Channel diversity
An Integrated Model for Determining Channel Diversity
in a Multichannel Environment
I. Introduction
A quick survey of current trade press covering cable television industry
tells that prevailing concern of cable system operators and cable
networks,
especially those up-starting, lies at the impact of re-regulation and rate
roll-back. Reduced profits of a local systems will limit the number of
new services that the operators might intend to add and this, in turn,
will
result in difficulty of networks in achieving enough level of access to
the audience. Accumulating critical mass of subscribers, which is
usually
said to be 30 million, is critical for up-starting networks. There
are
numerous new and proposed basic cable networks, some up-starting and
some
spin-offs from existing networks, which plan to launch in no time
("New
network ...", Cablevision Aug 8, 1994). These networks try to promote
themselves to the system operators in addition to consumers despite the
hostile regulatory environment in anticipation of so-called
'500-channel'
television environment. Established networks are no exception in
competing
with one another to increase access coverage1. Access through local
system carriage is the foundation of business for a cable programming
network -whether it is basic, premium, or pay-per-view network.
Economies
of scale are realized from the network's point of view since first
copy
costs of program acquisition and programming (putting programs together
as
a schedule) per viewer gets smaller as it achieves higher coverage.
Coverage rate of a network depends on each operator's system programming
decision -- how many channels to put together as a bundle and which
networks to select.
Even though spectrum scarcity is not a problem for cable television
industry, not every network, even the best, is guaranteed to have 100%
coverage. First of all, there is a physical limit due to the channel2
capacity of the kind of wire used for cabling. This technological
capacity
is the primary limitation on how many networks a system can have in its
system programming line-up. A system with channel capacity of 12
cannot
carry more than 12 networks without going through some costly plant
upgrade.
Then the following questions arise. Is technological capacity only factor
determining the number of programmed channels(channel capacity carrying
programming networks) television viewers can receive? Will
'500-channel TV
environment' be panacea for increasing the array of services viewers
receive since all 500 channels will be programmed with various kinds of
broadcasting and narrowcasting networks? If there are factors
affecting
the number of programming networks a local cable system can carry
other th
an channel capacity, finding out what they are and in what way they
affect
cable system programming in current environment are important not only
to
better understand the current state of cable television industry but
also
to make more realistic predictions about the '500-channel world' --
television without technological limit.
This study borrows theoretical framework from a model developed to
understand newspaper industry. We can always learn lessons from other mass
media industries. A comparison can be made between bundling diverse
sections together in a newspaper and bundling different programming
networks together in a local cable system. Before introducing full model,
a simplified example seems appropriate at this point. A small local
town
newspaper (for example, 'Evanston Review') cannot have sections such
as
'Woman news' or 'Good eating', which are included in big city papers
such
as 'Chicago Tribune'. 'Evanston Review' is not technologically
limited
from including those minor-interested sections in its paper. However,
'Evanston Review' is economically constrained from doing so. Because
with
smaller subscriber base 'Evanston Review' relies on, it is simply not
economically viable to have all the diverse sections that 'Chicago
Tribune'
would have. Smaller subscriber base means less subscription revenue and
less advertising revenue that can sustain such diverse contents.
Are small cable systems free from such economic constraints that 'Evanston
Review' faces in this example? Will technological upgrade solve all the
constraints and make even systems serving small number of cable
subscribers
carry as many programming networks as big urban systems have?
Economic constraint of the local market is not solved by technological
advance alone since an operator must decide the number of channels to
program based on marginal revenue and marginal cost its own market
incurs.
That is, an operator would add a network as long as the marginal revenue
potentially earned from adding an additional network covers the
marginal
cost (such as added operating cost and transaction cost) it incurs.
Even
though there are more programming networks than available channel
capacity
in the systems, some operators might choose to program only certain
number
of channels that is economically viable leaving some portion of
channel
capacity unused.
