O N O M I C S
, L.L.C.
-55declines in revenues per line, and one -- Allegiance -- actually experienced an increase in
revenues per line over this period.
B.
CTC Communications
Another CLEC that relies heavily on resale is CTC Communications. CTC
provides local and long distance telephone, and high-speed data services,86 and it leases
97 percent of its network lines through resale agreements. CTC has been very aggressive
in adding capital assets. In the first quarter of 1998 CTC reported only $1.7 million in
capital assets, but it expanded steadily to over $195 million in assets by the fourth quarter
of 2000. During the period, revenues were rising steadily from $12.8 million to $62.3
million. Thus, capital assets were growing at about 43 percent per quarter, while revenues
were growing at about 14 percent per quarter. Given the difference in the growth rate of
assets over revenues, CTC has now revised its business model, adding new lines only
after it has signed on new customers.87 The revised plan was announced at a time when
CTC’s stock price had fallen from a high of over $50 to around $5.
Over-expansion is clearly a major source of CTC’s problems, and this is
obviously one of the reasons for its new deployment strategy, but another problem is its
reliance on resale. A simple resale strategy has caused serious problems for many
CLECs, most notably AT&T. If AT&T finds resale unprofitable, then there is no reason
to think that a smaller firm, such as CTC, would be able to build a sustainable business
by reselling ILEC services.
86. CTC Communications Corp., U.S. Securities and Exchange Commission, Form 10-K, (June 29,
2000), at 1.
87. Too Many Lines, supra note 27, at 20.
C
R I T E R I O N
E
C O N O M I C S
, L.L.C.
-56C.
Teligent
On May 21, 2001 Teligent filed for Chapter 11 bankruptcy protection. Trading of
the firm’s stock was halted on NASDAQ at 56 cents per share. Fourteen months prior to
the bankruptcy filing Teligent’s stock was trading at nearly $100 per share, and the firm
was seen as potentially one of the most powerful CLECs in the industry. 88 The sharp
drop in its stock price left Teligent unable to secure sufficient funding to remain solvent.
The crash in Teligent’s stock price, and the subsequent financial squeeze left the
company over $1.6 billion in debt, more than $800 million of which derived from year
2000 operations.89
The reason for Teligent’s failure was over-expansion, but of a type different from
most other CLECs. Teligent’s business model was to provide voice and data services
over a fixed wireless system, thus avoiding the last mile access problem that plagued so
many CLECs. A fixed wireless system consists of a rooftop antenna that transmits a radio
signal to a receiver outside of the building. Data is then transferred to and from the end
user to the telecom’s optical network over the air rather than through copper wires. This
strategy avoids the last mile access problem, but it can be very costly.90
Teligent ran into problems when it tried to build networks in large numbers of
new markets all at once and relied too heavily on debt financing for the necessary capital
expenditures. Many of Teligent’s new markets might have eventually been very
profitable because it would have offered a service far different from that of the ILECs,
but its poor debt management resulted in a financial squeeze and subsequent bankruptcy.
88. Yuki Noguchi, “Teligent Files for Chapter 11 Protection; Move Adds to Doubt On Broadband’s
Role,” Washington Post, (May 22, 2001), at E1.
89. Elizabeth Douglas, “Teligent Is Latest Telecom to Fail, File for Chapter 11,” Los Angeles Times,
(May 22, 2001) at Business, Part3, 3.
C
R I T E R I O N
E
C O N O M I C S
, L.L.C.
-57The lesson to be taken from Teligent’s failure is that building local networks takes time,
and that markets must be added at reasonable rates so that profits from existing markets
can ease the cost of adding new markets thereby avoiding a drain of capital reserves.
D.
