.1} .. .é

$.33, _ u g .w l. ,

 

 

 

LIBRARY
Michigan State
University

 

This is to certify that the
dissertation entitled

DETERMINANTS OF CABLE SYSTEM DIVERSIFICATION
- AN INVESTIGATION OF THE US. CABLE SYSTEMS

presented by

FANG LIU

has been accepted towards fulﬁllment
of the requirements for the

Ph. D degree in MEDIA AND INFORMATION
STUDIES

 

 

 

L

ﬂ”—

” /’/ Ma7jor7Professor’s Signature
(Wm/£6

Date

MSU is an Affirmative Action/Equal Opportunity Institution

.—.—o—c-a-n-¢-n_.-.—.—o-.—---.—g—n-n-n-o-o—o-o-o-.‘.-¢‘.—---—._.—--.-o—o—.-o-n—o—n-g—n—nn-3-0-n-n—nmn-n-.----¢_n‘n---o-o—--<_...

 

PLACE IN RETURN BOX to remove this checkout from your record.
TO AVOID FINES return on or before date due.
MAY BE RECALLED with earlier due date if requested.

 

DATE DUE

DATE DUE

DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

6/01 c:/CIRC/DateDue.p65-p.15

 

 

DETERMINANTS OF CABLE SYSTEM PRODUCT DIVERSIFICATION
— AN INVESTIGATION OF THE US. CABLE SYSTEMS

By

Fang Liu

A DISSERTATION

Submitted to
Michigan State University
in partial fulﬁllment of the requirements
for the degree of

DOCTOR OF PHILOSOPHY

Department of Telecommunication, Information Studies and Media

2006

ABSTRACT

DETERMINANTS OF CABLE SYSTEM PRODUCT DIVERSIFICATION
— AN INVESTIGATION OF THE US. CABLE SYSTEMS

By
Fang Liu

Product diversiﬁcation has been happening in the cable industry on a large scale for
the past decade. The purpose of the study is to examine cable system diversiﬁcation into
non-traditional services: pay-per-view (PPV) television, high-speed intemet access, and
telephony. An econometrics model based on the Industrial Organization model (the [/0
model) is employed to explore the factors that have lead to variation in diversiﬁcation
among cable systems. Speciﬁcally, I look at how three sets of variables, namely, cable
system characteristics, market structure, and market demographics, inﬂuenced cable
system diversiﬁcation.

Binary logistic regressions were employed to identify factors that inﬂuenced cable
system diversiﬁcation into PPV, high-speed intemet access, and telephony, respectively.
Regression results indicate that systems owned by multiple system operators (MSOs) are
more likely to diversify into PPV, systems with larger basic subscriber bases are more
likely to diversify into PPV and high-speed Internet access service, systems operating in
franchise areas with more broadcast television stations receivable over-the-air (OTA) are
more likely to diversify into PPV and high-speed Internet access service, and systems
operating in franchise areas with more high-speed intemet service providers are less

likely to diversify into high-speed intemet access service.

Multiple linear regression was employed to identify the determinants of cable
system total degree of diversiﬁcation, which is deﬁned to be the number, from zero to
three, of the three non-traditional services offered by a cable system. Regression results
indicate that MSO ownership, the number of basic cable service subscribers, and the
number of broadcast television stations receivable OTA all had a statistically signiﬁcant
positive relationship to cable system total degree of diversiﬁcation, whereas the number
high-speed Internet service providers in a cable franchise area had a statistically

signiﬁcant negative relationship to cable system total degree of diversiﬁcation.

Copyright by
FANG LIU
2006

To my dear mother
Yongfen Gao
for many reasons

ACKNOWLEDGMENTS

This dissertation could not have been completed without the support and
inspiration of many people. I would like to thank my dissertation adviser, Dr. Steven S.
Wildman, for his guidance and mentorship. His suggestions and feedback helped me
immensely in my dissertation project. I would like to thank my dissertation committee
members, Dr. Charles Steinﬁeld, Dr. Johannes M. Bauer and Dr. Hairong Li, for their
advice. I also would like to thank my dear professor, Dr. Thomas A. Muth, for his belief
in me, and for all his help and encouragement.

My deepest gratitude must go to my family, dearest to my heart, for their
unconditional love and support. I would like to thank my parents, Yongfen Gao and
Hongfa Liu, for giving me the gifts of life and education and for allowing me to be as
ambitious as I wanted to. I thank my sisters, Hui Liu and Yan Liu, for providing me with
continuous source of encouragement while I pursued the goal of completing my doctoral
degree.

I thank my dear friends Jesse Barajas, Arvind Diddi, Ting Hong, and Bei Liu, for

their friendship, support, and help.

VI

TABLE of CONTENTS

LIST OF TABLES ............................................................................................................. IX
LIST OF FIGURES ........................................................................................................... XI
INTRODUCTION .............................................................................................................. 1
CHAPTER 1 CABLE SYSTEM DIVERSIFICATION ..................................................... 5
1.1 An Overview of the Cable Industry ........................................................................ 5

1.2 Cable’s Current Position in MVPD Services .......................................................... 7

1.3 Cable Provision of PPV Service ........................................................................... 10

1.4 Cable Operators in High-speed Internet Access Service ...................................... 11

1.5 Cable Provision of Telephone Service .................................................................. 13
CHAPTER 2 THE PRODUCT DIVERSIFICATION LITERATURE ............................ 15
2.1 Deﬁnitions of Product Diversiﬁcation .................................................................. 15

2.2 Diversiﬁcation Concepts ....................................................................................... l7
Motives for Diversiﬁcation ................................................................................... 17
Diversiﬁcation Relatedness .................................................................................. 18
Diversiﬁcation Modes ........................................................................................... 19

2.3 Theoretical Perspectives on Diversiﬁcation ......................................................... 20
The Market Power Perspective ............................................................................. 20

The Resource-Based Perspective .......................................................................... 21
Agency Theory ...................................................................................................... 22

2.4 Measures of Product Diversiﬁcation .................................................................... 23

2.5 Diversiﬁcation Studies of Media and Telecommunication Industries .................. 26
CHAPTER 3 THEORETICAL FRAMEWORK .............................................................. 30
3.1 The 1/0 Model ....................................................................................................... 30
The Bainsian [/0 Model ........................................................................................ 30

The New [/0 Model .............................................................................................. 32

VII

3.2 Applications of the [/0 Model in Media Diversiﬁcation Research ...................... 34

CHAPTER 4 STUDY DESIGN ....................................................................................... 36
4.1 Sampling ............................................................................................................... 36
4.2 Dependent and Independent Variables ................................................................. 38

System Characteristics .......................................................................................... 39
Market Structure ................................................................................................... 42
Market Demographics ........................................................................................... 46
4.3 A Framework for the Analysis of Cable System Diversiﬁcation ......................... 47

CHAPTER 5 DATA ANALYSIS .................................................................................... 51
5.1 Descriptive Analysis ............................................................................................. 51
5.2 Regression Results for Cable System Diversiﬁcation (Direct Effects-Only Model)
..................................................................................................................................... 53

Cable System Diversiﬁcation into PPV ................................................................ 54
Cable System Diversiﬁcation into High-speed Internet Access Service .............. 55
Cable System Diversiﬁcation into Telephone Service ......................................... 56
Cable System Total Degree of Diversiﬁcation ..................................................... 57
Summary ............................................................................................................... 59

5.3 Binary Logistic Regression for Cable Systems’ Desicions to Upgrade to HF C... 61

5.4 Regression Results of Cable System Diversiﬁcation Indirect Effects (Model with

Direct and Indirect Effects) ......................................................................................... 66
Cable System Diversiﬁcation into PPV ................................................................ 68
Cable System Diversiﬁcation into High-speed Internet Access Service .............. 69
Cable System Diversiﬁcation into Telephone Service ......................................... 70
Cable System Total Degree of Diversiﬁcation ..................................................... 71
Summary ............................................................................................................... 73

5.5 Comparison of the Direct Effects-Only Model and the Model with Both Direct
and Indirect Effects ..................................................................................................... 75
CHAPTER 6 DISCUSSIONS .......................................................................................... 77
6.1 Discussion of the Results ...................................................................................... 77
6.2 Suggestions for Future Research .......................................................................... 81
6.3 Limitations of the Study ........................................................................................ 82
REFERENCES ............................................................................................................... 108

VIII

LIST OF TABLES

TABLE 4-1 DEPENDENT VARIABLES AND INDEPENDENT VARIABLES ......... 84

TABLE 5-1 RELATIVE FREQUENCY OF CABLE SYSTEM PRODUCT
DIVERSIFICATION AND HFC UPGRADES ................................................................ 87

TABLE 5-2 RELATIVE FREQUENCY OF CABLE SYSTEM TOTAL DEGREE OF

DIVERSIFICATION ........................................................................................................ 88
TABLE 5-3 DESCRIPTIVE STATISTICS OF THE SAMPLE ...................................... 89
TABLE 5-4 CORRELATION MATRIX FOR THE VARIABLES ................................ 90

TABLE 5-5 CABLE SYSTEM DIVERSIFICATION INTO PPV (DIRECT EFFECTS
ONLY-MODEL) ............................................................................................................... 91

TABLE 5-6 CABLE SYSTEM DIVERSIFICATION INTO HIGH-SPEED INTERNET
ACCESS SERVICE (DIRECT EF F ECTS-ONLY MODEL) .......................................... 92

TABLE 5-7 CABLE SYSTEM DIVERSIFICATION INTO TELEPHONE SERVICE
(DIRECT EFF ECTS-ONLY MODEL) ............................................................................ 93

TABLE 5-8 CABLE SYSTEM TOTAL DEGREE OF DIVERSIFICATION (DIRECT
EFFECTS ONLY-MODEL) ............................................................................................. 94

TABLE 5-9 CHANGE IN PROBABILITY OF CABLE SYSTEM DIVERSIFICATION
FOLLOWING ONE UNIT INCREASE IN THE STATISTICALLY SIGNIFICANT
INDEPENDENT VARIABLES (DIRECT EFF ECTS-ONLY MODEL) ........................ 95

TABLE 5-10 THE EFFECTS OF SYSTEM CHARACTERISTICS, COMPETITION
FROM BROADCAST TELEVISION, AND MARKET DEMOGRAPHICS ON CABLE
SYSTEMS’ UPGRADE TO HFC .................................................................................... 96

TABLE 5-11 THE EFFECTS OF THE SYSTEM CHARACTERISTICS, MARKET
STRUCTURE, AND MARKET DEMOGRAPHIC VARIABLES ON CABLE
SYSTEMS’ UPGRADE TO HF C .................................................................................... 97

TABLE 5-12 CABLE SYSTEM DIVERSIFICATION INTO PPV (MODEL WITH
BOTH DIRECT AND INDIRECT EFFECTS) ................................................................ 98

IX

TABLE 5-13 CABLE DIVERSIFICATION INTO HIGH-SPEED INTERNET ACCESS
SERVICE (MODEL WITH BOTH DIRECT AND INDIRECT EFFECTS) .................. 99

TABLE 5-14 CABLE DIVERSIFICATION INTO TELEPHONE SERVICE (MODEL
WITH BOTH DIRECT AND INDIRECT EFFECTS) .................................................. 100

TABLE 5-15 CABLE SYSTEM TOTAL DEGREE OF DIVERSIFICATION (MODEL
WITH BOTH DIRECT AND INDIRECT EFFECTS) .................................................. 101

TABLE 5-16 CHANGE IN PROBABILITY OF CABLE SYSTEM DIVERSIFICATION
FOLLOWING ONE UNIT INCREASE IN THE STATISTICALLY SIGNIFICANT
INDEPENDENT VARIABLES (MODEL WITH BOTH DIRECT AND INDIRECT
EFFECTS) ....................................................................................................................... 102

LIST OF FIGURES

FIGURE 1.1 CABLE INDUSTRY INFRASTRUCTURE EXPENDITURE: 1996-2005
(IN s BILLIONS) ........................................................................................................... 103

FIGURE 1-2 CABLE INDUSTRY REVENUE FROM SUBSCRIPTION-BASED

VIDEO PROGRAMMING SERVICE: 1997-2005 (IN $ MILLIONS) ......................... 104
FIGURE 1-3 CABLE HIGH-SPEED INTERNET CUSTOMERS: 2000-2005 (IN

MILLIONS) .................................................................................................................... 105
FIGURE 1-4 CABLE PHONE CUSTOMERS: 2000-2005 (IN MILLIONS) .............. 106

FIGURE 4-1: A FRAMEWORK F OR ANALYZING CABLE SYSTEM
DIVERSIFICATION ...................................................................................................... 107

XI

INTRODUCTION

Current technological developments in media and telecommunications industries,
convergence of markets, and liberalizing legislation have removed many Of the
conventional entry barriers affecting media markets. As traditional market boundaries
and barriers began to blur, competition in these markets has greatly increased (Doyle,
2002). In response to increased competition, many media and telecommunications ﬁrms
have adopted Operations-based strategies, such as integration, diversiﬁcation, niche
products and internationalization (Picard, 2004).

In the past two decades, diversiﬁcation in media and telecommunications industries
has proceeded at a rapid pace. Many media and telecommunications ﬁrms have
diversiﬁed into new product or geographic markets in order to expand their product
portfolios, to diversify the sources of revenue, to reduce risk, to gain access to content, or
to overcome regulatory barriers (Albarran & Porco, 1990; Chan-Olmsted & Chang, 2003;
Picard & Rimmer, 1999). As a result, the literature on various aspects of diversiﬁcation,
including structural analysis of diversiﬁed ﬁrms, determinants of ﬁrms’ diversiﬁcation,
and the relationship between diversiﬁcation and ﬁrm performance, has grown over years.

Although there are plenty of studies of diversiﬁcation in the media literature, there is
a need for ﬁrrther study. First, most studies of diversiﬁcation in media and
telecommunications industries are cross-industry research, and thus pay little attention to
diversiﬁcation within one particular industry. Second, although diversiﬁcation of some
media and telecommunications industries such as the newspaper industry (Picard &
Rimmer, 1999) and the publishing industry (Kranenburg, Hagedoorn & Pennings, 2004)

has been addressed in the literature, that of other major media and telecommunications

industries, in particular, the cable television industry, has not been adequately studied.
Last, determinants of media and telecommunications ﬁrms’ diversiﬁcation have not been
extensively studied in the literature. Only a few studies have explored determinants of
media ﬁrm diversiﬁcation (Albarran & Porco, 1990; Dimmick & Wallschlaeger, 1986).

The purpose of the present study is to examine cable system diversiﬁcation into non-
traditional services. In this study, the terrn cable system diversiﬁcation is used to mean
cable systems’ offering other services in addition to subscription-based video
programming services. Three services that cable systems have diversiﬁed into are: pay-
per-viewl (PPV) television, high-speed intemet access, and telephony. In particular, this
study seeks to apply the Industrial Organization Model (the I/O model) in examining
cable system diversiﬁcation. The 1/0 model argues that the structure of the economic
market affects the conduct and performance of participants. This study aims to identify
the determinants of cable system diversiﬁcation by examining the relationship between
market structure and cable system diversiﬁcation strategies. Speciﬁcally, I look at how
three sets of variables, namely, cable system characteristics, market structure, and market
demographics, inﬂuence cable system diversiﬁcation.

Diversiﬁcation in the cable industry is worthy of investigation for the following
reasons. First, cable television is the dominant multichannel video programming
distributor (MVPD). As of June 2004, cable operators served approximately 69.4 percent
of all MVPD subscribers (FCC, 2006). Second, diversiﬁcation has been happening in the
cable industry at a large scale. In addition to the traditional video programming service,

many cable operators have offered interactive services, cable telephone service, and high-

 

] PPV is a pay cable service that allows cable television subscribers to access movies and special one-time
only events.

speed intemet access service in the past decade. Last, diversiﬁcation is an important
conceptual and practical issue for cable managers. The forces behind cable systems’
diversiﬁcation are of general management interest to practitioners in the industry.

This paper is organized as follows. Chapter 1 analyzes the cable industry’s current
positions in MVPD, PPV, high-speed intemet access, and telephone services. In this
chapter, an analysis of developments in technology, changes in regulation, and major
competitors to cable systems in providing each of the above services is provided.

The literature on diversiﬁcation and industrial organization is reviewed separately in
Chapter 2 and Chapter 3. Chapter 2 provides a review of deﬁnitions, motives, forms, and
measures of product diversiﬁcation, theoretical views on diversiﬁcation, the literature on
diversiﬁcation in media and telecommunications industries, and the literature on cable
systems’ product offerings. Chapter 3 ﬁrst provides a historical overview of the I/O
model, followed by a review of applications of the [/0 model in diversiﬁcation research,
especially diversiﬁcation studies of media and telecommunication ﬁrms.

In chapter 4, the design of the study is described. Chapter 4 ﬁrst describes the study’s
stratiﬁed random sampling technique, then the study’s dependent and independent
variables. Three binary dependent variables (0, l) are used to measure cable system
diversiﬁcation into PPV, high-speed intemet access, and telephony, respectively. The
fourth dependent variable measures a cable system’s total degree of diversiﬁcation into
the above three services. Also, how three sets of independent variables - cable system
characteristics, market structure, and market demographics — are operationalized and

constructed, and their anticipated relationships with the dependent variables, are

discussed. A framework for the analysis of cable system diversiﬁcation is presented at the
end of Chapter 4.

Data analysis method and results are reported in Chapter 5. Binary logistic
regressions were employed to identify factors that inﬂuenced cable systems’ decisions to
upgrade to higher capacity hybrid networks of ﬁber optic and coaxial cable, and cable
system diversiﬁcation into PPV, high-speed intemet access, and telephony, respectively.
Multiple linear regression was employed to identify the determinants of cable system
total degree of diversiﬁcation. In Chapter 6, I discuss the regression results, the
contributions and limitations Of the present study, and offer suggestions for future

research.

CHAPTER 1
CABLE SYSTEM DIVERSIFICATION

1.1 An Overview of the Cable Industry

Cable television began in 1948 as an alternate television service to households where
reception Of broadcast television station signals receivable over-the-air (OTA) was poor.
In cable television’s early days, cable systems only retransmitted local broadcast
television station signals. Beginning in the 19603, cable systems started experimenting
with importation of distant signals. Cable television took off in the 1970s when satellite
technology enabled video signals to be transmitted economically via satellites
nationwide. Today, the cable industry has grown into a multi-billion industry.

A cable system needs to obtain a cable franchise from a local franchising authority
(city, town, or county) to operate in that local area. A cable system pays a cable franchise
fee, which is usually 5 percent of a cable system’s gross revenue, to a local government
in exchange for the privilege of using the public right of way (Head, Spann & McGregor,
2001). Offering cable television service also requires a signiﬁcant investment in
infrastructure. At the outset, a cable system must make a substantial investment in
facilities construction, receiving and distribution equipment, and wiring neighborhoods
and households (Massey & Baran, 2001). For cable systems offering a wider variety of
services such as high—speed intemet access and cable telephone services, there is a need
to upgrade network infrastructure.

Broadband is a general term used to described high bandwidth equipment or systems.
A broadband system is able to deliver multiple channels and services to users or
subscribers. Since the passage Of the Telecommunications Act Of 1996, the cable industry

has spent billions of dollars upgrading their systems to meet rising demand for broadband

applications such as digital television, high-speed intemet access, and cable telephony
(NCTA, 2005, see Figure l-l). The upgrading is fairly expensive. In 2005 alone, the
cable industry invested $9.56 billion in upgrades. Between 1996 and 2004, the cable
industry invested about $106 billion in upgrading its facilities. The upgrades involved
rebuilding more than one million miles of cable plant with ﬁber optic technology (N CTA,
2005). It equates to about $1,300 per customer spent to upgrade cable systems and launch
new broadband services (NCTA, 2004).