Ultimately, the willingness for consumer to subscribe to cable depends on
the quantity and the quality of the programming the local cable
operator in
its area offers. On the other hand, cable operators adjust quantity and
quality of service in addition to marketing strategies and
infrastructure
investments so as to maximize returns under each circumstance(Hazlett,
1994). An operator decides how many and what kind of services its
system
offers based on the economic condition of local market in which a
system
is located as well as the demographics of the service area (assessment
of
potential and actual subscriber preferences). Although 100% coverage
(that is 62% of total TV households in the U.S. as of this writing) is
ideal for every programming network, the rate cannot but be constrained
by
economic factors affecting local systems' carriage decisions. Cable
systems need operational flexibility -- so that they can properly respond
to market changes and regulatory structure changes--, and for that
reason
alone some systems might not want to maximize the number of programmed
channels up to the channel capacity(Solomon, 1989). Some networks are
bound to be left out from being carried. Varying rate of coverage
determines the competitiveness of the programming networks. 1993 Myers
Reports Survey of Cable Operator Executives on Basic Networks shows
that
the majority number of the up-starting networks are only considered by
less
than 5% of the systems surveyed. To some extent, competition among the
wholesalers of cable programs, that is, programming networks, is
determined
by the system-level programming decisions. Chipty(1993) listed three
kinds of effects local system's carriage decision had on the
profitability
of programming networks: advertising revenue, popularity of the
network
(externality), and quality of programs. Studying economic factors
affecting local cable system programming3 will be a starting point in
understanding why only a limited number of networks survive in cable
television industry.
It has been often assumed that the number of programmed channels of the
local system increases and thus more networks become viable with
higher
coverage as long as channel capacity increases to contain them. It was
the
kind of optimism that prevailed in the Sloan commission report (Sloan
Commission, 1971) in the early days of cable television, which
resurrected
recently with the '500-channel' scenario. This study attempts to find
out
economic factors affecting local system operators' system programming
d
ecisions. And as we have many more emerging video delivery technologies
of
multichannel nature which emulate the programming, operation and
management of cable television, it is meaningful to examine the behavior of
'old' industry at this point of time.
The concern has been that cable operators, with its market power
(resulting from natural monopoly position), would have the incentives to
restrict the number of channels programmed and to reduce the number of
sources and discriminate against competitive source. (Thorpe, 1985) The
reasoning behind cable regulations has been that without competition
or
regulation cable operators would be able to charge monopolistic prices
to
the subscribers and to control the content of a large number of
program
channels (Noam, 1985). This study provides opportunity to empirically
examine such claims and to provide benchmark for regulatory concerns.
The following section reviews why studying the determinants of programmed
channels in a cable system is important along with Wildman model of
newspaper content diversity which will be used as a major theoretical
framework. Section III deals with literature review. Section IV
proposes
an integrated model for determining channel diversity (defined in
section
II), which is empirically tested in the subsequent sections. Section
V as
a brief description of data and method. Section VI reports results of
stat
istical analysis and section VII concludes with discussing
implications of
the result.
II. Theoretical framework
1. Channel diversity
Programs and the audience they attract are the products by which media
industry's performance is measured. Diversity in programming has been
one
way to evaluate the television industry. Diversity is not only
important
to increased consumer surplus, but also has the political value of
providing citizens with more information and access to a wider range of
viewpoints. However it is hard to define or measure diversity. There
are
numerous ways to approach diversity.
Scholars of communication and economics have tried to measure diversity in
television network programming in relation to economic factors such as the
level of market concentration, competition among networks, and the number
of channels over many years. To list a few, Greenberg and Barnett
(1971)
examined the relationship between program type and the number of
channels;
Dominick and Pearce (1976) found market concentration inversely
correlated
with diversity; while Litman (1979) found just the opposite. These
scholars measured diversity by the number of program types (which are
commonly stated as program genres) available within a given time and the
range of viewing options available within a program type.
Diversity is multi-dimensional in terms of the scope of the definition and
also across time. It can be the number of program types or it can be the
number of options available within a particular type. It can be
examined
across channels at a given point of time or longitudinally.
The following studies have examined the diversity of cable programming.
Wildman and Lee(1989) examined program repetition rates and also
devised a
diversity index based on an industry definition of the program genre.
Waterman and Grant (1991) analyzed programming origin, subject, and
format
of 3 over-the-air broadcasting networks and 34 cable networks during
1986.
De Jong and Bates (1991) defined diversity as the number of networks
cable system operators carry. Waterman (1986) offered an explanation
for
why direct pricing of programs to viewers and increased channel
capacity
could not bring the 'narrowcasting' of 'high-culture' programming into
being. He pointed to a shortage of demand - a shortage of viewers with
high willingness-to-pay - as the critical factor. And he also pointed
out
that increased channel capacity was not used for new and fresh kinds
of
programming but for program repetition within a channel and intermedia
repetition (sequential windowing) of mass-appeal programming.