NorthPoint Communications
Before declaring bankruptcy and then selling its network assets to AT&T in
March 2001, NorthPoint Communications was one of the largest DSL providers in the
nation with approximately 100,000 customers. NorthPoint’s business model was to be a
be a wholesale supplier of DSL, using ILEC UNEs and selling the service to Internet
service providers who in turn enrolled the end users.91 This business model may have
made sense to the extent that NorthPoint could have captured a better margin by being
the initial producer of the service while avoiding the costs of retailing. Unfortunately, the
bursting of the Internet bubble in the stock market led to financial constraints on
NorthPoint’s clients, such as Telocity. As a result, NorthPoint had to revise downward its
third quarter 2000 earnings statement, reducing reported revenue from $30 million to $24
million because about 30 percent of NorthPoint’s clients where delinquent in paying their
bills.92
After the revised earnings statement, Verizon promptly cancelled a deal to
purchase NorthPoint due the company’s financial disarray.93 By the time the Verizon
deal had fallen through, the capital markets had sharply reduced the flow of funds to the
failing Internet firms. NorthPoint was consequently left with a partially completed
network and a huge shortfall of capital funding because it had not pursued additional
90. Id.
91. Elizabeth Douglas, “100,000 Subscribers of NorthPoint DSL Face Disconnection,” Los Angeles
Times, (March 28, 2001) at C3 (“NorthPoint DSL Face Disconnection”).
C
R I T E R I O N
E
C O N O M I C S
, L.L.C.
-58financing, counting on the Verizon deal to be completed.94 NorthPoint was forced to file
for bankruptcy protection, and eventually to sell its network elements to AT&T.
Interestingly enough, in the AT&T deal with NorthPoint, AT&T required
NorthPoint to suspend operations, ensuring that it would not have to honor contracts with
NorthPoint’s ISP clients. AT&T stated that it preferred to offer the entire service itself,
rather than acting as a wholesale agent of DSL service. 95 By providing the entire DSL
service itself, AT&T was avoiding the problem that brought NorthPoint down, namely
the failure of Internet service providers to pay their bills.
E.
Focal Communications
In 1997, Focal Communications derived over 80 percent of its total revenues from
reciprocal compensation. With uncertainty looming over a possible FCC decision to
reduce reciprocal compensation, Focal was forced reduce its dependence on these
revenues. Focal reduced its reliance on reciprocal compensation to 30 percent of revenues
in the year 2000, and hopes to reduce this figure to 15 percent of revenues in 2001. These
efforts were not sufficient to keep its stock price from declining by 80 percent in the first
half of the year 2000 as the financial markets reflected a continuing concern over cash
flow problems stemming from reliance on reciprocal compensation.96
Other companies have recognized the folly of building a business strategy on the
arbitrage opportunities presented by reciprocal compensation. For example, Intermedia
Communications reduced its expectations of revenue in 2000 as a result of expected
92. Peter S. Goodman, “Verizon Terminates Deal to Buy Stake in NorthPoint,” Washington Post,
(November 30, 2000), at E9 (“Verizon Terminates Deal”).
93. NorthPoint DLS Face Disconnection, supra note 91, at C3.
94. Verizon Terminates Deal, supra note 92, at E9
95. NorthPoint DLS Face Disconnection, supra note 91, at C3.
96. Disconnect for Upstart, supra note 40, at 1.
C
R I T E R I O N
E
C O N O M I C S
, L.L.C.
-59changes in reciprocal compensation fees.97 The expected change in fees came as a result
of state court rulings recommending the reduction of reciprocal compensation rates. This
reduction in expected revenues from reciprocal compensation was cited as one reason
why Broadwing abandoned its negotiations to buy Intermedia. As a result, the value of
Intermedia’s shares fell 14 percent in one day.98
Possibly a bigger problem than the direct loss of revenues from reciprocal
compensation is the indirect loss of revenues from poor network design resulting from
reliance on reciprocal compensation revenues. Focal initially designed its network around
extracting reciprocal compensation revenues. As a result, 100 percent of Focal’s access
lines were UNE lines, while the industry average was approximately 33 percent UNE and
36 percent on-net in the second quarter of 2000.99 Focal’s CLEC competitors were
adding their own components and building their own lines while Focal continued to lease
UNEs from the ILECs. This is a poor business strategy because Focal is even now unable
to offer product quality different from the ILECs while some CLECs are able to offer
superior service. In the long term, customers are more likely to prefer a CLEC to an ILEC
if the CLEC can offer better service, lower cost, or a combination of the two. Focal is
unable to offer service or cost improvements over the ILECs, because Focal’s entire
network is based on UNEs.
97. Intermedia Communications Inc., U.S. Securities and Exchange Commission, Form 10Q,
(November 14, 2000), 14.