Cable has dominated the multichannel video programming market for years. Now,
technological innovations have made cable diversiﬁcation into PPV, high-speed intemet
access, and telephony technically feasible and economically attractive, and regulatory
initiatives have removed most of the legal barriers that separate cable and other
telecommunications industries (Doyle, 2002). These new developments allow cable
companies to explore growth Opportunities elsewhere through diversiﬁcation. In the
following sections of this chapter, the cable industry’s current positions in MVPD, PPV,

high-speed intemet access, and telephone services are examined.

1.2 Cable’s Current Position in MVPD Services

The cable industry has enjoyed its leading position in the multichannel video
programming market for years. Cable is the largest MVPD in the US. — as of June 2005,
69.4 percent of MVPD subscribers received video programming from a cable Operator
(FCC, 2006). However, cable’s share of MVPD subscribers has been declining in recent
years. Cable operators face increasing competition in the multichannel video
programming market. Today, the US. television consumers can choose from a variety of
MVPDs, from cable companies to Direct Broadcast Satellite (DBS) service providers,
and in some markets telephone companies (FCC, 2006).

DBS is the major competitor for cable in the multichannel video programming
market. The penetration of DBS service was impeded in its early years because there
were legal limitations on satellite companies’ ability to transmit local broadcast television
station signals (Bates & Chambers, 2004). On November 19, 1999, Congress enacted the
Satellite Home Viewers Improvement Act (SHVIA). SHVIA Offers DBS service
providers the option of providing local broadcast station signals to subscribers living in
the stations’ local market area, and also grants DBS service providers the authority to
provide distant or national broadcast programming to subscribersz. DBS penetration has
been increasing since. The growth Of DBS penetration can be partially attributed to the
passage of SHVIA (FCC, 2002).

DBS has become the most serious threat to cable television in MVPD services. In
June 2004, DBS had over 23.97 million subscribers, which represented a 24.34 percent of
all MVPD households (NCTA, 2005). The top two US. DBS companies are DirecTV

and EchoStar (marketed as the DISH Network). DirecTV is the largest DBS provider and

 

2 http://www.fcc.gov/mb/shva/

the second largest MVPD. DirecTV served 14.67 million MVPD subscribers as of June
2005; the second largest DBS provider and third largest MVPD EchoStar served
approximately 11.45 million MVPD subscribers as of June 30, 2005 (NCTA, 2005).

Another competitor to cable in MVPD services is the telephone companies. The
Telecommunications Act of 1996 gave telephone companies the option of offering video
programming service service3 . Some large US. local exchange carriers (LECs) have
made substantial progress in rolling out multichannel video programming service in the
past several years. For example, Verizon Communications received franchises from many
local communities, and began Offering multichannel video programming service under
the brand name “FiOS” in some of the local communities that they are franchised to serve
(FCC, 2006).

Telephone companies have obtained statewide television ﬁ'anchises in several states.
In August 2005, the Texas legislature passed a bill that would allow telephone companies
to apply for a statewide franchise to provide television services to cities throughout Texas
(Haugsted, 2005, August 10). Statewide franchise bills are advancing in Florida,
California, New Jersey and several other states. Also, telephone companies now are
lobbying for a national franchise license for providing video programming services. The
passage of a national video franchising scheme would help telephone companies avoid
the long franchise negotiation procedure, and thus accelerate telephone companies’ entry
into the video programming market (Hearn, 2006, March 13).

The cable industry’s revenue has been growing. From 1997 to 2005, the cable
industry’s revenue from the subscription-based video programming market increased

from $24.96 billion in 1997 to $37.54 billion in 2005 (FCC, 2002, 2006, see Figure 1-2).

 

3 The Telecommunications Act of 1996, TITLE III - Cable Services, Section 302

In constant dollars, it increased by 27.30 percent from 1997 to 2005 (in 2005 dollars,
$24.96 billion from 1997 is worth $29.49 billion, using the GDP deﬂator). U.S. GDP
increased by 27.94 percent over the same period, measured in year 2000 dollars.
However, confronted with increased competition, the cable industry has experienced
declining market share in the multichannel video programming market. Cable’s market
share of the multichannel video programming market decreased from 95 percent in 1994

to 72 percent in 2004 (FCC, 2005).

1.3 Cable Provision of PPV Service

Many cable systems Offer PPV services that allow cable subscribers to watch movies
and sport events for a one-time fee at a scheduled time (Head, Spann & McGregor,
2001). The signal on each cable PPV channel is scrambled until a cable subscriber
chooses to view programming on that channel. A cable subscriber can order PPV
programming either by telephone or by remote control. Cable PPV services depend on
addressability — after receiving an order from a subscriber, a video server at the cable
headend activates an addressable converter near the subscriber’s television set (Head et
al., 2001). The addressable converter then descrambles the PPV program ordered for its
duration.

DBS service providers such as Dish Network and DirecTV Inc provide PPV services
that compete with cable PPV services. The digitally enhanced DBS channels allow DBS
service providers to Offer their subscribers more PPV programming choices than can be
Offered by cable operators Offering analog services, and the DBS services are also able to
schedule more frequent start times for repeated programs, such as movies, than can their
cable competitors. Cable PPV is limited by the number of channels that can be allocated
to PPV programming. This is a disadvantage for cable PPV in competing with DBS PPV.
The upgrade to hybrid networks of ﬁber optic and coaxial cable allows cable systems to
expand their channel capacity. With these upgraded networks, cable companies are able

to provider their subscribers with more PPV programming choices.

10

1.4 Cable Operators in High-speed Internet Access Service

Many cable operators view providing high-speed intemet access service as an
important revenue source. Most cable operators provide hi gh-speed intemet access
service with one proprietary Internet Service Provider (ISP) created and owned by the
cable operator (NCTA, 2005). For example, Comcast Communications offers the service
under the “Comcast High-speed Internet” brand name; Charter offers the service under
the “Charter High-Speed” brand name; and Cox offers the service under the “Cox High
Speed Internet” brand name (N CTA, 2005).

The cable industry has a lead over other technologies in the residential high-speed
intemet access market. The number of cable modem customers has been increasing
steadily over the past years (NCTA, 2005, see Figure 1-3). As of June 2005, the cable
industry had 23 million high-speed intemet access service customers (NCTA, 2005).
However, cable’s leading position in this market has eroded as telephone companies and
other service providers have begun to provide competing services. In the past year, cable
Operators have seen erosion in their ability to Sign up new high-speed intemet access
service customers (The FCC, 2006).

The primary high-speed intemet access service competitors to cable companies are
the telephone companies that provide digital subscriber line (DSL) service. On June 27,
2005, the Supreme Court of the United States upheld a declaratory FCC ruling that
classiﬁed high-speed Internet access service provided by cable companies as an
information service’. This decision conﬁrmed the F CC’s ruling exempting high-speed

Internet access service provided by cable companies from common-carrier obligations. In

 

4 National Cable & Telecommunications Association et a1. V. Brand X lntemet Services, declaratory ruling
and notice of proposed rulemaking, l7 FCC Red 4798 (2002) (Cable Modem Declaratory Ruling and
NPRM).

ll

order to promote competition in the high-speed Internet market, the FCC redeﬁned high-
speed lntemet access service provided by telephone companies over DSL as an
information service at year end 20055 .

In response to price competition from DSL providers, cable operators lowered cable
high-speed intemet access service rates. According to Current Analysis Inc., residential
cable high-speed Internet access service rates have fallen from an average of $44.35 per
month at the end of the fourth quarter of 2003 to an average of $40.85 per month at the
end of the ﬁrst quarter of 2004. Also, some cable systems have begun to provide “higher-
speed” connections than DSL, especially download speeds, in order to maintain their

share of the high-speed lntemet access service market (Scanlon, 2005, January 24).

 

5 FCC Eliminates Mandated Sharing Requirement on Incumbents’ Wireline Broadband lntemet Access
Services. WC Docket Nos. 05-271 and 04-242. CC Docket Nos. 98-10, 95-20, 02-33, and 01-337.

12

1.5 Cable Provision of Telephone Service

The cable industry was separated physically and legally from the telephone industry
in its early years (Bates & Chambers, 2004). Cable operators used to only provide
multichannel video programming services, whereas telephone companies used to only
provide telephone services. The Telecommunications Act of 1996 removed both cable
and telephony’s local monopoly status, and encouraged those industries to compete
directly with one another6. With the passage the Telecommunications Act of 1996, cable
companies nationwide have been certiﬁed as competitive local exchange carriers
(CLECs).

The development of technology and the trend of deregulation of telecommunication
industries have opened up the telephone market to cable companies. Cable companies
have recently expressed an increased interest in providing telephone service. They
invested in upgrading their systems to meet the technological requirements for providing
telephone service. The upgrades include two-way ﬁber lines that can send and receive
data, large quantities Of devices to convert a phone’s signals into digital cable signals, and
equipment to link cable systems to the existing public phone networks (Healey, 1998,
October 27).

Larger multiple system operators (MSOS) such as Cox Communication, Cablevision,
MediaOne, Time Warner, and Comcast have begun to offer telephone service to
customers. Most cable operators provide telephone service through a subsidiary
speciﬁcally created and owned by the cable Operator (FCC, 2006). For instance, in

Virginia, Cox Communication Offers cable telephony to its customers through its wholly

 

6 Telecommunications Act of 1996, TITLE III — Cable Services, Section 302 & Section 303

13

owned subsidiary, Cox Telcom Virginia, Inc7.

While some cable systems have Offered traditional circuit-switched telephone service
for years, many cable systems have recently launched voice over lntemet protocol (VOIP)
service. The increasing availability Of VOIP cable telephone services has attracted
thousands of new subscribers (N CTA, 2005). Despite of the difﬁculties that cable
companies confront as new entrants in the telephone market, the number Of cable
telephone customers has been growing. The number of cable telephone phone customers
has increased from 1.5 million at year end 2001 to 5.6 million at year end 2005 (NCTA,

2005, see Figure 1-4).

 

7 Retrieved July 11, 2006 from: http://www.cox.com/fairfax/telephone/faqs.asp

l4

CHAPTER 2
THE PRODUCT DIVERSIFICATION LITERATURE

Diversiﬁcation emerged in the beginning of the 20th century as a result of ﬁrms’
efforts to supply entire lines of inter-related products (Didrichsen, 1972). For several
decades, product diversiﬁcation has been a popular strategy pursued by ﬁrms all over the
industrialized world (Bengtsson, 2000; Geringer, Tallrnan & Olsen, 2000). In this chapter,
a review of deﬁnitions, motives, relatedness, modes, and measures Of product
diversiﬁcation, and the literature on diversiﬁcation in media and telecommunications
industries is presented.

2.1 Deﬁnitions of Product Diversiﬁcation

Product diversiﬁcation is a multidimensional phenomenon that has been studied
from different perspectives (Kranenbrug, 2004). Therefore, diversiﬁcation researchers
have included in the deﬁnitions of product diversiﬁcation references to various
dimensions of product diversiﬁcation, such as the criteria for ﬁrms’ diversiﬁcation, the
method of entering the new businesses, the driving forces behind diversiﬁcation, and the
levels of relatedness between a ﬁrrn’s base business to its target businesses (Brost &
Kleiner, 1995).

Product diversiﬁcation has been deﬁned in various ways in the literature. Product
diversiﬁcation in its broadest sense can be deﬁned as entering into a new business
activity by an existing business entity (Brost & Kleiner, 1995). In the early deﬁnitions,
scholars assumed industry boundaries as given, and viewed product diversiﬁcation as an
increase in the number of industries in which ﬁrms operate (Ramanujam & Varadarajan,

1989). For example, Kamien and Schwartz (1975) deﬁne product diversiﬁcation as the

15

extent to which ﬁrms classiﬁed in one base industry produce goods classiﬁed in another
industry.

Media scholars use the term product diversiﬁcation to refer to the expansion of
media companies’ business activities across different product markets (Kranenbrug,
Hagedoom & Pennings, 2004). Subscribing to this deﬁnition, media markets can be
segmented by a variety of criteria, more or less narrowly deﬁned. One possible means to
segmenting media markets is by the four-digit NAICS codes, such as wired
telecommunication carriers (5171), cable and other program distribution (5175), and
lntemet service providers (5181)8.

For the purpose of this study, cable system diversiﬁcation is deﬁned as a cable
system offering other services in addition to subscription-based video programming
services. It means if a cable system has provided a service other than subscription-based
video programming services, the system has diversiﬁed into that service. In the present
study, the focus is on the three major services that cable systems have diversiﬁed into:

PPV, high-speed intemet access, and telephony.

 

8 http://www.census.gov/epcd/naicsOZ/naicodOZ.htrn

l6

2.2 Diversiﬁcation Concepts

The diversiﬁcation literature covers a variety of topics: what are the motives for
ﬁrms to diversify? What is relatedness of diversiﬁcation? What are the modes of
diversiﬁcation? Should ﬁrms diversify through internal market and product development,
or external acquisition, or merger, or joint venture, or venture capitalization? (Brost &
Kleiner, 1995; Oster, 1999). In this section, I look at several issues involving
diversiﬁcation, including motives for diversiﬁcation and diversiﬁcation relatedness and
modes.

Motives for Diversiﬁcation

One of the central topics of diversiﬁcation research is ﬁrms’ motives for
diversiﬁcation (Brost & Kleiner, 1995). Scholars suggest a number of different motives
for ﬁrm diversiﬁcation: the search for growth, proﬁtability, economies of scope and
synergies, ﬁnancial risk reduction, and increased market power (Amit & Livnat, 1988;
Ansoff, 1965; Rumelt, 1974; Stimpert & Duhaime, 1997).

Firms may diversify to seek growth opportunities and proﬁt. For ﬁrms operating in
mature industries with slow growth rates, diversiﬁcation has been considered the only
way to promote growth (Oster, 1999). Scholars suggest that ﬁrms operating in industries
with slow growth rates, and also ﬁrms operating in declining industries characterized by
low proﬁtability and few growth Opportunities, pursue diversiﬁcation to improve their
proﬁtability (Delios & Beamish, 1999; Rumelt, 1974; Stimpert & Duhaime, 1997).

F irrns may pursue diversiﬁcation strategies to create synergies and economies of
scope (Amit, 1988; Oster, 1999). Synergy exists ‘where it is more advantageous to

combine two or more activities than to undertake them separately’ (Smith, 1985, p.69).

17

The concept of synergy is based in part on economics of scope, which exist when the cost
of joint production is less than the cost of producing each output separately (Amit &
Livnat, 1988; Brost & Kleiner, 1995). Economies of scope may be achieved through
combining manufacturing facilities, sharing a common sales force, or advertising jointly
(Ansoff, 1965).

Firms may pursue diversiﬁcation for ﬁnancial motives (Amit, 1988). The ﬁnancial
motive can be viewed from a portfolio perspective. Managing a portfolio Of different
products is expected to spread a ﬁrm’s risk across a range of products, markets, and
industries (Ito & Rose, 2004). Firms can reduce the ﬁnancial risks associated with
depending solely on a single Operation by diversifying into products and markets that are
not perfectly correlated with their primary products and markets (Amit, 1988; Brost &
Kleiner, 1995; Montgomery, 1985; Oster, 1999).

Scholars suggest that ﬁrms use diversiﬁcation as either a reactive strategy or
proactive strategy. Firms may diversify in reaction to government policies, performance
problems, or uncertainties about ﬁrture cash ﬂow or because of various managerial
motivations (Lubatkin & Rogers, 1989; Prahalad and Bettis, 1986; Ramanujam and
Varadarajan, 1989). Others consider that diversiﬁcation is used more as a proactive
strategy to limit resource dependency and ensure organizational survival (Dimrnick &
Wallschlaeger, 1986; Lubatkin & Rogers, 1989).

Diversiﬁcation Relatedness

Firms can diversify into related or unrelated businesses, or diversify by expanding

into new product or geographic markets. Rumelt (1974) refers to relatedness as the logic

and extent by which a ﬁrm’s different lines of business are connected to each other.

18

A ﬁrm’s businesses could be related either because they share markets, distribution
systems, product and process technologies, or manufacturing facilities, or because they
rely on common technologies, managerial capabilities and routines and repertoires (Grant
1988; Rumelt, 1974; Winter, 1987). Cable systems’ diversiﬁcation into PPV, high-speed
intemet access and telephone services are related diversiﬁcation. Cable systems’
diversiﬁed businesses are related to their base business of video programming primarily
through the sharing of a common distribution network.

Diversiﬁcation Modes

Diversiﬁcation can take a variety of forms, including diversiﬁcation through internal
market and product development, external acquisition, merger, joint venture or venture
capitalization, and licensing of new business activities (Brost & Kleiner, 1995).
Diversiﬁcation by acquisition requires a signiﬁcant degree of ﬁnancial commitment and
ﬁnancial health, while diversiﬁcation by expansion typically requires less, but still

signiﬁcant, up-front ﬁnancial commitments (Yip, 1982).

19

2.3 Theoretical Perspectives on Diversiﬁcation

Rumelt (1974) emphasized the study of product diversiﬁcation strategy and
generated an important line of diversiﬁcation research. Most of the new theory
development with regards to diversiﬁcation strategy extends Rumelt’s (1974) original
diversiﬁcation strategy studies. The topic of diversiﬁcation has received considerable
attention in the industrial organization, strategic management, and ﬁnance literatures. The
industrial organization literature mostly compares the proﬁtability of diversiﬁed and
undiversiﬁed ﬁrms, the strategic management literature emphasizes the beneﬁts of
diversifying into related businesses, and the ﬁnance literature mostly examines the
motives for diversifying into unrelated businesses (Amit, 1988).

Scholars have considered this topic from a range of different theoretical perspectives.
Montgomery (1994) summarized the three main theoretical perspectives that have been
advanced to explain diversiﬁcation strategies: (1) The market power perspective, which
suggests that ﬁrms use diversiﬁcation to increase their market power; (2) The resource-
based perspective, which argues that ﬁrms diversify to exploit unutilized resources
available to the ﬁrm; and (3) Agency theory, which views diversiﬁcation as a strategy
chosen by managers who are looking for more power and prestige.

The Market Power Perspective

Market power exists when a ﬁrm is able to sell its products at prices above
competitive levels. Diversiﬁcation was long seen as the consequence of the growth and
development of large, successful companies, because they can transfer their accumulated
resources from their primary market to new markets (Porter, 1985). Early literature on

diversiﬁcation based on industrial organization economics asserted that diversiﬁed ﬁrms

20

could exploit market power in selling their products and buying their raw materials (Amit
& Livnat, 1988). Economists’ initial interests in diversiﬁcation reﬂected concerns with its
potential anti-competitive effects: large diversiﬁed ﬁrms may have strategic options that
are not available to their more specialized competitors (Oster, 1999).

Different from other perspectives on diversiﬁcation, the market power view
emphasizes the anti-competitive consequences of diversiﬁcation, rather than the motives
for diversiﬁcation or efﬁciencies or inefﬁciencies that may be involved in diversiﬁcation
(Montgomery, 1994). The market power perspective suggests that diversiﬁed ﬁrms use
diversiﬁcation strategies to increase their market power through cross-subsidization,
mutual forbearance, and reciprocal buying activities (Piscitello, 2004). Diversiﬁcation
’ scholars posit that these practices may lead to reduced competition and increased industry
concentration. However, the literature only provides limited support for the view that
diversiﬁcation increases market power (Oster, 1999).