As mentioned above, diversity is a complex concept and it can be measured
in many different ways. Whichever way it is measured, defining
diversity
requires a lot of subjective judgement and has unavoidable
limitations. It
is hard to refute the argument that every program is different and at the
same time one can say that all programs are essentially the same.
Owen(1977) argued, while pointing out the limitations of counting the
genres/formats as a way to measure diversity in television programs, that
it is unreasonable to assume that consumers do not obtain some
positive
utility from having choices among substitute programs within the same
basic
format and that diversity of format/genre is unrelated to any measurable
economic index of consumer well-being. Information about consumer
demand
for program genres/formats is very difficult to collect.
Following Owen's view on diversity, this research regards the number of
channels programmed by an operator as one way of measuring diversity
in a
multichannel television. In this definition, channel includes the
numbers
of over-the-air channels, distant broadcast signals, basic and premium
cable networks, public access channels, and pay-per-view channels--the
number of total channels programmed by an operator. Each network
carried
is assumed to be equivalent to 1 hypothetical unit of diversity.
Assuming
no two identical programs are aired at the same time by more than one
cable
channel in a system, one additional channel programmed by an operator
means that a subscriber/viewer has one more program to choose from at a
given point of time.
2. Wildman model of newspaper section-bundling
Wildman model(1991, lecture; 1994) of newspaper content diversity
addresses why sections of a paper are added or dropped. This model
provides a framework which can be applied to the examination of the
diversity in cable system programming.
Let Fi = First copy cost of section i
MCi = Marginal cost per copy of section i
Si = Fraction of subscribers that read section i
C = Circulation of the newspaper
Ri = Revenue generated by a reader of section i.
When deciding whether to add section i to the paper, a newspaper publisher
will equate revenue generated by adding section i with the cost it incurs
and will add a section if the cost is equal or smaller than the
revenue
expected. This condition is written as
Ri * Si * C = Fi + (MCi * C).
If we rearrange this equation, we have critical threshold proportion of
subscriber on which the decision for adding a section depends on. Let
this
critical fraction of subscribers be Si.
Si = (MCi / Ri) + (Fi / Ri*C).
This final equation tells that as cost relative to revenue goes up the
critical threshold also goes up. Moreover, it suggests that with large
circulation critical fraction of the readers that makes a particular
section financially feasible goes down.
According to the model, larger circulation reduces critical fraction of
the readers that makes a section financially viable. For example,
while a
paper with circulation of 100,000 needs only 10% of its readers'
interest
in the new, say, 'womenews' section which it considers adding, another
paper with 10,000 readers need 100% of its readers' interest in the
newly
proposed content to offer a comparable addition. With large group of
readers, more diverse sections can be published. In addition, each
section
comes to have higher quality since more money is invested in anticipation
of subscription and advertising revenue based on the size of readers.
These factors make the paper with large circulation more attractive to
the
audiences with wide ranging interests. Bundling of diverse features
in a
paper can be compared to bundling of diverse programming networks in
a
cable system despite a number of differences between the two media. At
the
heart of the analysis is a more general model of inter-temporal inter
-media flow of media products (Wildman, 1994).
This model of newspaper section-bundling suggests that the primary
relationship between the number of programmed channels and the number of
cable subscribers in the system area should be examined. The model
predicts that, as the number of subscriber increases, the number of
programmed channels offered in a system will increase holding all other
factors constant. The larger the subscriber base, an operator would
put
more video programming up to the point where marginal
revenue-subscription
revenue and advertising revenue- becomes zero.
An operator can expand revenue by the ancillary services such as
pay-per-view and subscription to premium channels. The amount of ancillary
revenue becomes larger when subscriber base is large. An operator might
still want to increase the number of programmed channels as long as
the
cost of adding a channel is justified by just maintaining current
subscription rate (that is, reducing the churn rate). The relationship
between subscriber base and channel diversity should be measured
holding
the channel capacity constant since it poses a limit to the number of
channels possibly programmed.
An integrated model proposed and tested in the study is an application of
the theoretical reasoning explaining newspaper content diversity. In
determining the primary relationship between subscriber base and channel
diversity of system programming, economic factors affecting revenue
and
cost of local cable systems will be incorporated in the integrated
model.