98. Kris Hundley, “Intermedia Revenues Come Up Short,” St. Petersburg Times, (July 12, 2000), E1.
99. “Telecom Services—CLECs,” Credit Suisse First Boston, (Sept. 12, 2000), at 19.
C
R I T E R I O N
E
C O N O M I C S
, L.L.C.
-60VII. CONCLUSION
The empirical analysis reported in Section V and the snapshots of the successful
and unsuccessful CLECs in Section VI point in the same direction. Those entrants who
deliberately built out their own networks, carefully analyzing competition and consumer
demand prior to entry, were able to increase revenues and continue to attract capital.
Several of the more successful CLECs combined resale and the leasing of unbundled
network elements with the construction of their own networks, but none of these firms
rely exclusively on UNE or resale, and these firms added more facilities based elements
over time in order to improve upon the product the ILECs offer. The fact that some firms,
such as McLeod and Allegiance, were able to employ a resale and/or UNE strategy as
part of their business plan provides strong refutation that the large incumbent telephone
companies are in some way responsible for the recent spate of CLEC failures.
Since December 1999, the CLEC share of the nation’s access lines has expanded
rapidly. As of December 2000, the CLECs had 8.5 percent of the country’s access lines
and were growing rapidly. Unfortunately, many of the entrants were not able to survive
the large decline in the market for high-technology equity shares that began in March
2000. These companies generally had faulty business plans that were exposed when a
declining stock market severely reduced their ability to raise capital. The ensuing shakeout of entrants has been described as “only natural” by the chairman and CEO of
Allegiance, who pointed out that the overheated capital markets of 1999 and early 2000
created an environment in which “no business plan [was] to weak or management team to
C
R I T E R I O N
E
C O N O M I C S
, L.L.C.
-61inexperienced to get funded.”100 Even industry veterans agree that the recent spate of
CLEC failures is due to their own failings.
Virtually every exercise in deregulation or market liberalization leads to a wave of
entry followed by a wave of bankruptcies. This was the experience in trucking and airline
deregulation—two industries in which technology has been rather stagnant. Given the
rapid changes in technology in telecommunications and the fact that there are few
historical models of competition in local telephone service, the likelihood of failed entry
is surely much greater in this market. Nevertheless, the good news is that some entrants
are succeeding and growing and that local markets are steadily become more competitive.
100. “CLEC Representatives Have Doubts About FCC’s ‘Recp Comp’ Order,” TR Daily, May 15,
2001.
C
R I T E R I O N
E
C O N O M I C S
, L.L.C.
-62APPENDIX 1. ECONOMETRIC ANALYSIS OF CLECS
A. Analysis of Individual, Publicly Traded CLECs
To analyze the performance of individual CLECs and the industry as a whole, the
ideal model would attempt to analyze the determinants of access line and revenue growth.
Unfortunately, the only data available on the individual CLEC networks are the data that
the CLECs supply themselves. Obviously, the incentive here is for strong CLECs to
publish fairly detailed data on their networks, while the weaker firms report little, if any,
line numbers. For this reason, I formulate a model of CLEC performance that can be
applied to the publicly available data. Specifically, I am interested in the rate at which
CLECs convert investments in assets into revenues and the importance of the CLEC’s
network design and choice of customers—business or residential— in that conversion
process.
To be specific, I begin by differentiating a specific CLEC from others with an
index, i. I also refer to time periods in quarters with the index t. I define the logarithm of
a firm’s revenues in time period t as lrevt. Similarly, I define the logarithm of a CLEC’s
capital assets in time period t as lcapt. Letting N denote the total number of CLECs in the
analysis, I create dummy variables—that is, variables taking on the values of 1 or 0, to
indicate the specific CLEC that the data points correspond to. I write the N-1 dummy
variables as follows:
CLECt1 = 1 if the data correspond to the 1st CLEC, 0 otherwise
CLECt2 = 1 if the data correspond to the 2nd CLEC, 0 otherwise
CLECtN-2 = 1 if the data correspond to CLEC #N-2, 0 otherwise
CLECtN-1 = 1 if the data correspond to CLEC #N-1, 0 otherwise
C
R I T E R I O N
E
C O N O M I C S
, L.L.C.