The Resource-Based Perspective

The strategic management literature examines diversiﬁcation strategy with respect to
corporate structure, internal processes and systems, and ﬁrnctional policies (Brost &
Kleiner, 1995). The resource-based view is one of the commonly accepted perspectives
on frrrn diversiﬁcation in strategy research (Montgomery, 1994; Piscitello, 2004;
Ramanujam & Varadarajan, 1989). This perspective argues that rent-seeking frrrns
diversify in response to unutilized resources — resources that exceed what is required for
the long-run proﬁtable operation of a production process (Teece, 1982). The resource-

based perspective suggest that a ﬁrm has an incentive to diversify as long as

21

diversiﬁcation provides a more proﬁtable way of employing its excess resources,
(Montgomery, 1994).

The resource-based perspective acknowledges the importance of ﬁrms’ resources in
shaping their diversiﬁcation strategies. It suggests that the portfolios of diversifying ﬁrms
are critical in predicting the resource characteristics of the industries into which they will
diversify (Montgomery & Hariharan, 1991). Leveraging resources across product lines
may produce economies of scope due to competencies that can be applied to multiple
lines of business (Geringer, Tallman & Olsen, 2000). The resource-based perspective
predicts that ﬁrms choose to enter markets that are close to their existing product lines
(Montgomery & Hariharan, 1991; Piscitello, 2004; Robins & Wiersema, 2003).

Agency Theory

Agency theory is another widely accepted perspective on ﬁrm diversiﬁcation in the
strategic management literature (Denis, Denis & Sarin, 1999). Agency theory suggests
that separation of ownership (shareholders) and control (managers) may cause corporate
assets to be deployed to beneﬁt managers rather than shareholders. Scholars applying this
perspective argue that diversiﬁcation strategies represent corporate decisions where there
exists a fundamental conﬂict of interest between managers and shareholders (Denis,
Denis & Sarin, 1999). According to agency theory, the adoption Of diversiﬁcation
strategies is predicted by personal beneﬁts that managers can derive from diversiﬁcation.
Some studies in this research stream ﬁnd a negative relationship between large
shareholdings by management and diversiﬁcation levels (Amihud & Lev, 1981; Denis,

Denis & Sarin, 1999).

22

2.4 Measures of Product Diversiﬁcation

The two most accepted measures of product diversiﬁcation are continuous business
count measures and categorical measures (Kranenbrug, Hagedoom & Pennings, 2004).
The difference between the two is that the former approximate the amount, or degree, of
diversiﬁcation while the latter approximate the type of diversiﬁcation, or the relationship
that a ﬁrm’s business units have to its base business (Reed & Sharp, 1987).

Continuous business count measures are developed from Standard Industrial
Classiﬁcation (SIC) codes, in which each of a ﬁrrn’s establishments is classiﬁed
according to its primary classiﬁcation or activity, and are considered to be more objective
(Hall & John, 1994). Researchers use both unweighted and weighted continuous business
count measures (Montgomery, 1982). In measuring ﬁrm diversiﬁcation with unweighted
continuous business count measures, researchers simply count the number of different
products or businesses a ﬁrm is operating in, while with weighted continuous business
count measures, researchers weight the products or businesses by relative size of sales or
assets.

The unweighted measures are simple to compute from data that are readily available
and do not require detailed business level sales data. At the same time, the major
weakness of the unweighted measures is that they are based on the assumption that SIC
codes are measured on an interval or ratio scale with equal distances between adjacent
SIC codes (Hall & John, 1994; Montgomery, 1982; Rumelt, 1982). The advantage of the
weighted-measures over the unweighted measures is that they take into consideration
different levels Of ﬁrm involvement in the lines of business it Operates in (Montgomery,

1982).

23

The categorical approach is a subjective way to measure diversiﬁcation, which is
based on the conceptualization of the core activities of a company, ranging on a spectrum
from single business, to dominant business, to related business, to unrelated business. For
example, Rumelt (1974) proposes a set of diversiﬁcation categories, which include single
business, dominant business, related constrained, related linked, and unrelated business.
Rumelt’s classiﬁcation system uses two steps to assign a ﬁrm to l of 10 diversiﬁcation
categories: (I) Assign a ﬁrm to a diversiﬁcation category based on the percentage of its
sales that can be attributed to a discrete business; and (2) Further differentiation is made
based on the pattern of linkages among the ﬁrm’s businesses (Rumelt, 1974). This two-
step process requires ﬁrm-speciﬁc information on activities such as marketing,
distribution, R&D, technology, production, and also ﬁrm history (Montgomery, 1982).

The two approaches both have their advantages and disadvantages. The major merit
of the continuous business count approach is that it is a continuous, quantitative measure
that can identify and measure differences in diversiﬁcation, both between ﬁrms and
across time (Grant, Jammine & Thomas, 1988). Categorical measures are superior to
continuous business count measures in the sense that they consider both qualitative and
quantitative data in making classiﬁcation decisions and thus overcome a weakness
inherent in continuous business count measures. However, the major disadvantages of the
categorical approach are that it demands detailed business-level information such as data
on sales or revenues from activities, and that it is largely based on understanding the
underlying logic behind the ﬁrm’s intentions and the assumed relatedness between

businesses. It thus has untested reliability and researcher-speciﬁc subjective

24

classiﬁcations (Kranenbrug, Hagedoom & Pennings, 2004; Hall & John, 1994; Lubatkin,

Merchant & Srinivasan, 1993).

25

2.5 Diversiﬁcation Studies of Media and Telecommunication Industries

With advances in technologies and the trend of convergence of different information
and communications markets, media companies have adapted to these changes and
responded to create or to sustain their competitive advantages. For several decades,
product diversiﬁcation has been a highly popular strategy pursued by many media ﬁrms
(Lacy, 2004). Picard (2002) points out that media ﬁrms’ diversiﬁcation was mostly
caused by the market growth and market share problems in individual media industries.

In the context of media research, diversiﬁcation can be conceptualized as the variety
of businesses in which a media ﬁrm engages, the diversity of its product Offerings, or the
range of geographic regions in which a media ﬁrm does business (Dimrnick, 2004).
Some scholars use the term ‘diagonal integration’ - common ownership across different
media sectors — to refer to diversiﬁcation (Doyle, 2004). Interests in various aspects of
product diversiﬁcation, such as structural analysis of diversiﬁed ﬁrms, the determinants
ﬁrms’ diversiﬁcation strategy, and the relationship between diversiﬁcation and economic
performance, have grown over the years in both media industry and academia.

Some studies analyze the structure of diversiﬁed media ﬁrms and industries. For
example, Dimmick and Wallschlaeger (1986) initially apply Rumelt’s (1974) index to
investigate the relationship between prior diversiﬁcation of TV network parent
companies and their future diversiﬁcation into the new media. In a later study, Albarran
and Porco (1990) adopt the Dirnrnick and Wallschlaeger (1986) index of diversiﬁcation
to measure the degree of diversiﬁcation of corporations operating pay cable channels
(Time Incorporated, Viacom Incorporated, Walt Disney Company and Playboy

enterprises). Kranenburg, Hagedoom and Pennings (2004) examine the diversiﬁcation

26

strategies of large publishing companies and conclude that the publishing companies
mostly diversify into related activities and businesses.

Another stream of study of product diversiﬁcation in media and telecommunication
industries examines the relationship between ﬁrms’ diversiﬁcation strategies and
performance (Jung, 2003; Jung & Chan-Olmsted, 2005; Kranenburg, Hagedoom &
Pennings, 2004). Jung (2003) examines the degrees of product diversiﬁcation by media
conglomerates since the Telecommunications Act of 1996 and tests the impact of product
diversiﬁcation on their ﬁnancial health. Jung and Chan-Olmsted (2005) examine the
impact of media ﬁrms’ diversiﬁcation on their ﬁnancial performance, based on the
product and geographic diversiﬁcation activities and performance of the top 26 media
ﬁrms from 1991 to 2002. Kranenburg, Hagedoom and Pennings (2004) investigate the
relationship between diversiﬁcation and performance among large publishing companies
that differed in both product and international diversiﬁcation. They ﬁnd that the large
diversiﬁed publishing companies do not outperform the more focused publishing
companies.

Last, a few studies in the literature address the determinants of media companies’
diversiﬁcation strategies. Scholars propose a variety of determinants for corporate
diversiﬁcation. Albarran and Porco (1990) posit that the degree of corporate activity
involving mergers and acquisition has an impact on media companies’ diversiﬁcation.
Eisenmann (1999) concludes that two dimensions of organizational form - a ﬁrm’s level
of diversiﬁcation and its CEO’s status as an agent versus owner-manager —- predict a
cable company’s propensity to either expand horizontally through acquisition or to exit

the cable industry. Also, scholars suggest that media companies’ future diversiﬁcation is

27

inﬂuenced by their prior diversiﬁcation experience. For example, Dimmick and
Wallschlaeger (1986) suggest that TV network parent companies that had more prior
diversiﬁcation experience are more likely to diversify into the new media. In recent
studies, determinants Of diversiﬁcation for media companies have been examined in a
more systematic manner. Using a case study approach, Chan-Olmsted and Chang (2003)
analyze the top seven global media conglomerates’ product and geographic
diversiﬁcation strategies and propose a set of determinants that might affect the extent,
relatedness, and mode of diversiﬁcation for global media conglomerates.

The cable industry is one of the major US. telecommunication industries. Some
aspects of cable system diversiﬁcation have been explored in the literature. Woroch
(1997) accesses the strategic merits of different ways that cable television might enter
into telecommunications, including direct entry, acquisition and joint venture. Albarran
and Porco (1990) study the degree of diversiﬁcation of four media corporations operating
pay cable channels and examine the relationship between the degree of corporate
diversiﬁcation and the degree of corporate involvement in acquisitions and ventures.

F ofana (1997) investigates whether the geographic market diversiﬁcation of the largest
US. cable operators affects pricing behavior.

Many other aspects of the cable industry have been addressed in the media
economics literature, especially, factors that determine cable system program offerings
(Chipsy, 2001; Dertouzos & Wildman, 1990; Goolsbee & Petrin, 2004; Kim, 1997).
These studies indicate that a variety of factors, including system speciﬁc characteristics
(e. g., the length of the system plant, the number of homes passed, age of the headend,

channel capacity, and whether or not a system is owned by an M80), market

28

demographic characteristics (e. g., ethnic audience composition, projected population
growth, employment rate, income, home ownership, and VCR ownership), and market
structure (the number of broadcast television station signals receivable OTA, DBS

penetration) predict cable system program Offerings, prices, and subscriptions.

29

CHAPTER 3
THEORETICAL FRAMEWORK

Technological advances enable cable systems to provide other services in addition to
traditional subscription-based video programming services. With their infrastructure in
place, cable systems can provide these services at a relatively low incremental cost
compared to introducing such services without a pre-existing network infrastructure.
Some of them did, some of them did not, and those that did diversiﬁed to differing
degrees. In this study, an econometrics model based on the [/0 model is employed to
explore the factors that have lead to variation in cable system diversiﬁcation.

3.1 The 1/0 Model
The Bainsian [/0 Model

Industrial organization is the study of the functioning of markets (Tirole, 1988). The
[/0 model focuses on the linkages between strategy and the external environment. A
prime example of this approach is Porter’s analysis of industry structure and competitive
positioning, which speciﬁes that the structure of the industry in which a ﬁrm chooses to
competes determines the state of competition and the context for corporate strategies, and
thus the proﬁtability of individual corporate strategies (Porter, 1979, 1980).

The ‘Old’ I/O model, or the Bainsian I/O framework of analysis, assumes a causal
link from market structure to conduct to performance. The Structure-Conduct-
Performance Paradigm (SCP Paradigm) was ﬁrst proposed by Bain (1968). The SCP
paradigm was the dominant paradigm in industrial organization from 1950 until the
19705. Analyzing industries using the SCP Paradigm often involves three factors: market

structure, market conduct, and market performance (Bustema, 1988).

30

Tirole (1988) deﬁnes a market as “ involving either a homogenous good or a group
of differentiated products that are fairly good substitutes (or complements) for at least
one good in the group and have limited interaction with the rest of the economy” (p. 13).
Market structure refers to the number and size distribution of ﬁrms in a market
(Kranenburg & Hogenbirk, 2005). The key dimensions of market structure include: (1)
Concentration of buyers and sellers, which refers to the number and size distribution of
buyers or sellers in a given market; (2) Product diﬂerentiation, which refers to the
differences perceived by buyers to exist among the products available in a given market
(3) Barriers to entry, which refers to the Obstacles new sellers must overcome to enter a
given market; (4) Cost structures, which refers to the relationship between ﬁxed costs
and variable costs in a given market; and (5) Vertical integration, which refers to the
degree of the common ownership or control of successive stages of the production and
distribution process for a good or service (Albarran, 1996; Bustema, 1988).

There are four basic models of market structures: monopoly, oligopoly, monopolistic
competition, and perfect competition. The intensity of competition varies across market
structures. According to the I/O model, competition is strongest in perfectly competitive
markets, where there are a large number of buyers and sellers producing a homogenous
product; competition is nonexistent in a monopoly market, where a single seller
dominates the market (Albarran & Dimmick, 1996).

Market conduct refers to “the policies and behavior exhibited by sellers and buyers
in a market” (Albarran, 1996, p.37). Five speciﬁc areas of market conduct include (1)
Pricing behavior, which refers to the procedures used by sellers to determine prices; (2)

Product strategy, which refers to decisions made by sellers about the design, quality, and

31

package of a product; (3) Research and innovation, which refers to the efforts made by
sellers to improve or differentiate a product from their competitors’ products; (4) Plant
investment, which refers to the resources employed to create or acquire the physical plant
for the production of goods; and (5) Legal tactics, which includes the entire range of legal
actions utilized by a ﬁrm in a given market (Albarran, 1996; Bustema, 1988).

The New [/0 Model

An important emphasis of Bainsian type [/0 has been empirical testing of the I
‘structure-conduct-perforrnance’ hypothesis. These studies focus on the impact of
industry structure on industry performance. As a consequence, initial I/O analyses did not
pay too much attention to ﬁrm conduct. The Bainsian [/0 model is criticized because it
ignores the dynamic effects that ﬁrms’ strategic actions could have on market structure
and ignores various forms of feedback from performance to structure or conduct, and thus
cannot properly depict the relationships between market participants (Chan-Olmsted,
1997; Ramsted, 1997; Young, 2000). New l/O theorists’ contributions to the Bainsian I/O
framework are: (1) Lines of causality, which address the effects of conduct and structure
on performance; and (2) Feedback loops, which make the model more dynamic and
evolutionary (Scherer & Ross, 1990).

The new [/0 model developed through the 19805 and the 1990s employs advanced
oligopoly theory based on Game Theory. The new I/O model pays more attention to ﬁrm
decisions, conduct, and strategic interaction in markets (Lacy & Bauer, 2005). Especially,
it provides a better theoretical foundation for ﬁrm behavior in conditions of oligopoly

(Young, 2000).

32

A major implication of the new I/O theory is that the assumed causal relationship
between market structure and performance is weakened and the relationship is viewed as
decided by the nature of the strategic interaction between ﬁrms (Young, 2000). Though
the new l/O model does not completely overcome some of the criticisms of the traditional
[/0 model, it can provide scholars with a powerful tool to analyze the conduct of media

ﬁrms (Lacy & Bauer, 2005).

33

 

3.2 Applications of the I/O Model in Media Diversiﬁcation Research

The [/0 model has several strengths as an explanatory tool for understanding the
functioning of media markets, including: (1) The model provides a systematic way of
examining the participants in a media market; (2) The model provides a framework for
analyzing how various market participants interact with each other; (3) The model helps
us better understand why market processes may break down; and (4) The model helps us
better understand how market performance can be improved through means other than
direct governmental control (Bustema, 1988).

The [/0 model suggests that the external environment should be the primary
determinant of a ﬁrm’s strategic actions. According to the I/O model, a ﬁrm’s optimal
expansion strategy is governed by external factors, such as the characteristics of the
industry that the ﬁrm operates in, as well as internal factors that deﬁne the constraints and
opportunities placed on the ﬁrm by its resources (Delios & Beamish, 1999; Luo, 2002).
Scholars suggest a variety of major inﬂuences on ﬁrms’ decisions to diversify, including:
(1) The general environment (i.e., the legal-political-economic-technological-social-
ecological environment in which a ﬁrm Operates); (2) The industry’s competitive
environment, such as the structure of the base industry of a ﬁrm, slowing industry grth
rates, market share erosion in a ﬁrm’s traditional market, and the structure of target
markets; and (3) Speciﬁc characteristics of ﬁrms, such as a ﬁrm’s degree of prior
diversiﬁcation, ﬁrm size and ﬁnancial strength (Abell & Hammond, 1979; Kashlak &
Maheshkumar,l994; Ramanujam & Varadarajan, 1989).

The [/0 model has been applied to media diversiﬁcation research. Some studies

examine the determinants of ﬁrm diversiﬁcation. For example, Kashlak and

34

Maheshkumar (1994) explored two external variables that constrained Regional Bell
Operating Companies’ (RBOCs) diversiﬁcation decisions: (1) Market growth in a
RBOC’s core industry of telecommunications access provision; and (2) Regional
government regulation. They concluded that RBOCs’ diversiﬁcation strategies were
inﬂuenced by the external contingencies of their domestic regulatory environments and

growth in their core business.

35

 

CHAPTER 4
STUDY DESIGN

This chapter describes the two-stage stratiﬁed sampling scheme used for sample
selection, the construction of the dependent variables and independent variables, and the
anticipated relationships between the dependent variables and the independent variables.
Also, a framework for analyzing cable system diversiﬁcation is proposed.

4.1 Sampling

As of December 2005, there were about 7,505 operating U.S. cable systems. 850
cable systems were excluded from the database for this study because of missing data.
103 overbuild cable systems were also excluded from the study. The sample size for this
study is 327, which represents about 5 percent Of the 6,552 remaining cable systems. A
two-stage stratiﬁed sampling scheme was used to select the sample for this study. First,
cable systems were stratiﬁed by the size of their owners (the number of systems owned).
Cable system owners vary in terms of the number of systems they own, ranging from one
to nearly one thousand cable systems. At this stage, all cable systems were organized in a
list starting with all cable systems owned by the cable owner that owns most cable
systems (Charter Communications with 990 cable systems), then cable systems owned by
the cable owner that owns the second most cable systems (Cebridge Connections with
745 cable systems), etc. As a result, all the one-system owners are at the bottom of the
list, then all the two-system owners next above them, etc.

At the second stage, cable systems were further stratiﬁed by cable system Size as
measured by the number Of basic service subscribers. Cable systems in the sample vary in
terms Of the number of basic service subscribers, ranging from four to 1,400,000 basic

service subscribers. The systems Of each cable system owner were listed in descending

36

order according to size, with the cable system with the most basic subscribers at the top,
the cable system with the second most basic subscribers next, etc. The two-stage
stratifying strategy ensures that a randomly-selected sample will accurately represent the
ownership and size structure of the population of cable systems as a whole.

Systematic random sampling was then performed. The sampling ratio is 1/20, which
means for the stratiﬁed list of cable systems, every 20th cable system was selected. To

avoid any possible human bias in using this method, a random number between 1 and 20

 

should be selected as the ﬁrst cable system to be selected from the list. The random

number drawn was 16.