III. Previous research
Unfortunately, there has been little research to date which has focused on
systematic analysis of cable system programming. Paucity of previous
literature partly reflects lack of research in the area of cable
television
as a whole. The differences among systems in terms of the number of
available channels not in use and the reason behind it have not been
addressed yet. However, the factors affecting the profitability of cable
system (thus affecting marginal cost/revenue of adding a programmed
channel) have been explored in the following studies.
Evaluating viability of cable systems
Vogel (1990) points out that profitability of a cable system is usually
measured by population density of the area the system serves and cable
penetration figure since much of a system's operating cost is fixed and
independent of subscriber numbers and the only major variable costs are
drop charges and the cost of installing converter boxes. It means that
there is economies of scale realized by population density and system
size.
Some factors were found to positively affect basic penetration, which
results in revenue increase. The more stations a system carried, the
fewer
of each type of OTA(over-the-air) stations receivable, the older the
system was, the farther away the area was from OTA stations, the lower
the
price was, and the higher the average household income was, the higher
the
penetration rate was. Park(1971) also acknowledged that the cost
factors
might limit the number of services a system can provide. Work on the
costs
of cable systems suggested that penetration on the order of 40 to 50% were
necessary to support advanced local origination.
Baer & Park (1972) conducted financial projections for the Dayton Miami
Valley area and revealed that financial results varied depending on
subscriber penetration, monthly subscriber fee, and characteristics of
geographical area covered.
Woodard(1974) suggested a list of criteria by which a franchise area's
viability as a cable market could be evaluated. They were: Size in
terms
of the number of subscribers, cost of plant, number of franchise,
population density, penetration rate, community growth in terms of new
homes as potential subscribers, state of the economy of the community,
presence of local college or university both as a potential programming
source and as a measure of culturally up-scale audience, cost-of-living
index, estimated average subscriber billing, availability of FM radio
stations, franchise terms, potential advertising time sales, and
application expense. He also emphasized the availability and the number of
local television stations. He recommended cable systems to give the
subscribers maximum number of programmed channels as an incentive to
subscribe.
Kent G. Webb's (1983) book titled The economics of cable television is
devoted to various economic aspects of cable television. Especially
relevant to this study is econometric analyses estimating cost and demand
of cable television.
Sources of cost were categorized into equipment, programming cost,
operating expense, and franchise fee. Equipment cost consisted of headend,
distribution plant, subscriber interface, and studio for local
origination. Operating expense was mostly labor cost, especially technical
labor. He estimated operating cost and depreciation cost, in his
econometric equation, on the basis of miles of plant and size of the
system in terms of subscriber count. He concluded that cable system was
a
natural monopoly with declining average total cost with the number of
channels, number of subscribers, and size of the geographical area
(miles
of cable). He also pointed out that most dramatic is the declining
cost
per subscriber given a cable system of a fixed channel capacity and
length
of plant. Part of the reason was spreading headend cost. Since he
used
1982 data, we should be cautious in accepting the results as of today.
Among statistically significant socioeconomic variables determining demand
of basic cable were per capita income, index of home equipment, and
education. Among the most important determinants of demand for basic
cable
were the number, type, and quality of signals carried by the system
compared with those available over the air in the local market. Imported
signals resulted in increasing penetration and consumer surplus while
marginal cost of adding an imported signal was quite low.
Pacey(1985) tested a model of the demand for basic service. Her
independent variables were factors describing cable system, subscriber
demographics, and local market characteristics which she called
environmental characteristics. OTA signals were separately entered as
primary network, duplicative network, independent, educational, and
local
origination. What was unique about the study was that she included
pay
television characteristics in estimating demand for basic cable. Among
the
findings were that urban subscribers were more likely to be responsive to
subscription fee than rural and that demand for basic cable was quite
elastic with respect to the price of cable television.
Channel diversity of cable system programming
Eastman(1989) lists 4 elements affecting system programming: Legal
carriage requirements, technology, economics/cost which are license fee,
signal importation fee, spot availability, promotional support,
satellite
placement, and marketing considerations meaning demographic and
psychographic composition of coverage area local audience. Demographic
factors are commonly understood as a major determinant of the mix of
cable
networks.
Dejong & Bates's (1991) study on channel diversity defines diversity as
the number of channels. They tracked absolute and relative diversity
according to the definition of Levin(1971) at three points of time.