-63The first equation estimated is:
lrevt = a0 + a1lcapt-1*CLEC(t-1)1 + a2lcapt-1*CLEC2(t-1)2 + . . . + aN-2lcapt-1*CLEC(t-1)N-2 +
aN-1lcapt-1*CLEC(t-1)N-1 + ut
(1)
In equation 1, a0, a1, a2, . . . aN-2, aN-1 are the parameters to be estimated and ut is an
error term that is drawn from a random sample in each quarter. Equation 1 allows for the
possibility that each CLEC has a different rate of converting capital assets into revenues.
An efficient CLEC will have a rapid rate of conversion and hence a large, positive
regression coefficient. For example, suppose that the 3rd CLEC is particularly efficient. In
this case, a3 will be a large positive number. On the other hand, if the 10th CLEC is
inefficient, then a10 would be close to zero, or even negative.
B.
Analysis of CLEC Business Models
The above analysis compares the performance of one CLEC to another, but gives
little, if any, insight to the business practices that lead to a CLEC’s eventual success or
failure. To measure the effects of business strategy on performance, I estimated another
regression, this time grouping the CLECs based on network platform, customer base, and
use of reciprocal compensation. I begin by defining a number of dummy variables for
quarter t, shown in Table A-1.
C
R I T E R I O N
E
C O N O M I C S
, L.L.C.
-64Table A-1
Dummy Variables Used in the Regression Analysis
Variable
Definition
UNE
1 if the CLECs main line type is UNE
0 otherwise
1 if the CLEC’s main line type is resale
0 otherwise
1 if the CLEC’s main line type is On-net
0 otherwise
1 if the CLEC’s network is split roughly
between on-net and UNE or on-net and resale
0 otherwise
Resale
On-net
Facility
1 if the CLEC’s network is split roughly
between UNE and resale
0 otherwise
1 if the firm targets business customers
0 otherwise
1 if the firm targets residential customers
0 otherwise
1 if the firm is known to rely on reciprocal
compensation revenues
0 otherwise
UNEResale
Business
Residence
RecipComp
I then use these dummy variables in the model of lagged capital assets regressed
on revenues. I estimate and equation similar to equation, but I exclude the firm specific
dummy variables and include the business specific dummy variables. This is formally
written in equation 2 that follows:
lrevt = b0 + b1lncapt-1 + b2lcapt-1*Onnet +b3lcapt-1Une + b4lcapt-1*Resale +
b5lcapt-1*Facility + b6lcapt-1*Business + b7lcapt-1*Residence + b8lcapt-1*RecipComp +
b9Onnet + b10UNE + b11Resale + b12Facility + b13Business +
b15RecipComp + et
C
(2)
R I T E R I O N
E
C O N O M I C S
, L.L.C.
b14Residence +
-65The parameters b0, b2, . . . , b10, b15 are the regression coefficients to be estimated.
Note that the variables lcapt-1*UNEResale and UNEResale are excluded from the above
equation. The effect of UNEResale on lrev, though both the slope and constant term, is
calculated by setting the dummies Onnet, Une, and Resale equal to zero. To be clearer,
the parameter b1 gives the slope coefficient for a facilities based CLEC that serves a
combination of both residential and business customers, while the parameter b0 gives the
constant term for that CLEC.
In equation 2, a large, positive coefficient on b2 would mean that on-net platforms
lead to larger rates of conversion of capital assets to revenues than for UNE.
Alternatively, a negative value for b5 would mean that the transfer of capital to revenues
tends to be poor for a CLEC that targets only residential customers.
C.
The Data
To estimate equations 1 and 2, I use quarterly financial data from 1998 to 2000
reported to the SEC for a list of publicly traded CLECs. Not all CLECs in my sample
were publicly traded during all quarters in this time frame. Some CLECs had their initial
public offerings after 1998, and some CLECs were either bought out or filed for
bankruptcy before the end of 2000. For this reason my total number of observations for
the lrev variable is 431. Further, I am able to find only 372 observations for lcap,
resulting in a regression sample of 331 observations after lagging lcap one quarter.101
Below, I include summary statistics for each of my regression variables. Table A-2
includes summary statistics for the full regression sample, while Table A-3 includes
101. The 372 observations of lcap do not all have corresponding observations of lrev. Thus, the number
of lost observations due to lagging lcap is less than the number of firms in the sample.