37

4.2 Dependent and Independent Variables

This study focuses on three important services that cable systems have diversiﬁed
into: PPV, high-speed intemet access, and telephony. The unweighted business count
approach of product diversiﬁcation was used to measure cable system diversiﬁcation into
these services. A cable system’s diversiﬁcation into each service was measured by a
binary variable (0, 1), indicating whether or not a cable system has diversiﬁed into a
service. The four dependent variables in this study are: (l) D_PPV, a cable system’s
diversiﬁcation into PPV; (2) D_INT, a cable system’s diversiﬁcation into high-speed
intemet access service; (3) D_TEL, a cable system’s diversiﬁcation into telephony; and
(4) D_TOTAL, the sum of a cable system’s binary variables for diversiﬁcation into the
above three sets Of services (see Table 4-1). The ﬁrst three dependent variables measure
cable system diversiﬁcation into each individual service, whereas D_TOTAL measures a
cable system’s overall degree of diversiﬁcation. Dependent variable D_TOTAL was
treated as a continuous variable, and its value ranges from 0 to 3. D_TOTAL has value of
zero if a cable system did not diversify at all, one if a cable system diversiﬁed into at
least one Of the three services, two if a cable system diversiﬁed into any two of the three
services, and three if a cable system diversiﬁed into all three services.

As reviewed in the previous chapter, the determinants of cable system diversiﬁcation
can be located in the structure of the markets where they Operate as well as in cable
system characteristics that determine their constraints and Opportunities when they
diversify into other services. Three sets of variables that might inﬂuence cable system
diversiﬁcation are examined: system characteristics, market structure, and market

demographics. These variables are listed in Table 4-1.

38

System Characteristics

Three variables are included in the system characteristics category. Dummy variable
MSO has a value of one if a cable system is owned by an M80 and zero otherwise.
According to NCTA, MSO refers to a company that operates multiple cable systems. In
this study, if a cable operator operates more than one cable system, it is considered an
M80. On the other hand, an independent system refers to an individually owned and
operated cable system not part of an M80.

The nature of diversiﬁcation depends, in part, on the ﬁnancial resources available to
the ﬁrm (Chatterjee & Wemerfelt, 1991). Due to the substantial amount of funds required
for upgrading investments, cable systems often need external resources for funding. MSO
systems can beneﬁt from the ﬁnancial resources available from their parent companies.
For some MSO systems, diversiﬁcation is a corporate strategy. Many MSOS speciﬁcally
established subsidiaries to provide high-speed intemet access or telephone services at the
state level. Therefore, ownership of multiple cable systems inﬂuences strategic choice
regarding diversiﬁcation. The variation in diversiﬁcation between MSO systems and
independent systems, if any, can shed some light on the impact of ownership on cable
system diversiﬁcation strategies.

SUB refers to the number of a cable system’s basic service subscribers. SUB is
included because it indicates the potential demand for services offered by cable systems.
Cable systems with larger numbers of basic subscribers might ﬁnd it more proﬁtable to
diversify into other services than systems with smaller subscriber counts because they can
sell these services to their existing customers (NCTA, 2004). As a result, cable systems

with larger basic subscriber counts might be more motivated to upgrade their systems and

39

 

to diversify into other services.

DEN measures the population density of the territory served by a cable system as the
number of households per mile of cable plant. DEN is calculated by dividing the number
of households passed by a cable system by the number of miles of the cable system’s
plant. A mile of feeder plant costs the same whether it passes 10 households or 10,000
households. Therefore, the per household cost is lower for cable systems with higher
density to build and upgrade their systems. Cable systems with higher density might be
more likely to diversify into new services, motivated by a higher rate of return on
investment.

The dummy variable HFC indicates whether a cable system has upgraded to a higher
capacity hybrid ﬁber/coaxial (HF C) system or not. It indicates the technological status of
a cable system. HF C has a value of one if a cable system has upgraded to HF C system
and zero otherwise. MHz capacity of a cable system is an indicator of whether it has
already upgraded to modern HF C systems. Traditional coaxial cable systems typically
operated with 330 MHz or 450 MHz of capacity, whereas those that have expanded to
750 MHz or more are HF C systems. In this study, cable systems with a capacity of 450
MHz or less are classiﬁed as traditional coaxial systems, and those with 750 MHz or
more are classiﬁed as HFC systems.

According to industry experts consulted by this researcher, this method for
categorizing coaxial systems and HF C systems was reasonable. But a small minority of
the cable systems in the sample (about 5 percent) has a capacity of more than 450 MHz
and less than 750 MHz. Telephone interviews were conducted to ﬁnd out the upgrade

status of those systems. The telephone interviews show that the majority of those cable

40

systems have upgraded to HF C systems, and only a few have not. These systems were
classiﬁed according to their revealed status. Also, information on upgrade status for a few
cable systems could not be gathered either because the cable companies refused to release
information on their upgrade status or they could not be reached. HF C for these systems
was coded as missing data.

Fiber optics became economically viable in the early 19905, and cable systems began
using it to upgrade their systems. The cable industry adopted this technology for
upgrading a portion of its transmission plant to transform traditional coaxial systems into
HF C systems — systems consisting of a combination ﬁber optic and coaxial distribution
systems. Cable upgrades during the 19905 have focused on increasing both bandwidth
and two-way capability. Cable systems added more ﬁber optic cable to their existing
systems, and moved ﬁber nodes closer to the individual subscribers (Chiddix et al., 2000).

The cable industry typically thinks of a channel as a contiguous 6 MHz block of the
spectrum. The range of frequencies used by cable systems is from 5 MHz to 860 MHz.
Video channels are offered between the 50 — 860 MHz range, and the rest is used for
upstream communications from subscribers’ homes (Overview, 2006). To deliver data
services over a cable network, one television channel (in the 50 - 860 MHz range) is
typically allocated for downstream trafﬁc to subscribers’ homes, and another channel (in
the 5 - 42 MHz band) is used to carry upstream signals (Overview, 2006).

Cable system characteristics data were collected from The 2005 Television & Cable
F actbook (The F actbook), a standard data source. for the cable industry. The F actbook is
compiled from survey responses from approximately 8,000 US. cable franchises. For

each cable franchise listed, information is provided on its ownership, channel capacity,

41

length of plant, the number of homes with access to cable within a system’s franchise
area, the number of subscribers to basic cable, the number of broadcast television station
signals receivable OTA in a cable franchise area, and if a cable system offers PPV, high-
speed intemet access, and telephony, etc.

In the Factbook, each cable franchise area is listed under a township, city or county.
Zip codes for these cable franchise areas were collected from www.2ip-codes.com. Zip
code information was later used to collect market structure and market demographic
information for the cable systems in the sample. Most cable franchise areas in the sample
are associated with only one zip code, and a small percentage are associated with
multiple zip codes. For franchise areas with multiple zip codes, an averaging approach
was used to aggregate data for all market structure and market demographic variables
reported on a zip code leve19.
Market Structure

The market structure of cable systems’ primary business of video programming
services and of the services that they might diversify into could inﬂuence their
diversiﬁcation strategies. The structure of the video programming service market was
measured in terms of competition from broadcast and satellite television. Telephone
companies were not considered as competitors to cable companies in the video
programming service market because video programming services provided by telephone
companies are available only in a limited number of local communities.

Currently in the U.S., there are three primary technologies that deliver video

programming services to individual households: OTA broadcast television, cable

 

9 For example, for a cable franchise area with three zip codes, values of market structure and market
demographic variables reported on a zip code level is the average for the three zip code areas.

42

television, and DBS. The competition for cable in the video programming service market
is represented by the following variables: the number of broadcast television station
signals receivable OTA in a cable franchise area (STA), and DBS penetration at the state
level (DBS).

Where available, television households can choose between having access to only
OTA broadcast signals for free or obtaining the broadcast channels as well as cable
networks over a cable system or satellite for a monthly subscription fee. TV households
that do not subscribe to an MVPD service rely solely on OTA transmission of local
broadcast television signals. The FCC estimated that 15.36 million US. TV households
did not subscribe to an MVPD service as of 2004, representing 14 percent of all US. TV
households (FCC, 2005a).

Studies have found that OTA broadcast television has signiﬁcant competitive effects
on some aspects of cable service. For example, Dertouzos and Wildman (1990) found
that OTA signals inﬂuence cable systems’ market conduct such as program service
offerings and pricing. They conclude that the economic effects of OTA signals on cable
service in markets that receive ﬁve or more OTA signals are similar to those of overbuild
cable systems. More broadcast television station signals receivable OTA in a franchise
area means more competition for cable television. Competition from broadcast television
might lead to cable systems’ product diversiﬁcation by inﬂuencing cable systems’
performance in the video programming service market.

Following the rise of DBS as an alternative MVPD service, some studies have
explored the impact of satellite competition on cable program offerings (Goolsbee &

Petrin, 2004; US. General Accounting Ofﬁce (2000). In a report responding to

43

 

congressional requests, the US. General Accounting Ofﬁce found that in the calendar
year 1998, cable systems tended to provide more channels to subscribers in franchise
areas where DBS penetration was high, suggesting that cable systems responded to DBS
entry by increasing the number of channels they provide to consumers. Goolsbee and
Petrin (2004) examined the competitive effects of the introduction of DBS on cable
service by examining how cable prices and service characteristics respond to DBS. They
found that higher quality DBS in a franchise area is correlated with lower cable prices,
and they also found evidence of modest improvements in cable quality in response to
DBS entry.

As cable Operators respond to competition from DBS by lowering price and
improving quality, they might also respond by expanding their product portfolios to
generate new revenue sources. Diversiﬁcation could Offer cable Operators a lucrative
opportunity at a time when cable operators conﬁont increasing competition from DBS
companies in video programming markets (Woroch, 1997).

Cable system diversiﬁcation might be inﬂuenced by the structure of target markets.
In this study, the structure of cable systems’ target markets was measured in terms of
competition from other service providers. These structural factors are represented by the
following variables: the number of high-speed ISPs operating in the primary zip code
area associated with a cable system (INT) and the number of CLECs operating in the
primary zip code area associated with a cable system (TEL). These factors indicate the
competitiveness of the target markets that cable systems might diversify into.

Data on the number of high-speed ISPs and CLECs by zip codes were collected from

F CC’s 477 form. In order to determine the extent of broadband deployment and local

44

telephone competition throughout the U.S., in March 2000 the FCC adopted a semi-
annual reporting requirement — FCC Form 477 — for facilities-based providers of
broadband connections to end users, facilities-based providers of mobile telephone
services and all local exchange carriers. The FCC generated annual reports from the F orrn
477 that they collected. The F CC’s 2005 report, Number of High-speed lntemet Service
Providers by Zip Codes, was used to collect data on the number of high-speed ISPs
operating in the zip code areas associated with the cable systems in the sample. The

F CC’s 2005 report, Number of Competitive Local Exchange Carriers by Zip Codes, was
used to collect data on the number of CLECs operating in the zip code areas associated
with the cable systems in the sample.

The effect of the competitiveness of the target markets on cable system
diversiﬁcation is a question that has not been answered yet in the literature. On the one
hand, cable systems are less likely to diversify into a new market if that market is already
very competitive. On the other hand, cable systems might enter a new market and
compete with existing service providers in that market, due to the desire for new revenue
source regardless of how competitive it is, or due to the threat of competitor’s entry into
cable’ primary market of video programming service. Diversiﬁcation allows cable
companies to offer service bundles, including triple-play services — video, data and voice
— when their competitors do, and thus helps cable companies retain their existing
customers. These competing hypotheses are tested in this study.

Market Demographics
In order to control for variation in customers’ demands across cable franchise areas,

market demographic variables were included in the model. It is important to control for

45

market demographic variables because they reﬂect unobserved tastes for services
(Goolsbee & Petrin, 2004). Market demographic characteristics controlled for include:
percentage of Black or African American population (BAA), percentage of Hispanic
population (HIS), median household income (INC), and home ownership rate (OWN). I
also controlled for the geographic region in which a cable franchise was located —
northeast, midwest, south or western US. (REG).

Market demographic data at the zip code level were collected from American
Community Survey (ACS) provided by the US. Census Bureau. The ACS collects
survey data each year to measure key social, economic, and housing characteristics for
the US. population. In the ACS reports, data on basic housing and population
characteristics, such as ethnic audience composition, projected population growth rates,
median household income, and home ownership rates, are reported at both the county and
city levels. Demographic data for the zip code areas associated with the cable systems in

the sample were collected.

46

4.3 A Framework for Analyzing Cable System Diversiﬁcation

Based on the diversiﬁcation literature and the [/0 model, a framework for analyzing
cable systems’ upgrade to HF C and cable system diversiﬁcation is proposed in this study.
The framework proposes that three sets of independent variables — system characteristics,
market structure, and market demographics — inﬂuence cable systems’ decisions to
upgrade to HF C and cable system diversiﬁcation. These three sets of variables might
inﬂuence cable system diversiﬁcation in the following two ways: (1) Direct effects on
cable system diversiﬁcation; and (2) Indirect effects on cable system diversiﬁcation via
HFC. The framework is depicted in Figure 4-1. The solid lines in Figure 4-1 describe
direct effects of these variables on cable system diversiﬁcation, and the dashed lines
describe potential indirect effects of these variables on cable system diversiﬁcation
through HF C.

Cable systems upgrade to HF C for different reasons. Cable systems may invest in
upgrading to HF C because it is the best way to offer PPV, high-speed intemet access and
telephone services. Also, they may upgrade to HF C for some other motives, such as to
expand channel capacity and to create more tiered video programming options for their
subscribers. However, once a cable system has upgraded to HF C, the cost of upgrades
required for offering other services is considerably lower than it was before. It is possible
that cable systems upgraded to HF C for reasons other than diversifying into the three
services considered here, but subsequent possession of HF C plant inﬂuenced the
decisions regarding diversiﬁcation. These two sets of motives for cable systems’ upgrade
to HF C open up the possibility of direct effects and indirect effects of the other

independent variables on cable system diversiﬁcation. Based on the above two sets of

47

motives, two models were examined in this study: a direct effects-only model and a
model with direct and indirect effects. Two sets of regressions are employed in the
following sections to test these two models.

HF C is different from other independent variables in the sense that it is an
endogenous variable acting as an independent variable in this study. My framework
hypothesizes direct and indirect effects of the e other independent variables on cable
system diversiﬁcation. Cable system characteristics, market structure, and market
demographic variables may have only direct effects on cable system diversiﬁcation if a
cable system’s initial objective in upgrading to HF C was not diversiﬁcation into new
services. On the other hand, these variables may have direct and indirect effects on cable
system diversiﬁcation if a cable system’s initial Objective in upgrading to HFC was
diversiﬁcation into new services but factors that make diversiﬁcation into one or more of
the new services studied here attractive also made upgrading to HF C plant attractive at an
earlier point in time.

Market structure variables, including STA, DBS, INT, and TEL, might inﬂuence
cable systems’ decisions to upgrade to HF C and diversify. Cable system diversiﬁcation
might be inﬂuenced by the structure of cable’s primary business of video programming
service and that of its potential target markets. The competitiveness of cable’s primary
business is measured by STA and DBS. The competitiveness of cable’s target high-speed
intemet access market is measured by INT, and the competitiveness of cable’s target
telephone market is measured by TEL. The U0 model suggests that the structure of a
target market inﬂuences the attractiveness Of that market and also imposes constraints on

ﬁrms’ diversiﬁcation into that market. In addition to STA and DBS, INT might inﬂuence

48

cable system diversiﬁcation into high-speed intemet access service, and TEL might
inﬂuence cable system diversiﬁcation into telephone service. All of the four market
structure variables are proposed to have an inﬂuence on D_TOTAL, cable system total
degree of diversiﬁcation.

The set of market demographic variables includes BAA, HIS, EMP, INC, OWN
and REG. These variables indicate the variation in customers’ demands for services
provided by cable systems across cable franchise areas. The literature on cable system
product offerings suggests that cable product offerings are inﬂuenced by the demographic
characteristics of the markets that they operate in. The same market demographics that
inﬂuence cable product offerings might also inﬂuence cable system diversiﬁcation,
because diversiﬁcation is a means for cable systems to respond to the changes in the
environment by expanding the scope of their product offerings.

Based on the preceding analysis, the following six research questions are addressed:

RQl: What factors inﬂuence the likelihood that a cable system will upgrade to HFC?

RQ2: How do cable system characteristics including ownership, the number of basic
service subscribers, density, and whether upgraded to HF C inﬂuence cable system
diversiﬁcation into PPV, high-speed lntemet access, and telephone services?

RQ3: How does the competitiveness of the video programming market inﬂuence
cable system diversiﬁcation into PPV, high-speed lntemet access, and telephone services?
RQ4: How does the competitiveness of the local high-speed lntemet access market

inﬂuence a cable system’s decision on whether to diversify into high-speed lntemet

access service?

49

RQS: How does the competitiveness of the local telephone market inﬂuence a cable
system’s decision on whether to diversify into telephone service?

RQ6: What factors inﬂuence the number of services that cable systems diversify into?

50

CHAPTER 5
DATA ANALYSIS

5.1 Descriptive Analysis

Table 5-1, Table 5-2 and Table 5-3 present the descriptive statistics for the
dependent and independent variables. Table 5-1 shows that as of 2005, 149 (45.6 percent)
of the cable systems in the sample diversiﬁed into PPV, 164 (50.2 percent) diversiﬁed
into high-speed intemet access service, and only 27 (8.3 percent) diversiﬁed into
telephony.

Table 5-2 presents frequency of cable system total diversiﬁcation scores, D_TOTAL.
It shows that 132 of the cable systems in the sample (40.4 percent) had not diversiﬁed at
all, 73 (22.3 percent) diversiﬁed into only one of the three services considered in this
study, 103 (31.5 percent) diversiﬁed into two of the three services, and only a small
minority of the cable systems — 19 (5.8 percent) — diversiﬁed into all of the three services.
Most, but not all of the cable systems that diversiﬁed into telephony also diversiﬁed into
PPV and high-speed intemet access services.

Table 5-3 presents descriptive statistics for the independent variables. As regards the
system characteristics, the majority —- about 92 percent — of the systems in the sample
were MSO-systems, and 8 percent were independent systems. The average number of
basic service subscribers was 9,517, and the average plant density was 55
households/mile. Upgrading to HP C was not that common among the cable systems —
about 31 percent of the cable systems have upgraded to higher capacity HF C systems.

As for the market structure variables, on average, there were 8 broadcast television
station signals receivable OTA, 4 high-speed lSPs, and 4 CLECs in operation in cable

franchise areas. On average, DBS penetration was about 23 percent.

51

For the market demographic variables, on average, Aﬁican Americans and Hispanics
were 8.6 percent and 5.6 percent of the populations in zip code areas associated with
cable franchise areas respectively. The average home ownership rate was 61 percent.
Average median household income was $33,850. As for geographic distribution, 28 (8.6
percent) of the systems were in northeast U.S., 119 (36.4 percent) were in the midwest,
122 (37.3 percent) were in the south, and 58 (17.7 percent) were in the west.

Table 5-4 presents the correlation matrix for the variables.