Absolute diversity was operationalized by the number of different channel
types carried by a system divided by the total number of channel
types4 for
the cable industry. Relative diversity was operationalized by the number
of different channel types divided by the channel capacity of the
system.
Diversity5 was measured at three points of time, 1976, 1981, 1986,
roughly
responding to periods of high, moderate, and no regulation. They
found
that diversity increased over time. But, the growth in relative
diversity
was substantially less than that of number of channels, and the
relative
and absolute diversity measures indicated that the average cable
systems
offered less than half of its potential for diversity. The authors
believed that greater channel capacity and regulatory freedom fostered the
growth and expansion of cable.
However, the authors did not provide any answers as to why cable had not
lived up to its full potential, as to why a system operator had
channels
not in use when there were more than enough programming services to
fill up
to the channel capacity. Their research was limited to be descriptive in
nature. Channel capacity might remain an economically scarce commodity
despite the advancement of technology. This study aims to provide
answer
for that.
The relationship between channel capacity and channel diversity should be
considered with a caution that capacity is given as an exogenous
variable
once constructed. It is possible that a system with large capacity
programs only small number of channels because the overly optimistic
information estimated at the time of franchise application and
construction
can be corrected over the course of operation as more accurate information
on the profit potential of a specific market is revealed.
An econometric analysis of competitive effect of broadcast signals on the
performance of cable, controlling system and local market
characteristics
was done by Dertouzos & Wildman (1990). Performance of cable system
was
operationalized in three ways: subscriber counts, program service
offerings, and prices of these services. System-related control variables
were length of system, homes passed, age of headend, channel capacity,
and
whether or not the system is managed by an MSO (multiple system
operator).
Market demographics included ethnic composition of population, projected
population growth, employment, income, home and VCR ownership, and
geographical location of the system. The study concluded that five OTA
signals constituted effective competition to local cable system. The
effect of system and market characteristics on basic cable programming
was
also dealt in this study. Channel capacity, projected population
growth,
MSO, and to the lesser extent employment were statistically
significant
variables influencing the number of basic services the system offers.
Smaller systems expand the number of basic networks rapidly with increases
in channel capacity. As capacity grows beyond fifteen, the percentage
change in basic offerings increases less rapidly.6 These results of
cross-sectional analysis were similar to what Dejong & Bates (1991) found
in their analysis - that increases in channel diversity were less than
those in channel capacity over time. The results show that marginal
value
of an additional channel tends to decrease as channel capacity
expands.
Additional subscribers picked up by one more programming channel
should be
increasing at a decreasing rate. There are other economic factors
limiting
the number of services carried in a system, which this study will find
out. Rural areas, areas with higher population growth projection, MSO
managed systems and systems facing higher competition from OTA TV
carried
more basic networks.
In their another study, Dertouzos & Wildman (1993) emphasized that the
cost of running a cable system made critical differences between
markets
and that factors which accounted for the systematic difference between
systems should be accommodated when studying cable television.
According
to their analysis, system age largely dictates channel capacity which
is a
driving force behind channel diversity. In addition, younger systems
are
more likely to have more service and/or less expensive in regards to
its
basic service because operators can expect higher ancillary revenue
through
technological capacity such as PPV(pay-per-view) and local commercial
insertions.
Thorpe(1985) did a study of the effect of competition from non-cable
programming service on the market power of cable system. A part of the
study explored the factors affecting programming decisions of a system:
the total number of cable channels programmed and the number of pay
television programs offered. Population of the franchise area and channel
capacity were positively associated with the number of programs
offered.
Age of the system, TV market ranking, and household income were
negatively
associated. Competition from subscription TV, ownership affiliation
between the system and a pay TV programmer, MSO, presence of rate
regulation, and physical obstructions of television signals were included
in the analysis, but did not appear to influence the number of cable
programs offered.
IV. An integrated model for determining channel diversity of cable
television
1. Economic factors affecting revenue and cost of running local cable
systems
Following is a description of the factors affecting the profitability of
local cable systems. The factors are categorized into three groups:
Market/environmental characteristics, local system characteristics, and
marketing considerations/audience demographics. These factors are
included
in the model predicting channel diversity of a cable system and
empirically analyzed in the subsequent analysis.