C
R I T E R I O N
E
C O N O M I C S
, L.L.C.
-66summary statistics for the reduced sample of 221 observations where I was able to
identify the type of network platform.
Table A-2
Sample Characteristics of Variables in Analysis, Sample of 331 Observations
Variable
lrevt
17.084
Standard
Deviation
2.078
lcapt – 1
18.863
Onnet*lcapt – 1
Mean
Minimum
Maximum
9.286
19.832
1.557
14.347
21.574
1.2502
4.814
0
20.530
UNE*lcapt – 1
5.0134
8.372
0
21.444
Resale*lcapt – 1
1.5226
4.937
0
19.863
Facility*lcapt – 1
4.2559
8.309
0
21.574
UNEResale*lcapt – 1
.8321
3.981
0
21.364
Business*lcapt – 1
12.631
9.004
0
21.456
Residence*lcapt – 1
1.432
4.939
0
21.262
RecComp*lcapt – 1
2.3195
6.464
0
21.444
Onnet
.0634
.244
0
1
UNE
.2659
.442
0
1
Resale
.0876
.283
0
1
Facility
.2085
.406
0
1
UNEResale
.0423
.202
0
1
Business
.6677
.472
0
1
Residence
.0785
.269
0
1
RecComp
.1148
.319
0
1
C
R I T E R I O N
E
C O N O M I C S
, L.L.C.
-67-
Table A-3
Sample of 221 Observations Where Network Type is Known
Variable
Standard
Deviation
1.555
Mean
Minimum
Maximum
10.937
19.831
14.347
21.574
lrevt
17.413
lcapt – 1
19.282
1.503
Onnet*lcapt – 1
1.872
5.796
0
20.530
UNE*lcapt – 1
7.509
9.291
0
21.444
Resale*lcapt – 1
2.280
5.901
0
19.863
Facility*lcapt – 1
6.374
9.487
0
21.574
UNEResale*lcapt – 1
1.246
4.823
0
21.364
Business*lcapt – 1
13.735
8.776
0
21.444
Residence*lcapt – 1
1.399
5.036
0
21.262
RecComp*lcapt – 1
3.474
7.658
0
21.444
Onnet
.0950
.294
0
1
UNE
.398
.491
0
1
Resale
.131
.338
0
1
Facility
.312
.464
0
1
UNEResale
.063
.244
0
1
Business
.715
.452
0
1
Residence
.072
.260
0
1
RecComp
.171
.378
0
1
D.
Estimation Technique
I use the technique of ordinary least squares (OLS), a statistical method that is
widely used to estimate the parameters of linear equations, to estimate equations 1 and 2.
To give a formal description of the OLS estimator in this particular case, define Y as a
T*N x 1 column vector of data for the variable lrev. T is the number of time periods, and
N is the number of CLECs. Next, define X as a T*N x K vector of observations for the
C
R I T E R I O N
E
C O N O M I C S
, L.L.C.
-68right hand side variables, where K is the number of these variables. For example, in
equation 1, K is equal to N, because there are N-1 firm specific dummy variables, in
addition to a constant term. Finally, define U as a T*N x 1 vector of error terms drawn
from a random sample with zero mean. Then we can write the equation
Y = XB + U
(3)
In equation 3, B is a K x 1 vector of regression coefficients that we attempt to
estimate. The OLS estimator for B, call it ß, is the vector of parameter estimates yielding
a line that minimizes the sum of squared error terms. In equation form, this estimator is
written as:
ß = (X’X)-1X’Y
(4)
Thus, I apply equation 4 to the relevant data to obtain my estimates of the linear
coefficients of interest.
E.
Regression Results
1.
Controlling for Individual Firms
The results of the first regression analysis are presented in Table A-4. Note that I
do not include a few publicly traded CLECs such as Universal Access, Choice One, or
Pac West because their initial offerings where not until the year 2000 and there is little
data for these firms. For most other public CLECs, however, I do have sufficient
observations to conduct the analysis. Note the highly negative and statistically significant
estimated coefficients for firms such as SpeedUS.com, Advanced Radio, Allied Riser,
and Telocity. These results mean that increases in the capital assets by these firms did not
translate into increases in revenues. Not surprisingly, these firms are all performing
poorly. SpeedUS.com has stock prices of about 60 cents per share, down from a 52 week
C
R I T E R I O N
E
C O N O M I C S
, L.L.C.