52

5.2 Regression Results for Cable System Diversiﬁcation (The Direct Effects-Only
Model)

In this section, cable system diversiﬁcation was modeled as if only direct effects as
depicted in ﬁgure 4-1 were possible. This section presents the results of regressing cable
system diversiﬁcation into each of the three services (D_PPV, D_INT and D_TEL) and
cable system total degree of diversiﬁcation (D_TOTAL) on cable system characteristics
excluding HF C, market structure variables, and market demographic variables. Binary
logistic regressions were employed to identify factors that inﬂuenced cable system
diversiﬁcation into PPV, high-speed intemet access, and telephony, respectively.
Multiple linear regression was employed to explore the determinants of cable system total

degree of diversiﬁcation.

53

Cable System Diversiﬁcation into PPV

Diversiﬁcation into PPV is the most closely related type of diversiﬁcation for cable
systems. The primary difference between PPV and cable’s primary business of video
programming services is that PPV is not subscription-based, but priced per consumption
episode. Equation (1) assumes that a cable system’s diversiﬁcation into PPV (D_PPV), is
determined by cable system characteristics excluding HF C, the structure of the system’s
video programming service market, and the demographics of its franchise area.

(1) D_PPV = a0 + ouMSO + (1.28113 + orgDEN + ouSTA + asDBS + a6BAA + a7HIS
+ agEMP + aglNC + aloOWN + auREG + 61.

Results for the binary logistic regression for D_ PPV are described in Table 5-5.
Overall, the model correctly categorized 84.5% of the cases. The percent in the modal
category for D_PPV was 54.4%. The model correctly predicted 30.1% more of the cases
compared to the mode value frequency. MSO, SUB and STA were the only three
independent variables for equation (1) with statistical signiﬁcance at the .01 level or
higher. The odds ratio for M80 is 28.08, indicating that the odds of an M80 system
offering PPV service is about 28 times the odds of an independent system offering PPV
service, holding other independent variables constant. The odds ratio for SUB is 3.92. It
means that for every 1,000 increase in a cable system’s basic service subscribers, the
odds of a cable system offering PPV service increases by a factor of 3.92, holding other
independent variables constant. The odds ratio for STA is 1.22. It means adding a
broadcast television station receivable OTA to a cable franchise area increases the odds
Of a cable system offering PPV service by a factor of 1.22, holding other independent

variables constant.

54

Cable System Diversiﬁcation into High-speed Internet Access Service

Equation (2) assumes that D_INT, a cable system’s diversiﬁcation into high-speed
intemet access, is determined by cable system characteristics excluding HF C, the
structure of the system’s video programming service market, the structure of the local
high-speed intemet market, and the demographics of its franchise area.
(2) D_INT = a0 + GIMSO + azsUB + agDEN + ouSTA + asDBS + aélNT + 0L7BAA

+ agHIS + orgEMP + amINC + anOWN + alzREG + 62.

Results for the binary logistic regression analysis for D_INT are described in Table
5-6. Overall, the model correctly categorized 79.1% of the cases. The percent in the
modal category for D_INT was 50.2%. The model correctly predicted 28.9% more of the
cases compared to the mode value frequency. SUB, STA and INT were statistically
signiﬁcant at the level of .01 or higher. The odds ratio for SUB is 3.20. It means that for
every 1,000 increase a cable system’s basic service subscribers, the odds of the cable
system Offering high-speed intemet access service increases by a factor of 3.20, holding
other independent variables constant. The odds ratio for STA is 1.17. It means adding a
broadcast television station receivable OTA to a cable ﬁ'anchise area increases the odds
Of a cable system offering high-speed intemet access service by a factor of 1.17, holding
other independent variables constant. The odds ratio for INT is 0.61. It means that with
the addition of one more high-speed ISP to a cable franchise area, the Odds of the cable
system offering high-speed intemet access service decreases by a factor of 0.61, holding

other independent variables constant.

55

Cable System Diversiﬁcation into Telephone Service

Equation (3) assumes that D_TEL, a system’s diversiﬁcation into telephony, is
determined by cable system characteristics excluding HFC, the structure of the system’s
video programming service market, the structure of the local telephone market, and the
demographics of its ﬁ'anchise area.
(3) D_TEL = a0 + alMSO + GZSUB + a3DEN + ouSTA + asDBS + aéTEL + ouBAA

+ agHIS + agEMP + aloINC + omOWN + omREG + e;.

Results for the binary logistic regression analysis for D_TEL are presented in Table
5-7. Overall, the model correctly categorized 91.8% of the cases. The percent in the
modal category for D_TEL was 91.7%. The model correctly predicted only 0.01% more
of the cases compared to the mode value frequency. The model does not improve much
over just using modal category. There was no statistically signiﬁcant independent

variable for D_TEL in the model.

56

Cable System Total Degree of Diversiﬁcation

Equation (4) assumes that a cable system’s total degree of diversiﬁcation,
D_TOTAL, is determined by cable system characteristics excluding HFC, the structure of
the system’s video programming service market, the structure of the local telephone
service market, the structure of the local high-speed Internet market, and the
demographics of its franchise area.

(4) D_TOTAL = a0 + ouMSO + a2SUB + (13DEN + ouSTA + asDBS + a6INT
+ a7TEL + agBAA+ angS + aloEMP + anch + (1.120WN + ornREG + e4.

Multiple linear regression was employed to identify the determinants of cable system
total degree of diversiﬁcation. The distribution of SUB and INC shows some skewness to
the right, suggesting that a log version of the model might improve the ﬁt of the model.
Therefore, a log version of the model where LNSUB, the natural log of SUB, and LNIN C,
the natural log of INC, were used instead of SUB and INC was run and compared to the
original model. The log version of the model showed a better ﬁt.

The correlation matrix table, Table 5-4, shows a high positive, and statistically
signiﬁcant, correlation coefﬁcient for INT and TEL (R= 0.728). Also, in a regression
collinearity diagnosis, INT and TEL show low tolerance values"), 0.401 and 0.302,
respectively. Therefore, TEL was eliminated from Equation (4) to control for
multicollinearity. The revised model is presented in Equation (5):

(5) D_TOTAL = oro + onMSO + azLNSUB + a3DEN + a48TA + asDBS + a6lNT

+ (17BAA + agHIS + (lgEMP + aioLNINC + (InOWN + alzREG + 65.

 

'0 Tolerance values take values between 0 and l. Tolerance values closer to 0 indicate higher collinearity.

57

Results for the multiple linear regression analysis for D_TOTAL are presented in
Table 5-8. Overall, the model explained about 61.9 percent of the variation in the
dependent variable D_TOTAL. MSO, LNSUB, STA and INT were statistically
signiﬁcant at the .01 level or higher. Regression results show that MSO systems on
average scored .09 higher than independent systems on D_TOTAL, holding other
independent variables constant. SUB had the largest positive effect on D_TOTAL. For
every percent increase in a cable system’s basic service subscribers, D_TOTAL increases
by 0.73, holding other independent variables constant.

The coefﬁcient for STA was positive and signiﬁcant. For each additional broadcast
television station signal receivable OTA in a cable franchise area, D_TOTAL increases
by .17. To the contrary, INT, another market structure variable, had a statistically
signiﬁcant negative relationship to D_TOTAL. For each additional high-speed ISP in a

cable franchise area, D_TOTAL decreases by .11.

58

Summary

As shown in Table 5-1, 45.6 percent of the cable systems in the sample diversiﬁed
into PPV, 50.2 percent diversiﬁed into high-speed intemet access service, and 8.3 percent
diversiﬁed into telephony. How would one unit increase in the statistically signiﬁcant
independent variables inﬂuence the probability of a statistically representative cable
system’s diversiﬁcation into PPV, high-speed intemet access and telephone services?
Table 5-9 shows the estimated change in the probability of cable system diversiﬁcation
following one unit change in the independent variables that were statistically signiﬁcant
at the 0.01 level or higher for the model that only allows for direct effects.

For a statistically representative cable system in this sample, the probability of
offering PPV service increases from 0.456 to 0.959 if the ownership changes from
independent to MSO, holding other independent variables constant. The probability of a
statistically representative cable system in the sample diversifying into PPV increases
from 0.456 to 0.767 if the number of its basic cable service subscribers increases by
1,000, holding other independent variables constant. The probability of a statistically
representative cable system in the sample diversifying into PPV increases from 0.456 to
0.506 if one more broadcast television station receivable OTA is added to its cable
franchise area, holding other independent variables constant.

The probability of a statistically representative cable system in the sample
diversifying into high-speed intemet access service increases ﬁ'om 0.502 to 0.763 if the
number of its basic cable service subscribers increases by 1,000, holding other
independent variables constant. The probability of a statistically representative cable

system in the sample diversifying into high-speed intemet access service increases from

59

0.502 to 0.542 if one more broadcast television station receivable OTA is added to its
cable franchise area, holding other independent variables constant. The probability of a
statistically representative cable system in the sample diversifying into high-speed
intemet access service decreases from 0.456 to 0.382 with the addition of one more high-

speed ISP to its cable franchise area, holding other independent variables constant.

60

5.3 Binary Logistic Regression for Cable Systems’ Decision to Upgrade to HF C

In this section, the question of why cable systems invested in upgrading their
systems to HF C is explored. While interesting in its own right, the results of this analysis
are employed in the two-stage least squares regressions examining the economics of
cable system diversiﬁcation into PPV, high-speed intemet access, and telephony in the
following section. Binary logistic regression analysis was employed to examine which
cable system characteristics, market structure variables, and market demographic
variables inﬂuenced cable systems’ decisions whether to upgrade to HF C or not.

Cable systems may have been motivated to upgrade to HF C for different reasons.
Two plausible hypotheses regarding the motives for cable systems’ decisions to upgrade
to HF C were examined in this study. (1) Cable systems upgraded to HF C plant as part of
their plans for diversiﬁcation into one or more of the new services examined in this study:
PPV, high-speed intemet access, and telephony. (2) Alternatively, systems may have
upgraded to HF C for reasons unrelated to diversiﬁcation into one of these services.
However, once in possession of HF C plant, they may have found diversiﬁcation into one
or more of the new services more appealing when the opportunity arose at a later date.

Two equations corresponding to the two alternative hypotheses, equation (6) and
equation (7), were speciﬁed for analyzing cable systems’ decisions to upgrade to HF C .
Because the same (relatively slow-changing) factors describing systems and market
features may have inﬂuenced both types of decisions, the same system characteristics and
market demographic variables were included in both equations. The difference is in the
competition variables. Equation (6) only includes competition from broadcast television

as a competition factor inﬂuencing cable systems’ decisions to upgrade to HF C, whereas

61

equation (7) includes variables reﬂecting competition from broadcast television,
competition ﬁ'om DBS companies in the multichannel video programming market, and
the competitiveness of the local high-speed intemet and local telephone service markets.

Equation (6) assumes that whether a cable system has upgraded to HF C (HF C=1) is
determined by cable system characteristics, the number of broadcast television station
receivable OTA in its cable franchise area, and the demographics of its franchise area.
(6) HFC= a0 + ouMSO + OZSUB + (13DEN + a4STA + asBAA + a6HIS

+ a7EMP + agINC + a90WN + aioREG + e6.

Odds and odds ratios are important basic terms in binary logistic regression. Odds is

the ratio of the probability something is true divided by the probability that it is not. Odds

ratio is the ratio of two odds. For binary logistic regressions, the odds ratio is a summary

measure of the effect size of the independent variable(s) on the dependent variable’s odds.

For example, if the odds ratio for an independent variable is 3, we may say that the odds
that the dependent variable is a one increases by a factor of 3 with one unit increase in the
independent variable, holding other independent variables constant.

The odds ratio takes values between zero and inﬁnity. An odds ratio of one means
that there is no change in the dependent variable due to a change in the independent
variable. An Odds ratio larger than one indicates a positive relationship between an
independent variable and a dependent variable, whereas an odds ratio smaller than one
indicates a negative relationship between an independent variable and a dependent
variable. An Odds ratio close to zero or inﬁnity indicates a larger change.

Table 5-10 presents the logistic regression coefﬁcient, standard error of the

coefﬁcient, the Wald statistic, probability level, and the Odds ratios for each independent

62

variable. Overall, the model correctly categorized 83.5% of the cases. The percent in the
modal category for HF C was 69.0%. The model correctly predicted 14.5% more of the
cases compared to the mode value frequency. SUB, STA and OWN were the only three
independent variables for equation (6) with statistical signiﬁcance at the .01 level or
higher. The odds ratio for SUB is 1.18, indicating that for every 1,000 increase in the
number of basic service subscribers, the odds of a cable‘system’s upgrade to HF C system
increases by a factor of 1.18, holding other independent variables constant. The odds ratio
for STA is 1.17. It means adding a broadcast television station receivable OTA to a cable
franchise area increases the Odds of a cable system upgrading to HP C system by a factor
of 1.17 , holding other independent variables constant. The odds ratio for OWN is 0.02. It
means for every percentage increase in the home ownership rate in a cable franchise area,
the odds of a cable system’s upgrade to HFC system decreases by a factor of 0.02,
holding other independent variables constant.

Equation (7) assumes that whether a cable system has upgraded to HF C (HF C=1) is
determined by cable system characteristics, the structure of the system’s video
programming service market, the structure of the system’s target markets, and the
demographics of its franchise area.

(7) HFC: a0 + onMSO + a2SUB + (13DEN + 0L4STA + asDBS + aélNT + a7TEL
+ agBAA + agHIS + aloEMP + anlNC + a120WN + a13REG + e7.

Results for the binary logistic regression for HF C are described in Table 5-11.
Overall, the model correctly categorized 83.5% of the cases. The percent in the modal
category for HF C was 69.0%. The model correctly predicted 14.5% more of the cases

compared to the mode value frequency. The odds ratio for SUB is 1.20, indicating that

63

for every 1,000 increase in the number of basic service subscribers, the odds of a cable
system’s upgrade to HF C system increases by a factor of 1.20, holding other independent
variables constant. The odds ratio for STA is 1.17. It means adding a broadcast television
station receivable OTA to a cable franchise area increases the Odds of a cable system
upgrading to HF C system by a factor of 1.17, holding other independent variables
constant. The odds ratio for OWN is 0.01. It means for every percentage increase in the
home ownership rate in a cable franchise area, the odds of a cable system’s upgrade to
HF C system decreases by a factor of0.01, holding other independent variables constant.
In addition to SUB, STA, and OWN, INT was an independent variable for equation (7)
with statistical signiﬁcance at the .05 level. The odds ratio for INT is 0.76. It means
adding a high-speed ISP to a cable franchise area decreases the odds of a cable system
upgrading to HF C system by a factor of 0.76, holding other independent variables
constant.

The difference in the results for the binary logistic regression analysis for HF C using
equation (6) and equation (7) is that the results for equation (7) reveal that the number of
high-speed ISPs in a local market had a statistically signiﬁcant negative relationship to
HF C. It indicates that the competitiveness of the high-speed intemet access service
market inﬂuenced cable systems’ decisions to upgrade to HF C. This result is consistent
with the difference in the results for the binary logistic regression for D_INT using
equation (2) and the results for the multiple linear regression for D_TOTAL using
equation (5) in the sense that when included, a variable representing the competitiveness
of high-speed intemet access service had a statistically signiﬁcant negative relationship to

cable systems’ decisions to upgrade to HFC in a HFC upgrade regression or to diversify

64

into high-speed intemet access service in the corresponding diversiﬁcation regression. 1f
cable systems upgrade to HF C in part to facilitate diversiﬁcation into high-speed intemet
service and if the likelihood of diversifying into high-speed intemet service diminishes
with the number of competitors offering this service (strength of competition), then we
should expect the likelihood of HF C upgrades to diminish as competition in high-speed

intemet access increases.

65

5.4 Regression Results for Cable System Diversiﬁcation (The Model with Direct and
Indirect Effects)

In this section, cable system diversiﬁcation was modeled as if both direct and
indirect effects were possible. This section presents the results of regressing cable system
diversiﬁcation into each of the three services (D_PPV, D_INT and D_TEL) and cable
system total degree of diversiﬁcation (D_TOTAL) on cable system characteristics
including HF C, market structure variables, and market demographic variables.

As mentioned in Chapter 4, HF C is an endogenous variable acting as an
independent variable in the diversiﬁcation equations. As a result, the error terms in the
diversiﬁcation equations are not independent of the error term in the HF C equation.
Therefore, two-stage least squares regression analysis was used for the set of
diversiﬁcation regressions in this section to address the issue of endogeneity that would
arise if cable systems upgraded to HF C for reasons other than diversifying into the three
services considered here, but subsequent possession of HF C plant inﬂuenced the
decisions regarding diversiﬁcation. Endogeneity is an issue because many of the factors
inﬂuencing cable systems’ diversiﬁcation decisions are likely to inﬂuence their decisions
to upgrade to HF C. Each of these factors changes relatively slowly over time, and this
study employs a single year’s observations. Therefore, simultaneous equations techniques
were employed to explore the possibility of endogeneity.

Two-stage least squares analysis includes two regression analyses in two steps:
(1) Regress each endogenous variable acting as an independent variable on all the
exogenous variables in the system of simultaneous equations, and save the predicted

values for these endogenous variables; and (2) Regress the dependent variable on these

66

predicted values and other exogenous variables included in the system of simultaneous
equations (Kennedy, 1998).

Two-stage least squares regression analysis was employed for each of the four
dependent variables measuring diversiﬁcation. HF C was ﬁrst regressed on all the
exogenous variables in the system of simultaneous equations at step one, and predicted
values for HF C, PRE_HF C, were calculated using equation (7). At step two, a
diversiﬁcation variable was regressed on PRE_HF C and the other exogenous variables

included in the system.

67

Cable System Diversiﬁcation into PPV

Equation (8) assumes that a cable system’s diversiﬁcation into PPV (D_PPV), is
determined by cable system characteristics, the predicted values for HFC, the structure of
the system’s video programming service market, and the demographics of its franchise
area. Binary logistic regression was employed to identify important factors inﬂuencing
cable system diversiﬁcation into PPV.
(8) D_PPV = a0 + ouMSO + a2SUB + a3DEN + a4PRE_HF C + assTA + aoDBS

+ a7BAA + agHIS + agEMP + aloINC + or; IOWN + onsz + 63.

Results for the binary logistic regression for D_ PPV are presented in Table 5-12.
Overall, the model correctly categorized 85.1% of the cases. The percent in the modal
category for D_PPV was 54.4%. The model correctly predicted 30.7% more of the cases
compared to the mode value frequency. MSO, SUB and STA were the only three
independent variables for equation (8) with statistical signiﬁcance at the .01 level or
higher. The odds ratio for MSO is 25.49, indicating that the odds of an MSO system
offering PPV service is more than 25 times the odds of an independent system offering
PPV service, holding other independent variables constant. The odds ratio for SUB is
3.82. It means that for every 1,000 increase in a cable system’s basic service subscribers,
the odds of a cable system offering PPV service increases by a factor of 3.82, holding
other independent variables constant. The odds ratio for STA is 1.20. It means adding a
broadcast television station receivable OTA to a cable franchise area increases the odds
of a cable system offering PPV service by a factor of 1.20, holding other independent

variables constant.

68

Cable System Diversiﬁcation into High-speed Internet Access Service

Equation (9) assumes that D_INT, a cable system’s diversiﬁcation into high-speed
intemet access, is determined by cable system characteristics, the predicted values for
HF C, the structure of the system’s video programming service market, the structure of
the local high-speed intemet market, and the demographics of its franchise area.
(9) D_INT: a0 + alMSO + GZSUB + a3DEN + a4PRE_HF C + assTA + agDBS

+ a7INT + orgBAA + orgHIS + amEMP + anINC + on;OWN + omREG + e9.