-69high of $8, and Allied Riser’s stock currently trades at around 70 cents per share, down
from $20.50 per share. Trading on Advanced Radio has been halted , and Telocity was
bought by Hughes at share prices 82 percent below its IPO value.102
The estimated coefficients for the two strongest CLECs, Time Warner and
McLeod both have positive coefficients, as one would expect. Note, however, that a
number of CLECs that are currently performing poorly have positive and statistically
significant coefficients, and therefore, this analysis does not fully sort out the successful
from the unsuccessful firms. Nonetheless, it does provide insight into a single problem
that contributed to the failure of some of these firms.
102. See Table 3, and Appendix 2.
C
R I T E R I O N
E
C O N O M I C S
, L.L.C.
-70Table A-4
The Productivity of Capital Assets in Generating Revenues
Variable
Estimated Coefficient
Lagged Log Cap Assets
------Adelphia
Business Solutions
0.372
for:
Allegiance Telecom Inc.
-0.0072
Allied Riser
-0.132
Advanced Radio
-0.219
US LEC Corp
0.027
CoreComm Ltd.
-0.0026
Convergent
0.057
Covad
-0.023
CapRock
0.051
CTC Communications Corp.
0.046
Electric Lightwave Inc.
0.0042
Focal Communications.
0.018
GST Telecommunications
0.021
ICG Telecommunications
0.041
Intermedia Communications
0.065
Inter-Tel Inc.
0.105
ITC DeltaCom Inc.
0.036
McLeod USA Inc.
0.088
Metromedia
-0.025
Mpower
-0.0201
Network Access
-0.083
Network Plus CP
0.036
NorthPoint
-0.063
North Pittsburgh
-0.0048
Net 2000
-0.047
Primus
0.103
RCN Corp.
0.035
RMI.Net
-0.011
RSL
0.124
Rhythms
-0.090
SpeedUS.Com
-0.351
Teligent Inc.
-0.091
Telocity
-0.151
Time Warner TLC
0.026
World Access
0.054
Winstar
Communications
-0.051
XO
0.027
Inc. Comm. (Nextlink)
ZTEL
-0.005
Non Firm Specific:
Time Trend
0.137
Constant Term
9.05
Sample Size
R2 (goodness of fit)
C
White-Huber t-statistic
------4.46
-0.29
-4.03
-7.45
1.04
-0.08
1.93
-0.88
1.81
1.63
0.18
0.71
0.89
1.76
2.81
3.71
1.53
3.77
-1.07
-0.53
-2.71
1.30
-2.33
-0.19
-1.28
4.29
1.49
-0.29
5.13
-3.24
-10.14
-3.79
-4.09
1.10
1.95
-2.16
1.18
-0.15
5.61
5.91
331
0.81
R I T E R I O N
E
C O N O M I C S
, L.L.C.
-712.
Analysis of Business Practices
While the above analysis gives insight into the efficiency—or non-efficiency, as the case
may be—of specific CLECs in converting capital investments into revenues, it does not
provide insight into why a CLEC will succeed or fail. In order to better determine a
CLEC’s likely outcome, I now take into account a number of specific business practices
that should affect a CLEC’s performance. To be specific, I include information on resale,
UNE leasing, reciprocal compensation, and the customer base (business, residential, or
both). This information is incorporated into the regressions through the use of dummy
variables, previously described in Table A-1. Additionally, I multiply these dummy
variables by the lagged logarithm of capital assets, an estimation technique that is
tantamount to simultaneously estimating a different linear relationship for each type of
CLEC.
If a firm targets both businesses and residents, the “Business” and “Residence”
dummy variables are both assigned a value of zero. The characteristics of these firms are
obtained from analysts’ reports, financial reports to the SEC, or other public information.
In addition, the UNE, Resale, and On-net variables are based on data from statistics
provided in the Telecom Services—CLECs report published by Credit Suisse First Boston
in April 11, 2001 and June 5, 2000. My characterization of the CLEC for each of these
variables is in Table 5 in the text.
C
R I T E R I O N
E
C O N O M I C S
, L.L.C.
-72Table A-5
The Role of Business Practices in Generating Revenues
(Dependent Variable: Log Revenue in period t)
Variable
Estimated
Coefficient
0.275
lcapt – 1
T-stat
White-Huber T
1.80
0.75
Onnet*lcapt – 1
2.602
7.53
6.71
UNE*lcapt – 1
0.717
3.61
1.90
Resale*lcapt – 1
0.272
1.26
0.70
Facility*lcapt – 1
1.352
4.67
2.46
Business*lcapt – 1
-0.126
-1.00
-1.07
Resident*lcapt – 1
0.220
1.14
0.76
RecComp*lcapt – 1
-0.135
-0.87
-1.27
Onnet
-53.448
-7.89
-6.83
UNE
-15.085
-3.92
-1.97
Resale
-6.023
-1.50
-0.78
Facility
-28.791
-4.94
-2.52
Business
2.759
1.15
1.28
Residence
-3.592
-0.96
-0.63
RecComp
2.888
0.94
1.41
Cons
13.163
4.36
1.75
Sample Size
221
2
R (Goodness of fit)
0.63
In the regression reported in Table A-5, the coefficients for the constant term and
lcapt-1 should be interpreted as representing a mixture of the resale and UNE strategy.103
103. Specific dummy variables cannot be included for this “mixed” strategy because it would make the
calculation of the coefficients impossible. In the language of econometrics, the matrix would become
“singular”.
C
R I T E R I O N
E
C O N O M I C S
, L.L.C.
-73Probably the most striking results from Table A-5 are the regression coefficients for the
on-net dummy variable and the interaction between that variable and the capital assets
variable. The coefficient on the interaction term is positive and statistically significant,
meaning that we are highly confident in our ability to estimate this coefficient. Further,
the coefficient is 2.602, an extraordinarily large value. Because the revenue and assets
variable are in logs, or “percent form”, the 2.602 means that a one percent increase in
capital assets for a CLEC with primarily on-net lines, yields an increase in revenues that
is 2.602 percent greater than revenue growth for the average CLEC.
Simply put, firms with on-net lines are able to transfer assets into revenues much
more efficiently than a CLEC with another type of platform. The coefficient on the
variable Onnet, which is equal to –53.45, reflects the high startup cost for a CLEC with
primarily on-net lines. Obviously, if a CLEC decides to build a network with mostly onnet lines, the initial fixed cost is much greater than for the typical CLEC. For this reason,
the on-net CLEC must wait until it has deployed its own facilities before it can begin
realizing large incremental increases in revenues from a state of the art network. Thus,
building a primarily on-net system is efficient in the long term, but costly in the short
term.
The above facts are even more evident when we explore the effect of UNE and
resale lines on revenues. When combined with the on-net strategy, a resale or UNE
strategy yields above average revenue growth for each increase in fixed assets, but the
growth rate is only 1.352 percent above average for each percentage point increase in
revenues, as indicated by the coefficient on Facility*lcap. Use of a predominantly resale
strategy permits revenue growth that is only 0.272 percent above average for each
C
R I T E R I O N
E
C O N O M I C S
, L.L.C.
-74percentage point increase in fixed assets. A strategy based primarily on UNE leasing
generates a revenue increase that is only 0.762 percent above average for each percentage
point increase in capital assets, and a combination of both UNE and resale yields an
increase in revenues of 0.275 from a one percent increase in capital assets. 104 In addition,
the coefficient for UNE is both large and positive, implying the initial, average revenue
growth for a reseller or a UNE type CLEC is larger than for the average CLEC. The
above analysis indicates, however, that the long term gains from UNE leasing or resale
are much smaller than that experience from building an on-net base of lines.
These results highlight the fact that a CLEC’s long term growth prospects are
maximized by building its own network. Reselling and leasing an ILEC’s network
elements may be a good way to get a foot in the door, so to speak, but it is a much better
strategy when combined with building out one’s own facilities. Without its own facilities,
an entrant has added little of value to the industry. This statement is readily evident in the
poor revenue performance for the CLECs that rely on reselling, and to a lesser extent to
those that rely on UNEs.