Results for the binary logistic regression analysis for D_INT are described in Table
5-13. Overall, the model correctly categorized 81.2% of the cases. The percent in the
modal category for D_INT was 50.2%. The model correctly predicted 31.0% more of the
cases compared to the mode value frequency.SUB, INT and PRE_HF C were statistically
signiﬁcant at the level of .01 or higher. The odds ratio for SUB is 2.58. It means that for
every 1,000 increase a cable system’s basic service subscribers, the odds of the cable
system offering high-speed intemet access service increases by a factor of 2.58, holding
other independent variables constant. The odds ratio for INT is 0.65. It means that with
the addition of one more high-speed ISP to a cable franchise area, the odds of the cable
system offering high-speed intemet access service decreases by a factor of 0.65, holding
other independent variables constant. The odds ratio for PRE_HFC is 364.10, indicating
that the odds of an HF C system offering high-speed intemet access service is more than
364 times the odds of a traditional coaxial system offering high-speed intemet access

service, holding other independent variables constant.

69

Cable System Diversiﬁcation into Telephone Service

Equation (10) assumes that D_TEL, a system’s diversiﬁcation into telephony, is
determined by cable system characteristics, the predicted values for HF C, the structure of
a system’s video programming service market, the structure of the local telephone market,
and the demographics of its franchise area.
(10) D_TEL = a0 + ouMSO + azsUB + a3DEN + a4PRE_HF C + GSSTA + ousDBS

+ a7TEL + orgBAA + a9HIS + amEMP + anINC + (anWN + anREG + em.

Results for the binary logistic regression analysis for D_TEL are presented in Table
5-14. Overall, the model correctly categorized 91.8% of the cases. The percent in the
modal category for D_TEL was 91.7%. The model correctly predicted only 0.01% more
of the cases compared to the mode value frequency. PRE_HFC was the only independent
variable with statistical signiﬁcance for equation (10) at the level of .01. The odds ratio
for PRE_HF C is 9.08, indicating that the odds of an HF C system offering telephone
service is more than 9 times the odds of a traditional coaxial system offering telephone

service, holding other independent variables constant.

70

Cable System Total Degree of Diversiﬁcation

Equation (11) assumes that cable system total degree of diversiﬁcation, D_TOTAL,
is determined by cable system characteristics, the predicted values for HF C, the structure
of a system’s video programming service market, the structure of the local telephone
service market, the structure of the local high-speed lntemet market, and the
demographics of its ﬁanchise area. In equation (11), LNSUB, the natural log of SUB, and
LNINC, the natural log of INC, were used instead of SUB and INC.

(1 l) D_TOTAL = oro + alMSO + azsUB + (13DEN + a4PRE_HFC + a5STA + orgDBS +
a7INT+ orgTEL + orgBAA+ amHlS + anEMP + anlNC + (113OWN
+ (114REG + er I-

The correlation matrix table, Table 5-4, shows a high positive, and statistically
signiﬁcant, correlation coefﬁcient for INT and TEL (R= 0.728). Also, in a regression
collinearity diagnosis, INT and TEL show low tolerance values, 0.346 and 0.266,
respectively. Therefore, TEL was eliminated from Equation (11) to control for
multicollinearity. The revised model is presented in Equation (12):

(12) D_TOTAL = a0 + alMSO + azLNSUB + a3DEN + a4PRE_HFC + assTA + a6DBS
+ a7lNT+ agBAA+ agHIS + aloEMP + anLNINC + a120WN + a13REG + e12.

Results for the multiple linear regression analysis for D_TOTAL are presented in
Table 5-15. Overall, the model explained about 60.0 percent of the variation in the
dependent variable D_TOTAL. MSO, LNSUB, PRE_HFC, STA and INT were
statistically signiﬁcant at the .01 level or higher. Regression results show that MSO
systems on average scored .082 higher than independent systems on D_TOTAL, holding

other independent variables constant. SUB had the largest positive effect on D_TOTAL.

71

For every one percent increase in a cable system’s basic service subscribers, D_TOTAL
increases by 0.62, holding other independent variables constant. HF C systems on average
scored .16 higher than traditional coaxial systems on D_TOTAL, holding other
independent variables constant.

The coefﬁcient for STA was positive and signiﬁcant. For each additional broadcast
television station receivable OTA in a cable franchise area, D_TOTAL increases by .13.
To the contrary, INT, another market structure variable, had a statistically signiﬁcant
negative relationship to D_TOTAL. For each additional high-speed ISP in a cable
franchise area, D_TOTAL decreases by .12. This is consistent with the results for the

regression for D_INT using equation (9).

72

Summary

Table 5-16 shows the estimated change in the probability of cable system
diversiﬁcation following a one unit change in the independent variables that were
statistically signiﬁcant at the 0.01 level or higher for the model that allows for both direct
and indirect effects. For a statistically representative cable system in the sample, the
probability of offering PPV service increases from 0.456 to 0.955 if the ownership
changes from independent to MSO, holding other independent variables constant. The
probability of a statistically representative cable system diversifying into PPV increases
from 0.456 to 0.762 if the number of its basic cable service subscribers increases by
1,000, holding other independent variables constant. The probability of a statistically
representative cable system diversifying into PPV increases from 0.456 to 0.501 if one
more broadcast television station receivable OTA is added to its cable franchise area,
holding other independent variables constant.

The probability of a statistically representative cable system diversifying into high-
speed intemet access service increases from 0.502 to 0.722 if the number of its basic
cable service subscribers increases by 1,000, holding other independent variables
constant. The probability of a statistically representative cable system diversifying into
high-speed intemet access service decreases from 0.502 to 0.396 with the addition of one
more high-speed ISP to its cable franchise area, holding other independent variables
constant. For a statistically representative cable system in this sample, the probability of
offering high-speed intemet access service increases from 0.502 to 0.997 if the upgrade
status changes from traditional coaxial cable plant to HF C plant, holding other

independent variables constant.

73

For a statistically representative cable system in this sample, the probability of
offering telephone service increases from 0.083 to 0.451 if the upgrade status changes
from traditional coaxial cable plant to HF C plant, holding other independent variables

constant.

74

5.5 Comparison of the Direct Effects-Only Model and the Model with Both Direct and
Indirect Effects

Table 5-9 and Table-16 show the regression results for the direct effects-only model
and the model with both direct and indirect effects, respectively. By comparing these two
tables, we can see that the regression results for the model with both direct and indirect
effects show that HF C inﬂuenced cable systems’ decisions to diversify into high-speed
intemet access and telephone services, and the odds ratios for the statistically signiﬁcant
independent variables, including MSO, SUB, STA, and INT, did not change much from
their values in the direct effects-only model. This suggests that although the model with
both direct and indirect effects allows for the possibility of indirect effects, it did not
provide evidence that the indirect effects were very important in determining cable
system diversiﬁcation into the three services examined in this study. Also, regression
results show that the factors that inﬂuenced cable systems’ decisions to upgrade to HF C,
including SUB, STA, and INT, also inﬂuenced cable systems’ decisions to diversify into
the three services considered in this study.

In the direct effects-only model, STA had a statistically signiﬁcant and positive
inﬂuence on D_INT in the equation for cable system diversiﬁcation into high-speed
intemet access service. To the contrary, in the model with direct and indirect effects, STA
was not a statistically signiﬁcant independent variable in the equation for cable system
diversiﬁcation into high-speed intemet access service, whereas HF C had a positive
inﬂuence on cable system diversiﬁcation into high-speed intemet access service. This
suggests a plausible explanation for this ﬁnding is that once cable systems have upgraded
to HF C, they would diversify into high-speed intemet access service regardless of the

competitiveness of their primary business of video programming. Another plausible

75

explanation is that cable systems upgraded to HF C in part so they can offer high-speed

intemet access service.

76

CHAPTER 6
DISCUSSIONS

6.1 Discussion of the Results

Technological advances and regulatory initiatives have made competition in the
multichannel video programming service market possible to a degree that was once
assumed to be economically infeasible. The cable industry now faces more competition
than ever before. Cable’s primary business of video programming service is confronted
with increased competition from satellite companies and telephone companies. The old
system of cable monopoly is gradually being replaced with a competitive marketplace.

This study is an application of the 1/0 model to cable system diversiﬁcation. The
model used in this study focuses on interaction between cable systems and the other
service providers when cable systems have the opportunity to diversify into PPV, high-
speed intemet access, and telephone services. It examines how cable systems diversify in
response to the structure of both their primary markets and their target markets, and thus
explores the sensitivity of cable systems’ conduct to market structure at the cable
franchise level. It thereby contributes to our understanding of the forces behind the cable
industry’s diversiﬁcation. Results for the set of regressions corresponding to the direct
effects- only model are discussed in the following paragraphs.

SUB had a positive effect on HF C, D_PPV and D_INT. These results suggest that
the number of basic cable service subscribers is one of the factors that inﬂuenced cable
systems’ decisions to upgrade to HFC and to diversify into PPV and high—speed intemet
access services. Potential revenues that can be generated from providing PPV and high-
speed intemet access services to existing customers might have increased the motivation

for cable systems to upgrade and diversity into these services.

77

STA had a positive effect on HF C, D_PPV and D_INT. My regression analysis
indicates that cable systems operating in franchise areas with more broadcast television
signals are more likely to upgrade to HF C and to diversify into PPV and high-speed
intemet access services. More broadcast television receivable OTA implies more local
channels for cable systems to carry, which increases the value of additional bandwidth.
Also, more broadcast television signals means more competition for cable television.
OTA television is a substitute for cable television for some customers. Confronted with
more competition from OTA television, cable systems might be responding by upgrading
and diversifying in order to create new revenue sources.

My regression analysis indicates that MSO systems are more likely to diversify into
PPV service than independent systems. Cable operators that provide PPV services obtain
PPV programming from PPV networks and major suppliers of PPV programming to
cable systems like Viewer’s Choice and Request TV, and split the viewing fees with
them based on a negotiated percentage. This suggests a plausible explanation for
this ﬁnding is that MSOs are more likely to be able to negotiate a higher percentage for
themselves and achieve higher buy-rates because they have a larger subscriber base and
thus more bargaining power compared to independent systems.

In this study, two competing hypotheses, that the effect of competitiveness of the
target markets on cable system diversiﬁcation into these markets is either positive or
negative, were considered. INT, the number of high-speed ISPs in a cable franchise area,
which measures the competitiveness of the high-speed intemet market, had a negative
coefﬁcient in the equation for cable systems’ upgrade to HF C and cable system

diversiﬁcation into high-speed intemet. It suggests that cable systems are less likely to

78

upgrade to HF C and to diversify into high-speed intemet access service when there are
more competitors in the high-speed intemet market. More high-speed ISPs in a cable
franchise area means more competition for a cable system. A plausible explanation for
these results is that a lower potential market share, especially where other high-speed
ISPs have already acquired a substantial market share, and thereby a lower rate of return
on investment might discourage a cable system operating in a local market with more
high-speed lSPs from investing in HF C and diversifying into highs-speed intemet access
service. To the contrary, the market structure variable that measures the competitiveness
of a cable system’s local telephone market, did not have a statistically signiﬁcant
coefﬁcient in the equation for cable systems diversiﬁcation into telephony. It suggests
that a cable system’s decision to diversify into telephony was not inﬂuenced by the
competitiveness of its local telephone market. None of the factors included in the
equation had a signiﬁcant coefﬁcient in the direct effects-only model. In the set of
regressions corresponding to the model with both direct and indirect effects, HF C was the
only independent variable that had a statistically signiﬁcant coefﬁcient in the equation for
cable systems diversiﬁcation into telephony; also, HF C systems are more likely to
diversify into telephone service than traditional coaxial systems. The primary advantage
of digital cable telephone service over traditional circuit-switched service for cable
operators is its cost advantage. It costs between 17 percent and 25 percent less to serve a
digital cable telephone service subscriber than a traditional circuit-switched cable
telephone service subscriber (Cable telephony, 2006). This cost advantage is a plausible

explanation for why HF C cable systems are more likely to provide telephone service.

79

My analysis indicates that cable systems are more likely to upgrade to HF C in cable
franchise areas with lower home ownership rates. According to the US. Census Bureau’s
Housing Vacancy Survey (US, Census Bureau, 2005), as of 2004 the home ownership
rate was 53.8 percent for central city households, 75.4 percent for suburban households,
and 76.4 percent for rural households. Home ownership rate is the lowest in central cities
where rental apartment living is more often chosen. It costs cable systems less per
household to upgrade to HFC in central cities because household unit density is higher
there. Having compared the expense of cable plant upgrades to the number of potential
subscribers in central cities and in urban areas with higher rates of home ownership, cable
systems operating in central cities may have expected a higher return on investment than

those in urban areas.

80

6.2 Suggestions for Future Research

I suggest future research explore the management implications of diversiﬁcation for
cable systems, especially the impacts of pursuing diversiﬁcation on cable companies’
organization structures. Diversiﬁed cable systems have to adapt to the requirements of
providing multiple services. Studies in this area can be useful for cable managers looking
for guidance in managing their diversiﬁed organizations.

Also, I suggest that future research assess the competitive effects of telephone
companies’ entry into the traditional video programming market on cable systems’
product offerings and pricing policies. There is little discussion about the effects of
telephone companies’ entry into the video programming market on cable systems’
product offerings and pricing policies in the literature perhaps because they are just
starting to do so. With more telephone companies providing video programming services,
it would be meaningful to assess whether the F CC’s objective of promoting competition
in MVPD services has been achieved. I suggest that future research investigate how
telephone’s entry into video inﬂuences the cable industry’s product offerings and pricing
policies in the multichannel video programming market.

Last, I suggest that future research look at the new services that cable systems have
diversiﬁed into in addition to those examined here. Cable systems’ investments in
upgrading their systems have enabled them to provide more new services such as
streaming media, interactive television, and video-on-demand. It would be meaningful to
study cable system diversiﬁcation into these new services once cable systems provide

them on a large scale.

81

6.3 Limitations of the Study

Four limitations of this study should be noted. First, DBS penetration data at the state
level instead of at the cable franchise area level were used for the cable systems in the
sample. The FCC only provided DBS penetration data at the state level, and no data
sources on DBS penetration by cable franchises areas were available that could be
obtained with the budget available for this study. Ideally my analysis could be replicated
with franchise level data on DBS penetration in the future.

The second limitation is the study’s imprecision in data matching: cable system
characteristics data were reported at the cable franchise area level, but market structure
and market demographic data were reported at the zip code level. For cable franchise
areas associated with multiple zip codes, simple averaging instead of a more reﬁned
weighting was used to aggregate data. As mentioned in Chapter 4, some cable franchise
areas are associated with multiple zip codes. For those cases, a weighted sum with
weighting by population in each zip code should be used for the cable franchise area, and
this weighting would be applied to all variables reported on a zip code level. However,
information on the distribution of cable system subscribers in each zip code was not
available for this study.

The third limitation is that one year’s Observations were used to draw inferences
about motives for events that occurred sequentially over time. Two-stage least squares
techniques were employed to address the issue of endogeneity that would arise if cable
systems upgraded to HF C for reasons other than diversifying into the three services
considered here but subsequent possession of HF C plant inﬂuenced the decisions

regarding diversiﬁcation. In the two-stage least squares regressions, the other

82

independent factors that might inﬂuence cable system diversiﬁcation were used for the
equation for HF C, but they may not be adequate for identifying the factors that
inﬂuenced cable systems’ HFC upgrade decisions because it is impossible to
determine when the upgrade to HFC was complete using one year’s data. Systems
that upgraded recently and those that did so considerably earlier are of necessity
treated the same, but the expectations for revenues from new services may have
been different for those that upgraded most recently and those that upgraded much
earlier.

The fourth limitation is that all the cable operators with two or more systems are
treated the same by using a simple binary MSO dummy variable. It is possible that the
largest MS05 are different from the other MSOS when they make their diversiﬁcation
decisions. This issue could have been addressed if the total number of basic cable
subscribers for the largest MSOs had been employed as an independent variable in this

study.

83

Table 4-1 Dependent Variables and Independent Variables

Category Variable

Dependent D_PPV
Variables

D_INT

D_TEL

D_TOTAL

Independent
Variables

Operationalization

Binary variable =1 if a
system offered PPV
service, 0 otherwise
Binary variable =1 if a
system offered high-
speed lntemet access
service, 0 otherwise
Binary variable =1 if a
system offered voice
telephone service, 0
otherwise

The sum of D_PPV,

D INT and D TEL

Expected Source

Sign
n.a.
n.a.

n.a.

n.a.

[1]

[1]

l1]

[1]

 

System MSO
characteristics

SUB
LNSUB

DEN

HF C

PRE_HFC

Dummy variable = 1 if
a system is MSO-
owned, 0 otherwise
The number of a cable
system’s basic service
subscribers (in
thousands)

Natural log of SUB
The number of
households per mile of
cable plant

Dummy variable = 1 if
a system has a capacity
of 750 MHz or higher,
and 0 if a system has a

capacity of 450 MHz or

lower; systems with
MHz capacity between
450 and 750 were
coded based on
telephone interview
results.

Predicted values for
HF C

[1]

[11

[1]

[1]

 

84

Table 4-1 (cont’d)

 

Market STA
structure
DBS
INT
TEL
Market BAA

Demographics

HIS

EMP

INC

LNIN C
OWN

The number of over-
the-air broadcast
television receivable in
a cable franchise area
Penetration of DBS at
the state level

The number of high-
speed lntemet access
service providers in a
zip code area that
associate with the cable
franchise area

The number of CLECs
a zip code area in a zip
code area that associate
with the cable franchise
area

Percentage of Black or
African American
population in a zip
code area that associate
with the cable franchise
area

Percentage of Hispanic
population in a zip
code area that associate
with the cable ﬁanchise
area

Employment rate in a
zip code area that
associate with the cable
franchise area

Median household
income in a a zip code
area that associate with
the cable franchise area
(in thousands)

Natural log of INC
Home ownership rate
in a zip code area that
associate with the cable
franchise area

+/-

+/-

n.a.

n.a.

n.a.

n.a.

n.a.
n.a.

[1]

[2]
l3]

[4]

[5]

[5]

[5]

[5]

 

85

Table 4=1 (cont’d)
REG Dummy variable = 0 if n.a. [5]
a cable system is in the
midwest, 1 if in the
northeast, 2 if in the
south, and 3 if in the
west

[1] The Television & Cable Factbook (2005). Cable vol. 1&2

[2] NCTA (February 2005). The video market is fully competitive: Almost 27 million
consumers now subscribe to cable’s competitors.

[3] FCC Form 477 (2005) Number of High-Speed Service Providers by Zip Codes

[4] FCC Form 477 (2005) Number of Competitive Local Exchange Carriers by Zip Codes
[5] US. Census (2004), American Community Survey.

86

Table 5-1 Relative Frequency of Cable System Product Diversiﬁcation and HF C
Upgrades

Frequency Percent

 

PPV 149 45.6
lntemet 164 50.2
Telephone 27 8.3

HF C 100 3 1 .0

 

87

Table 5-2 Relative Frequency of Cable System Total Degree of Diversiﬁcation
Score Frequency Percent Cumulative

 

Percent
0 132 40.4 40.4
1 73 22.3 62.7
2 99 30.3 93.0
3 23 7.0 100.0
Total 327 100.0

 

88

Table 5-3 Descriptive Statistics for the Sample

 

 

N Minimum Maximum Mean Std.
Deviation

M80 327 0 1 0.92 .27
SUB 327 0.004 460 9.52 37.39
DEN 297 2 744 55.44 60.35
HF C 310 0 1 0.31 .46
STA 320 l 23 7.80 3.47
DBS 327 9.00 39.00 23.00 5.00
INT 315 0 16 3.52 2.72
TEL 315 0 19 3.50 3.60
BAA 316 0 83.20 8.63 15.89
HIS 316 0 84.60 5.34 10.79
EMP 314 12.90 86.50 59.42 9.53
INC 316 13.46 112.76 33.85 12.65
OWN 315 1.00 93.00 61.00 .13
REG(O) 28 n.a n.a. n.a. n.a.
REG(l) 1 l9 n.a n.a. n.a. n.a.
REG(2) 122 n.a n.a. n.a. n.a.
REG(3) 58 n.a. n.a. n.a. n.a.