Turning attention to the choice of consumer base, the results in Table A-5 suggest
no significant difference between a strategy that concentrates on business customers and
one that targets residences. The coefficient for “resident” is actually greater than the
coefficient for “business,” but neither is statistically significant.
104. For reasons explained in the previous footnote, this deduction is based on the size of the estimate
of lcapt-1 in Table A-5.
C
R I T E R I O N
E
C O N O M I C S
, L.L.C.
-75APPENDIX 2. ARTICLES CITING CLEC BANKRUPTCY FILINGS AND ACQUISITIONS
Al Lewis, Even The $ 20 Million Man Couldn’t Save Convergent, THE DENVER POST
(Apr. 22, 2001) at K-01.
Convergent Communications Announces Business Plan to Accelerate EBITDA
Breakeven; -Expects to Reach EBITDA Breakeven by Year-End; -Closes Sale of Voice
Business, PR NEWSWIRE, (Jan. 29, 2001).
Covad 2000 Financials to be Reported and 10-K Filed the Week of May 7; Covad
Receives
Nasdaq
Delisting
Letter,
BUS.
WIRE
(Apr.
23,
2001).
Daniel Bogler, Richard Waters, Ebbers Has Good Reason To Dig Deep For Intermedia:
Predators Are Said To Be Circling WorldCom - Which may Explain Its Over-The-Odds
Bid, FIN. TIMES (LONDON) (Mar. 23, 2001) at 33.
George C. Ford, McLeodUSA Buys Dallas, Texas-Based Fiber Optic Company to
Increase Empire, THE GAZETTE (CEDAR RAPIDS) (Dec. 8, 2000).
IDT in Control at Teligent, THE WASH. POST (May 7, 2001) at E02.
Jennifer Davies, NorthPoint To Shut Off High-Speed Net Service; Bankrupt Company
Tells Clients To Seek Options, THE SAN DIEGO UNION-TRIBUNE (Mar. 24, 2001) at C1.
NewsEdge Reports Q1 2001 Operating Results; Content Solutions Continues to Show
Momentum; NewsEdge Electronic Publishing Technology Launched, BUS. WIRE, (May
14, 2001).
Peter Elstrom, If Anyone Can Save Excite. . . , BUS. WEEK (May 14, 2001) at 96.
Phil Porter, CoreComm Plans To Sell Businesses, THE COLUMBUS DISPATCH (Apr. 14,
2001), at 1E.
Reinhardt Krause, As Phone Start-Ups Fade, What Carriers Will Get The Spoils?
INVESTOR’S BUS. DAILY (Apr. 4, 2001) at 6.
SmartPipes Names Telecommunications Leader President and Chief Executive Officer;
Hank Nothhaft, former Chairman and CEO of Concentric and Vice Chairman of XO
Communications, Joins as Company Prepares to Launch Advanced IP Services, PR
NEWSWIRE (Apr. 23, 2001).
Richard Waters, Teligent Fails to Meet Creditors’ Deadline, FINANCIAL TIMES, Edition 2,
(May 22, 2001), at 17.
Time Warner Telecom Reports 73% Revenue Increase for the First Quarter of 2001;
GST Acquisition Completed and Integration on Track; -Eighth-Consecutive Quarter
C
R I T E R I O N
E
C O N O M I C S
, L.L.C.
-76Increasing Positive Recurring EBITDA -EBITDA Increased 44% Over First Quarter
2000, PR NEWSWIRE (May 7, 2001).
William Glanz, Bankrupt Communications Firm In Herndon, Va., Is Allowed to Borrow
$ 25 Million, THE WASH. TIMES (Apr. 11, 2001).
Bethany McLean, “Hear No Risk, See No Risk, Speak No Risk,” FORTUNE, 143(10),
(May 14, 2001), at 91-98.
Kris Hudson, “Telecom Completes Major Buy; Purchase Expands Reach of Metro Time
Warner,” DENVER POST, 2ND ED, (January 11, 2001), at C-1.
“Nextlink Pays $2.9 Billion for Concentic Network,” THE BUFFALO NEWS, CITY EDITION
(January 10, 2000), 1C.
C
R I T E R I O N
E
C O N O M I C S
, L.L.C.