 

89

 

.320 mL 09.8530: 381x 3.. En sine—om

 

 

Ulww< UIHZA. Ulﬁmr UIHOHEP. ZmO mg UmZ wand mu...» me ~24. Hm? >> $5 m5 ~20 032
Olav/x _ amt... .89.... @001. .Smi. bit. .30.}. 63...... LEI. 33mg... moon... #3.... .055 .03 .37.... hams... 33a...
UIHZA. _ has... .39.... .000 .Nao: Lows... 08...... hour... 35m... .39.... bout. .o—u bum .31.... hues... Loo...
Clam? _ Mama... boo bows... LE... .Nmoi. .29.... 339.... 98...... Moon... .05 Loo... bum boot. 303

UIHOHB _ .39.... bums... NEE... .omo: 33...... Jami. .wmqi. .52.... .oma Low Mama... bows... L5...
750 _ bum bum Low... bu.» .o _ N b3 .93 38m .2: .o E .owc 308
mg _ L8... .8%... but... 3N8: 98...... but... .39.... .T: boo .59.... :08
UmZ _ 96...... .39.... scam .59.... has... no; .30... ._ Z . 5m dow—
wmun _ .39.... 35%.... bums... bums... baa b: .33.... ._mm..._.. swnoi.
m5? _ 320...... .33.... .Amoa... bum .Nmoi. has... boos... .03
me _ 33%.... swag...” boo -bmo JUNE... snows... no;
~24. _ gums... bum No.2... .89.... 03...... 30.3
Hmb _ .wai. hoax... .39.... .uoma... some
>> _ .o: JNAO: 35%.... JEN
Em _ .OAN .Omm -. _ mu ...
ma _ .39.... .No—s...
~20 — boa:
OQZ _

 

... Coin—mac: mm amines" on :5 98 :29 APE—08
...... Dona—mac: mm mmmammnma 2 :5 Pc— _o<n_ APE—08

0
9

Table 5-5 Cable System Diversiﬁcation into PPV (Direct Effects-Only Model)

Nagelkerke’s R2 .707
Signiﬁcance <.0001

 

 

 

 

N=283
Variables B S.E. Wald df Sig. Odds Ratio 95% CI. for OR
(OR)
Lower Upper
MSO(1) 3.335 1.275 6.842 1 .009 28.076 2.307 341.668
SUB 1.365 .250 29.845 1 .000 3.918 2.400 6.394
DEN -.005 .003 3.105 1 .178 1.094 .989 1.001
STA .201 .074 7.335 1 .007 1.223 1.057 1.415
DBS -1.342 4.947 .074 1 .786 .261 .000 4244.435
BAA .024 .015 2.461 1 .117 1.024 .994 1.055
HIS .000 .019 .001 1 .980 1.000 .963 1.039
EMP .037 .025 2.183 1 .140 1.038 .988 1.091
INC .014 .031 .216 l .642 1.014 .955 1.077
OWN -.750 1.623 .214 l .644 .472 .020 11.363
REG 5.463 3 .141
REG(l) -.566 1.201 .222 1 .637 .568 .054 5.974
REG(2) -l.973 1.321 2.231 1 .135 .139 .010 1.852
REG(3) -.612 1.336 .210 1 .647 .542 .040 7.440
Constant 3.335 1.275 6.842 1 .009 28.076 2.307 341.668
Comparison of Actual and Predicted Results
Predicted Percent
Correct
0 1
Actual 0 136 14 90.7
1 30 103 77.4
Percent Correctly Predicted 84.5
Percent in Modal Category 54.4

 

91

 

Table 5-6 Cable System Diversiﬁcation into High-Speed lntemet Access Service (Direct
Effects-Only Model)
Nagelkerke’s .584

 

 

 

 

 

 

R2
Signiﬁcance <.0001
N=262
Variables B S.E. Wald df Sig. Odds Ratio 95% Cl. for OR
(OR)
Lower Upper
MSO(1) .073 .553 .018 1 .895 1.076 .364 3.183
SUB 1.162 .219 28.225 1 .000 3.196 2.082 4.906
DEN .003 .007 .247 1 .619 1.003 .991 1.016
STA .159 .066 5.748 1 .017 1.172 1.029 1.335
DBS 6.318 4.423 2.041 1 .153 5.685 .095 32.792
INT -.489 .125 15.400 1 .000 .613 .481 .783
BAA -.005 .013 .160 1 .689 .995 .971 1.020
H18 .003 .018 .026 l .871 1.003 .969 1.038
EMP -.002 .020 .012 l .913 .998 .960 1.038
INC .028 .028 1.014 1 .314 1.028 .974 1.085
OWN -.605 1.451 .174 1 .677 .546 .032 9.385
REG 3.579 3 .311
REG(l) -.648 .997 .422 1 .516 .523 .074 3.696
REG(2) -1.404 1.075 1.706 1 .191 .246 .030 2.019
REG(3) -1.318 1.168 1.273 1 .259 .268 .027 2.641
Constant -2.373 1.782 1.773 1 .183 .093
Comparison of Actual and Predicted Results
Predicted Percent
Correct
0 1
Actual 0 112 20 84.8
1 39 111 74.0
Percent Correctly Predicted 79.1
Percent in Modal Category 50.2

 

92

Table 5-7 Cable System Diversiﬁcation into Telephone Service (Direct Effects-Only

 

 

 

 

 

 

 

Model)
Nagelkerke’s .268
R2
Signiﬁcance <.0001
N=282
Variables B S.E. Wald df Sig. Odds Ratio 95% CI. for OR
(OR)
Lower Upper
MSO(1) -.873 .843 1.072 1 .301 .418 .080 2.181
SUB .008 .006 1.981 1 .159 1.008 .997 1.020
DEN .001 .004 .050 l .822 1.001 .994 1.008
STA .055 .077 .508 1 .476 1.057 .908 1.230
DBS -8.034 7.533 1.138 1 .286 .000 .000 836.588
TEL .112 .085 1.744 1 .187 1.118 .947 1.320
BAA -.023 .024 .899 l .343 .977 .932 1.025
HIS .005 .023 .042 1 .839 1.005 .961 1.050
EMP -.018 .032 .338 l .561 .982 .923 1.044
INC .042 .024 3.132 1 .177 1.043 .995 1.094
OWN -2.519 1.924 1.714 1 .191 .081 .002 3.499
REG .720 3 .868
REG(l) .258 .986 .069 l .793 1.295 .188 8.941
REG(2) .813 1.140 .508 1 .476 2.254 .241 21.056
REG(3) .326 1.253 .068 1 .795 1.385 .119 16.137
Constant -.301 2.518 .014 1 .905 .740
Comparison of Actual and Predicted Results
Predicted Percent
Correct
0 1
Actual 256 3 98.8
20 3 13.0
Percent Correctly Predicted 79.1
Percent in Modal Category 50.2
Percent Correctly Predicted 91.8
Percent in Modal Category 91.7

 

93

Table 5-8 Cable System Total Degree of Diversiﬁcation (Direct Effects-Only Model)
R2 0.619

 

 

 

Signiﬁcance <.0001
Standardized beta t Probability
coefﬁcient
(Constant) -1.978 .049
M80 .091 2.378 .018
LNSUB .732 13.055 .000
DEN -.026 -.632 .528
STA .171 3.511 .001
DBS .030 .607 .545
INT -.106 —2.088 .038
BAA -.040 -.869 .386
HIS -.026 -.604 .547
EMP .086 1.759 .180
LNINC .039 .691 .490
OWN -.023 -.529 .597
NORTHEAST -.045 -.959 .338
SOUTH -.070 - l .337 .182
WEST -.047 -.965 .335

 

94

Table 5-9 Change in Probability of Cable System Diversiﬁcation Following One Unit

Increase in the Statistically Signiﬁcant Independent Variables (Direct Effects-Only
Model)

 

 

 

 

Original Aﬁer One Unit Change
Probability“ MSO SUB" STA INT
D_PPV 45.6 95.9 76.7 50.6 --
D_INT 50.2 -- 76.3 54.2 38.2

 

"' Value for statistically average system
** One unit is one thousand basic service subscribers

95

Table 5-10 The Effects of System Characteristics, Competition from Broadcast
Television, and Market Demographics on Cable Systems’ Upgrade to HF C
Nagelkerke’s .50

 

 

 

 

 

R2
Signiﬁcance <.0001
N=263
Variables B S.E. Wald df Sig. Odds Ratio 95% CI. for OR
(OR)
Lower Upper
MSO( 1) .600 .694 .748 1 .387 1.823 .467 7.1 10
SUB .165 .040 17.303 1 .000 1.179 1.091 1.274
DEN .005 .004 1.583 1 .208 1.005 .997 1.013
STA .157 .061 6.600 1 .010 1.170 1.038 1.320
BAA -.002 .013 .017 1 .895 .998 .973 1.025
HIS -.024 .024 1.016 1 .313 .977 .933 1.023
EMP .025 .023 1.212 1 .271 1.025 .981 1.072
INC -.012 .020 .356 l .551 .988 .951 1.027
OWN -4.045 1.403 8.316 1 .004 .018 .001 .274
REG .959 3 .811
REG(l) -.691 .727 .905 1 .341 .501 .121 2.081
REG(2) -.585 .768 .580 l .446 .557 .124 2.511
REG(3) -.726 .850 .731 1 .393 .484 .092 2.557
Constant -l .443 1.602 .811 1 .368 .236
Comparison of Actual and Predicted Results
Predicted Percent
Correct

0 1
Actual 0 171 9 95.0
1 36 57 61.3
Percent Correctly Predicted 83.5
Percent in Modal Category 69.0

96

Table 5-11 The Effects of the System Characteristics, Market Structure, and Market
Demographic Variables on Cable Systems’ Upgrade to HF C

 

 

Nagelkerke’s .53
R2
Signiﬁcance <.0001
N=263
Variables B S.E. Wald df Sig. Odds Ratio 95% CI. for OR
(OR)
Lower Upper
MSO( 1) .510 .690 .546 l .460 1.665 .430 6.445
SUB .183 .046 16.203 1 .000 1.201 1.099 1.313
DEN .006 .005 1.405 1 .236 1.006 .996 1.016
STA .158 .064 6.035 1 .014 1.171 1.032 1.328
DBS 5.268 4.332 1.479 1 .224 193.957 .040 943510.331
INT -.275 .115 5.744 1 .017 .760 .607 .951
TEL .265 .093 8.117 1 .114 1.304 .986 1.565
BAA -.010 .015 .519 l .471 .990 .962 1.018
HIS -.027 .026 1.090 1 .296 .973 .925 1.024
EMP .039 .025 2.540 1 .11 1 1.040 .991 1.091
INC -.025 .022 1.327 1 .249 .975 .935 1.018
OWN -4.437 1.511 8.624 1 .003 .012 .001 .229
REG .575 3 .902
REG(l) -.456 .803 .322 1 .570 .634 .131 3.060
REG(2) -.553 .883 .393 1 .531 .575 .102 3.245
REG(3) -.768 1.027 .559 1 .455 .464 .062 3.475
Constant -2.823 1.819 2.408 1 .121 .059

 

Comparison of Actual and Predicted Results
Predicted Percentage

 

Correct
1
Actual 169 10 94.4
58 62.4
Percent Correctly Predicted 83.5
Percent in Modal Category 69.0

 

97

Table 5-12 Cable System Diversiﬁcation into PPV (Model with Both Direct and Indirect

Effects)

Nagelkerke’s R2 .707

Signiﬁcance <.0001

N=283

Variables B S.E. Wald df Sig. Odds Ratio 95% CI. for OR
(OR)

 

 

Lower Upper

 

 

MSO(l) 3.238 1.293 6.275 1 .012 25.493 2.023 321.208
SUB 1.339 .258 26.875 1 .000 3.816 2.300 6.332
DEN -.006 .004 2.798 1 .114 .994 .987 1.001
STA .184 .087 4.515 1 .034 1.202 1.014 1.425
DBS -1.608 4.993 .104 l .747 .200 .000 3565.730
BAA .024 .015 2.451 1 .117 1.024 .994 1.055
HIS .003 .021 .028 1 .868 1.003 .963 1.045
EMP .034 .026 1.665 1 .197 1.035 .982 1.090
INC .017 .032 .297 1 .586 1.018 .956 1.083
OWN -.274 2.054 .018 1 .894 .760 .014 42.619
REG 5.116 3 .164

REG(I) -.4501.283 .123 1 .726 .638 .052 7.888

REG(2) -1.8291.415 1.670 1 .196 .161 .010 2.572

REG(3) -.4671.434 .106 1 .744 .627 .038 10.412

PRE_HFC .812 2.214 .134 1 .714 2.251 .029 172.592

Constant -7.534 2.500 9.084 1 .003 .001

Comparison of Actual and Predicted Results
Predicted Percent
Correct

0 1
Actual 0 137 12 91.9
1 30 103 77.4
Percent Correctly Predicted 85.1
Percent in Modal Category 54.4

98

Table 5-13 Cable Diversiﬁcation into High-Speed lntemet Access Service (Model with

Both Direct and Indirect Effects)

 

 

 

 

 

 

Nagelkerke’s .554
R2
Signiﬁcance <.0001
N=262
Variables B S.E. Wald df Sig. Odds Ratio 95% CI for OR
(OR)
Lower Upper
MSO(I) -.328 .572 .328 1 .567 .720 .235 2.212
SUB .947 .238 15.912 1 .000 2.579 1.619 4.108
DEN .000 .007 .000 1 .991 1.000 .986 1.014
STA .045 .080 .319 1 .572 1.046 .895 1.223
DBS 4.374 4.592 .907 1 .341 79.340 .010 642838.944
INT -.428 .127 11.384 1 .001 .652 .508 .836
BAA -.005 .013 .178 1 .673 .995 .969 1.020
HIS .020 .018 1.214 1 .270 1.021 .984 1.058
EMP -.025 .022 1.298 1 .255 .975 .934 1.018
INC .044 .029 2.294 1 .130 1.045 .987 1.107
OWN 2.890 2.018 2.051 1 .152 18.002 .345 939.881
REG 1.901 3 .593
REG(l) .227 1.040 .048 1 .827 1.255 .164 9.626
REG(2) -.406 1.129 .129 1 .719 .667 .073 6.097
REG(3) -.219 1.224 .032 1 .858 .804 .073 8.856
PRE_HFC 5.897 2.280 6.688 1 .010 364.099 4.170 31793.489
Constant -4.080 1.921 4.510 1 .034 .017
Comparison of Actual and Predicted Results
Predicted Percent
Correct
0 1
Actual 0 115 17 87.1
1 36 114 76.0
Percent Correctly Predicted 81.2
Percent in Modal Category 50.2

 

99

Table 5-14 Cable Diversiﬁcation into Telephone Service (Model with Both Direct and

 

 

 

 

 

 

Indirect Effects)
Nagelkerke’s .292
R2
Signiﬁcance <.0001
N=282
Variables B S.E. Wald df Sig. Odds Ratio 95% CL for OR
(OR)
Lower Upper
MSO(l) -1.220 .887 1.892 1 .169 .295 .052 1.679
SUB .004 .006 .548 1 .459 1.004 .993 1.015
DEN -.001 .004 .036 1 .849 .999 .992 1.007
STA -.004 .085 .002 1 .965 .996 .844 1.176
DBS -7.244 7.145 1.028 1 .311 .001 .000 862.416
TEL .051 .088 .336 l .562 1.053 .885 1.251
BAA -.020 .023 .734 1 .391 .980 .936 1.026
HIS .019 .024 .663 1 .416 1.019 .973 1.068
EMP -.034 .033 1.066 1 .302 .966 .906 1.031
INC .043 .024 3.070 1 .180 1.044 .995 1.095
OWN -l.089 2.088 .272 1 .602 .337 .006 20.160
REG .694 3 .875
REG(l) .271 .954 .081 1 .776 1.311 .202 8.498
REG(2) .782 1.098 .507 1 .476 2.185 .254 18.780
REG(3) .242 1.219 .039 1 .842 1.274 .117 13.894
PRE_HFC 2.206 1.187 3.453 1 .063 9.080 .886 93.028
Constant -.263 2.557 .011 1 .918 .768
Comparison of Actual and Predicted Results
Predicted Percent
Correct
0 1
Actual 256 3 98.8
20 3 13.0
Percent Correctly Predicted 91.8
Percent in Modal Category 91.7

 

100

Table 5-15 Cable System Total Degree of Diversiﬁcation (Model with Both Direct and

 

 

 

Indirect Effects)
R2 0.600
Signiﬁcance <.0001
Standardized beta I Probability
coefﬁcient
(Constant) -2.329 .021
MSO .082 2.052 .041
LNSUB .615 7.909 .000
DEN -.044 -1 .026 .306
PRE_HFC .161 1.810 .071
STA .126 2.323 .021
DBS .031 .608 .544
INT -.116 -2.216 .028
BAA -.029 -.627 .531
HIS -.005 -.1 14 .909
EMP .023 .458 .647
LNINC .076 1.354 .177
OWN .005 .109 .913
NORTHEAST -.029 -.604 .547
SOUTH -.076 -1.432 . 153

WEST -.050 -.995 .320

 

101

Table 5-16 Change in Probability of Cable System Diversiﬁcation Following One Unit
Increase in the Statistically Signiﬁcant Independent Variables (Model with Both Direct
and Indirect Effects)

 

 

 

 

Original After One Unit Change
Probability“ MSO SUB" STA INT PRE_HF C
D_PPV 45.6 95.5 76.2 50.1 -- --
D_INT 50.2 -- 72.2 -- 39.6 99.7
D;TEL 8.3 -- -- -- -- 45. 1

 

* Value for statistically average system
** One unit is one thousand basic service subscribers

102

Figure 1-1 Cable Industry Inﬁ'astructure Expenditure: 1996-2005 (in $ billions)

18

 

16]
14 -
12-
10-

 

1997

  
  

1998

 
 

:'-: ‘.G' '9; 's: ' "
.mﬁuﬁgff ,
.4 .-

..- ~’-'
his
49/.“

Q's-unit..-
e I . ..
‘ l 2' t 5" I a
a .. "f;.":°‘;l:=

9'4 3'." '
in

“-.‘iti'ﬁir‘

. u
‘ :3 «1'4"

1999

 

 

 

 

2000 2001 2002 2003

Source: NCTA (2006)

103

2004

 
     
     
 
     
 
  

2005

 

Figure 1-2 Cable Industry Revenue from Subscription-Based Video Programming
Service: 1997-2005 (in $ millions)
$40,000

 

537,537
$36,335
$34,891
$35.00“ ' $32,916

330.919
$29,377
530-000 ‘ $28,124 _ ,

$26,914

  
   

$24,960
$25,000 - --—-

$20,000 <

 

$15,000 -

$10,000 -

 

$5,000 -

 

1997 1998 1999 2000 2001 2002 2003 2004 2005

 

 

           

 

 

 

Source: FCC (2002, 2006)

104

Figure 1-3 Cable High-speed lntemet Customers: 2000-2005 (in millions)

 

25

23.0

 

 

 

 

 

 

 

 

Source: NCTA (2005)

105

Figure 1-4 Cable Phone Customers: 2000-2005 (in millions)

 

 

   

 

 

 

   

2003 2004

2001

Source: NCTA (2005)

106

Figure 4-1: A Framework for Analyzing Cable System Diversiﬁcation

 

System Characteristics
- MSO

- SUB
- DEN

 

 

Market Structure
- STA
- DBS
- INT
- TEL

 

 

Market Demoggphics
- BAA

- HIS
- EMP
- INC
- OWN
- REG

 

 

 

 

107

 

Cable S stem
mm

- D_PPV
- D_INT
- D_TEL
- D_TOTAL

 

 

References

Abell, D. F ., & Hammond, J. S. (1979). Strategic market planning: Problems and
analytical approaches. Englewood Cliffs, NJ: Prentice-Hall.

Ahn, H. (2000). A structural analysis of overbuild competition and convergence in the
multichannel television industry. Doctoral dissertation, Michigan State University.

Albarran, A. B. (1996). Media Economics. Ames, Iowa: State University Press.

Albarran, A. B., & Dimmick, J. (1996). Concentration and economies of multiforrnity in
the communication industries. Journal of Media Economics, 9(4), 41-50.

Albarran, A. B., & Moellinger, T. (2002). The top six communication industry ﬁrms:
Structure, operation, and strategy. In R. G. Picard (Eds), Media ﬁrms: Structures,
operations, and performance (pp. 103-122). Mahwah, N.J.: Lawrence Erlbaum
Associates, Inc.

Albarran, A. B., & Porco, J. F. (1990). Measuring and analyzing diversiﬁcation of
corporations involved in pay cable. Journal of Media Economics, 3(1), 3-14.

Amihud, Y., & Lev, B. (1981). Risk reduction as a managerial motive for conglomerate
mergers. Bell Journal of Economics, 12, 605-617.

Amit, R., & Livnat, J. (1988). Diversiﬁcation strategies, business cycles and economic
performance. Strategic Management Journal, 9(2), 99-110.

Ansoff, H. I. (1958). A model for diversiﬁcation. Management Science, 4, 392-414.

Ansoff, H. I. (1965). Corporate Strategy. New York, NY: McGraw-Hill.

Bain, J. S. (1968). Industrial Organization (2“). New York, NY: Wiley.

Bates, B. J ., & Chambers, T. (2004). The economics of the cable industry. In A.
Alexander, J. Owers, R. Carveth, C.A. Holliﬁeld, & A. N. Greco (Eds), Media
economics: Theory and practice (3rd ed.) (pp. 173-192). Mahwah, NJ: Lawrence
Erlbaum Association.

Bengtsson, L. (2000). Corporate strategy in a small open economy: Reducing product
diversiﬁcation while increasing international diversiﬁcation. European

Management Journal, 18(4), 444.

Berry, C. H. (1975). Corporate growth and diversiﬁcation. Princeton, NJ: Princeton
University,

108

Booz, Allen, & Hamilton (1985.) Diversiﬁcation: a survey of European chief executives.
New York, NY: Booz, Allen and Hamilton, Inc.

Brost, M. A., & Kleiner, B. H. (1995). New developments in corporate diversiﬁcation
strategies. Management Research News, 18(3-5), 24-33.

Bustema, J. C. (1988). Concentration and the industrial organization model. In R. G.
Picard, et al (Eds). Press Concentration and Monopoly: New Perspectives on
Newspaper Ownership and Operation (pp.35-53). Press Concentration and
monopoly. Norwood, NJ: Ablex Publishing Corporation.

Cable telephony service revenues to hit $10 billion by 2009 (2006, February 8). In-Stat.
Retrieved March 10, 2006, from:
http://www.instat.com/press.asp?ID=l580&sku=ll\10502142MB

Chan-Olmsted, S. M., & Chang, B. H. (2003). Diversiﬁcation strategy of global media
conglomerates: Examining its patterns and determinants. Journal of Media
Economics, 16(4), 213-233.

Chatterjee, S., & Blocher, J. (1992). The continuous measurement of ﬁrm diversiﬁcation:
Is it robust? Academy of Management Journal, 35(3), 874-888.

Chatterjee, S., & Wemerfelf, B. (1991). The link between resources and type of
diversiﬁcation: Theory and evidence. Strategic Management Journal, 12(1), 33-
48.

Chiddix, J. A., Boyer, Darcie, Werner, Vecchi (2000). The evolution of the US
telecommunications infrastructure over the next Decade: Hybrid-ﬁber-coax
technology. In the proceedings of the 2000 IEEE Conference.

Chipty, T. (2001). Vertical integration, market foreclosure, and consumer welfare in the
cable television industry. American Economic Review, 91(3), 428-453.

Collis, D. J., & Montgomery, C. A. (1995). Competing on resources: Strategy in the
19905. Harvard Business Review, 73(4), 1 18-129.

Delios, A., & Beamish, P.W. (1999). Geographic scope, product diversiﬁcation, and the
corporate performance of Japanese ﬁrms. Strategic Management Journal, 20(8),
711-727.

Denis, D. J ., & Denis, D. K., & Sarin, A. (1999). Agency theory and the inﬂuence of
equity ownership structure on corporate diversiﬁcation strategies. Strategic
Management Journal, 20(11), 1071.

Dertouzos, J. N., & Wildman, S. S. (1990). Competitive eﬂects of broadcast signals on
cable. In R. G. N011 and M. E. Price (Eds.), A Communications Cornucopia: Market

109

Foundation Essays on Information (pp. 499-517). Washington DC: Brookings
Institution Press.

Didrichsen, J. (1972). The development of diversiﬁed and conglomerate ﬁrms in the
United States, 1920-1970. Business History Review, 46(2), 202—19.

Dimmick, J. (2004). Introduction. Journal of Media Economics, 1 7(2), 85-86.

Dimmick, J, & Wallschlaeger, M. (1986). Measuring corporate diversiﬁcation: A case
study of new media ventures by television network parent companies. Journal of
Broadcasting & Electronic Media. 30(1), 1.

Doyle, G. (2002). Understanding media economics. London: SAGE, 2002.

Eisenmann, T. R. (1999). Organizational form and risk taking in the US. cable television
industry. Harvard Business School working paper.

F arjoun, M. (1998). The independent and joint effects of the skill and physical bases of
relatedness in diversiﬁcation. Strategic Management Journal, 19(7), 611-630.

F ofana, B (1997). Diversiﬁcation and market power in the United States cable TV
industry. Doctoral dissertation, the University of Nebraska.

The FCC (1997). 1996 Report on Cable Industry Prices. Retrieved April 20, 2006 from:
http://www.fcc.gov/mb/csrptpg.html

The FCC (2002). Ninth Annual Assessment of the Status of Competition in the Market
for the Delivery of Video Programming. Retrieved November 25, 2005, from:
httpz/lwww.fcc.gov/mb/csrptpg.htm1

The FCC (2004). Tenth Annual Assessment of the Status of Competition in the Market
for the Delivery of Video Programming. Retrieved March 10, 2006, from:
http://hraunfoss.fcc.gov/edocs_publie/attachmatch/FCC-05-1 3A1 .pdf

The FCC (2005a). Media Bureau Staff Report Concerning Over-the-Air Broadcast
Television Viewers. Retrieved March 20, 2006, from:
httpz/lhraunfoss.fcc.gov/edocs_public/attachmatch/DOC-257073A1 .pdf

The FCC (2005b). Statistical Report on Average Rates for Basic Service, Cable
Programming Service, and Fxruipment. Retrieved April 20, 2006 from:
http://www.fcc.gov/mb/csrptpg.html

The FCC (2006). 12th Annual Assessment of the Status of Competition in the Market for
the Delivery of Video Programming. Retrieved April 20, 2006 from:
http://www.fcc.gov/mb/csrptpg.html

110

Fry, T. R. L., Brooks, R. D., Comley, B. R., & Zhang, J. (1993). Economic motivations
for limited dependent and qualitative variables model. Economic Record, 69(205),
193-205.

Geringer, J. M., Tallman, S., & Olsen, D. M. (2000). Product and international
diversiﬁcation among Japanese multinational ﬁrms. Strategic Management
Journal, 21(1), 51-80.

Goolsbee, A., & Petrin, A. (2004). The consumer gains from direct broadcast satellites
and the competition with cable television. Econometrica, 72(2), 351-381.

Grant, R. M. (1988). On ‘dominant logic’, relatedness and the link between diversity and
performance. Strategic Management Journal, 9, 639-642.

Grant, R. M. (1991). The resource-based theory of competitive advantage: Implications
for strategy formulation. California Management Review, 33, 114-135.

Hall, E. H., & John, C. H. (1994). A methodological note on diversity measurement.
Strategic Management Journal, 15(2), 153-168.

Haugsted, L. (2005, August 10). Texas Legislature OKs Telco Bill. Multichannel News.

Head, S. W., Spann, T., & McGregor, M. A. (2001). Broadcasting in America: A Survey
of Electronic Media (9th). Boston, MA: Houghton Mifflin Company.

Healey, J. (1998, October 27). Phone-via-cable service tested. San Jose Mercury News,
1F, 4F.

Heam, T. (2006, March 13). House Near Consensus on National Franchise. Multichannel
News.

Hitt, M. A., Ireland, R. D., & Hoskisson, R. E. (2000). Strategic management:
Competitiveness and globalization. Cincinnati, OH: South-Western College
Publishing Thomson Learning.

Ito, K., & Rose, E. L. (2004). An Emerging Structure of Corporations. Multinatinal
Business Review, 12(3), 63-83.

Jung, J. (2003). The bigger, the better? Measuring the ﬁnancial health of media ﬁrms.
The International Journal on Media Management, 5(4), 237-250.

Jung, J. (2005). Impacts of media conglomerates’ dual diversiﬁcation on ﬁnancial
performance. Journal of Media Economics, 18(3), 183-202.

111

Kamien, M. 1., & Schwarts, N. L. (1975). Market structure and innovation: A survey.
Journal of Economics Literature, 13, 1-37.

Kashlak, R. J ., & Joshi, M. P.(l994). Core business regulation and dual diversiﬁcation
patterns in the telecommunications industry. Strategic Management Journal,
15(8), 603-611.

Kennedy, P. (1998). A guide to econometrics (4th). Cambridge, MA: The MIT Press.

Kim, B. (1997). Determining the number of programming options in cable television.
Journal of Media Economics, 10(4), 23-27.

Kochhar, R. & Hitt, MA. (1998). Research notes and communications linking corporate
strategy to capital structure: Diversiﬁcation strategy, type and source of ﬁnancing.
Strategic Management Journal, 19(6), 601-610.

Kolo, C., & Vogt, P. (2003). Strategies for growth in the media and communications
industry: Does size really matter? The International Journal on Media
Management, 5(4), 251-261.

Kranenburg, H. V. (2004). Diversiﬁcation strategy, diversity and performance among
publishing companies. In the proceedings of the 6th World Media Economics
Conference, Montreal, Canada, 2004.

Kranenburg, H. V., & Hogenbirk, A. (2005). Market structure issues. In A. Albarran, A.
& S. M. Chan-Olmsted (Eds), The Handbook of Media Management and
Economics (pp. 325-344). Mahwah, NJ: Lawrence Erlbaum Associates, Inc.

Kranenbrug, H. V., Hagedoom, J ., & Pennings, J. (2004). Measurement of international
and product diversiﬁcation in the publishing industry. Journal of Media
Economics, 17(2), 87-104.

Lacy, S., & Bauer, M. J. (2005). Future directions for media economics research. In A.
Albarran, A. & S. M. Chan-Olmsted (Eds), The Handbook of Media Management
and Economics (pp. 655-674). Mahwah, NJ: Lawrence Erlbaum Associates, Inc.

Litman, B. R. (1985). Economic methods of broadcasting research. In J. R. Dominick &
J. E. Fletcher (Eds), Broadcast Research Methods (pp. 106-122). Boston, MA:
Allyn & Bacon.

Lubatkin, M., & Chatterjee, S. (1994). Extending modern portfolio theory into the

domain of corporate diversiﬁcation: Does it apply? Academy of Management
Journal, 3 7(1), 109-136.

112

Lubatkin, M, & Rogers, R. C. (1989). Diversiﬁcation, systematic risk, and shareholder
return: A capital market extension of Rumelt’s 1974 study. Academy of
Management Journal, 32, 454-465.

Luo, Y. (2002). Product diversiﬁcation in international joint ventures: Performance
implications in an emerging market. Strategic Management Journal. 23(1), 1-20.

Markids, C. C. (1995). Diversiﬁcation, restructuring, and economic performance.
Strategic Management Journal, 16(2), 101-118.

Massey, K., & Baran, S. J. (2001). Introduction to Telecommunications: Converging
Technologies. Mountain Views, CA: Mayﬁeld Publishing Company.

Mayo, J. W., & Otsuke, Y. (1991). Demand, pricing, and regulation: Evidence from the
cable TV industry. RAND Journal of Economics, 22(3), 396-410.

Montgomery, C. A. (1982). The measurement of ﬁrm diversiﬁcation: Some new
empirical evidence. Academy of Management Journal, 25(2), 299-307.

Montgomery, C. A. (1991). Corporate diversiﬁcation. Journal of Economic Perspectives,
8(3), 163-178.

Montgomery, C. A., & Hariharan, S. (1991). Diversiﬁcation entry by established ﬁrms.
Journal of Economic Behavior and Organization, 15, 71—89.

NCTA (2005). 2005 Mid-Year Industry Review. Retrieved January 10, 2006, from:
http://www.ncta.corn/industry_overview/CableMid-YearOverviewOSFINAL.pdf

NCTA (2006). 2006 Industry Overeview. Retrieved July 28, 2006, from:
http://i.ncta.corn/ncta_com/PDFs/NCTAAnnual%20Report4-06FINAL.pdf

NCTA (February 2005). The video market is fully competitive: Almost 27 million
consumers now subscribe to cable’s competitors. Retrieved March 10, 2006, from:
http:// www.gacable.com/images/ competition-in-video-market-1-2005.pdf

Oster, S. M. (1999). Modern Competitive Analysis (3rd). New York, NY: Oxford
University Press.

Overview of Cable Modem Technology and Services (2006). Cable Digital News.
Retrieved July 5, 2006, from: http://www.cab1edigitalnews.com/cmic/cmic1.htrnl

Peteraf, M. A. (1993). The cornerstones of competitive advantage: A resource-based
view. Strategic Management Journal, 14, 1 79-190.

Picard, R. G. (2004). Strategic Responses to Media Market Changes. JIBS Research
Report.

113

Picard, R.G., & Rimmer, T. (1999). Weathering a recession: Effects of size and
diversiﬁcation on newspaper companies. The Journal of Media Economics,
12(1), 1-18. -

Piscitello, L. (2004). Corporate diversiﬁcation, coherence and economic performance.
Industrial and Corporate Change, 13(5), 757—787.

Porter, M. E. (1979). How competitive forces shape strategy. Harvard business Review,
57(2), 137-145.

Porter, M. E. (1980). Competitive Strategy: Techniques for Analyzing Industries and
Competitors. New York: Free Press.

Porter, M. E. (1985). Competitive Advantage. New York, NY: Free Press.

Prahald, C. K., & Bettis, R. A. (1986). The dominant logic: A new linkage between
diversity and performance. Strategic Management Journal, 7(6), 485-501.

Prahalad, C. K., & Hamel, G. (1990). The core competence of the corporation. Harvard
Business Review, 68(3), 79-91.

Reed, R., Luffman, G. A., (1986). Diversiﬁcation: the growing confusion. Strategic
Management Journal, 7, 29-35

Reed, R., & Sharp, J. A.(1987) Conﬁrmation of the specialization ratio. Applied
Economics, I 9(3), 393-405.

Ramanujam, V., &Varadarajan, P. (1989). Research on corporate diversiﬁcation: A
synthesis. Strategic Management Journal, 10(6), 523-551.

Robins, J ., & Wiersema, M. F. (2003). The measurement of corporate portfolio strategy:
Analysis of the content validity of related diversiﬁcation indexes. Strategic
Management Journal, 24, 39—59.

Rumelt, R. P. (1974). Strategy, structure and economic performance. Cambridge, MA:
Harvard University.

Sampler, J. L. (1998). Redeﬁning indusu'y structure for the information age. Strategic
Management Journal, 19(4), 343-355.

Scanlon, M. (January 24, 2005). Here’s how cable will win the HSD war: Features beat
Price. Cable World.

Scherer, F. M., & Ross, D. (1990). Industrial market structure and economic performance
(3rd ed.). Boston, MA: Houghton Mifﬂin.

114

Smith, J. G. (1985). Business Strategy: An Introduction. Oxford: Basil Blackwell Ltd.

Stimpert, J. L., & Duhaime, I. M. (1997). Seeing the big picture: The inﬂuence of
industry, diversiﬁcation and business strategy on performance. Academy of
Management Journal, 40, 560-583.

Teece, D. J. (1982). Towards an economic theory of the multiproduct firm. Journal of
Economic Behavior and Organization, 3, 39-63.

Tirole, J. (1988). The Theory of Industrial Organization. Cambridge, MA: The MIT Press.

The Television and Cable F actbook (2005). Washington, D. C .: Warren Communication
News.

Umstead, R. T. (August 29, 2005). VOD Shares PPV Burdens But Bears More Promise.
Multichannel News.

US. Census Bureau (2005). Housing Vacancies and Homeownerships: Fourth Quarter
2004. Retrieved Jun 5, 2006, from:
http://www.census.gov/hhes/www/housing/hvs/qtr404/q404tab6.htrnl

US. General Accounting Ofﬁce (2000). The Effect of Competition From Satellite
Providers on Cable Rates. Retrieved November 24, 2006, from: http://
www. gao. gov/cgi-bin/getrpt?GAO/RCED-00~ 1 64

Vannoni, D. (2004). Causes and effects of multimarket activity: From theory to empirical
analysis. Managerial and Decision Economics, 25(3), 163.

Venkatraman, N., Grant, J. (1986). Construct measurement in organizational research: A
critique proposal. Academy of Management Review, I I, 71-87.

Wemerfelt, B. (1984). A resource-based View of the ﬁrm. Strategic Management Journal,
52, 171-180.

Winter, S. G. (1987). Knowledge and competence as strategic assets. In D. J. Teece,
(Eds), The Competitive Challenge: Strategies for Industrial Innovation and
Renewal(pp. 159-184). Cambridge, MA: Ballinger Publishing Company.

Wolf, B. (1977). Industrial diversiﬁcation and internationalization: Some empirical
evidence. Journal of Industrial Economics, 177- 1 91 .

Woroch, G. A. (1997). Turning the cables: Economic and strategic analysis of cable entry

into telecommunications. In E. Noam (Eds), Globalism and Localism in
Telecommunications (pp. 45-62 ). Elsevier Science.

115

Yip, G. S. (1982). Diversiﬁcation entry: Internal development versus acquisition.
Strategic Management Journal, 3, 331-345.

Young, D. P. T. (2000). Modeling media market: How important is structure. The
Journal of Media Economics, 13(1), 27-44.

116

   

IILIlllljlllllljj][[1]]