(1). Market/environmental characteristics
A. Competition from other media
Factors relating to the competition from other media such as VCR
penetration and the number of over-the-air stations
(networks/independent/public) have been used heavily in the previous
researches. An operator, facing higher competition from other media,
might
increase or decrease the quality of service including channel diversity.
In addition, there is a gap between urban and rural areas in terms of
general entertainment options. Urban environment will have a lot more
entertainment options alternative to cable television. Urban markets
should have comparatively abundant OTA services and incur higher
construction cost per mile.
B. Population
Population of the system area, population density, and expected population
growth rate are expected to be positively correlated with the channel
diversity.
Size of the total population in the area and expected population growth
rate measure potential market size (potential future revenue). Two
systems
serving same number of subscribers but with different growth rate or total
population might behave differently in determining the number of channels
to program. The one with larger potential subscriber has higher
incentive
to increase channel diversity since expected revenue from new
subscription
is higher.
C. Income
Median household income in 1989 is used to measure the income level of the
area. It is known that those with higher income are more likely to
subscribe cable. In addition, higher income audience increases
attractiveness of the medium to the advertisers.
D. Retail sales growth rate
Retail sales growth increases local advertising demand, resulting in the
increase of marginal revenue by adding a programmed channel. Average
retail sales growth rate of the area over last 5 years is used.
E. Projected population growth rate
Systems located in the area with higher population growth projection have
higher incentive to increase quality of service since potential market
is
expanding. Higher population growth will positively work for channel
diversity.
F. Average wage
Systems located in high wage areas face higher cost of operating cable
system. Higher average wage of the area will negatively affect channel
diversity.
G. VCR penetration rate and alternative multichannel media competition
VCR ownership rate can another competition for cable television.
Competition from other multichannel provider should obviously affect the
operators' management decisions. The direction of the influence is
ambiguous.
(2). System characteristics
A. Channel capacity
Channel capacity means the maximum number of channels that the wire can
transmit within the intended service area. In Dertouzos & Wildman
(1993),
channel capacity was significant factor in determining the number of basic
network a system carried. Channel capacity physically limit the possible
number of the service provided. Marginal value of available but not
in use
channel can be higher in a system with smaller channel capacity.
Webb(1983) found out that the operating expense (excluding programming
costs) was not measurably different for a 12-channel or 36-channel
system.
It means that the average operating expense per channel declined as the
number of channel increased.
According to 1993 Myers Report, 20.6% of 603 cable system executives
interviewed said that they had no channel capacity expansion plan at the
time: That corresponds to 42.0% of systems of under 10,000
subscribers and
26.4% of systems with less than 49 channel capacity respectively while
only 5.8% of the systems with 50,000 or more subscribers and 17.5% of
systems with 49+ channels had no plans for expansion. Systems serving
smaller number of subscribers are less likely to expand channel
capacity.
B. Age of the system
Older systems are expected to have smaller channel capacity. Moreover,
newer systems are likely to have improved amplifier which can enlarge
channel capacity. However, considering the accumulated marketing efforts
of older system, the penetration rate is likely to have large
subscriber
group (probably at the saturation stage).
Age of the system also makes difference in other technological
capabilities such as pay-per-view, addressability, and local commercial
insertions. With the basic cable rate held constant, system operators
have
higher incentive to increase number of channels offered in the basic
package because of the higher ancillary revenue from the
subscribers(Dertuozos & Wildman, 1993).
Older systems are more likely to be located in poorer OTA television
reception areas.
C. MSO management
It needs to be questioned whether MSO-managed systems are better for
consumer welfare in terms of programming service, that is increased
channel
diversity. If MSO systems are indeed better and efficient, channel
diversity should increase as the size of the MSO gets larger holding other
key factors constant.
D. Density of the area
Higher density means low cost of cabling and operation (lower units per
mile). Length of plant per household can be used as a proxy for
density.
E. Vertical integration
Systems vertically integrated with programming networks have higher
incentive to carry integrated network, which can add to the number of
channels programmed. But on the other hand they also have incentive to
reduce the number of networks competing with the integrated network for
viewership. In that case integrated relationship can negatively affect
channel diversity.
F. The proportion of addressable households
Addressability itself can be another measure of diversity. Systems with
large fraction of addressable households have higher incentive to
offer
services which require addressability.
(3) Marketing considerations / Demographics
In order to attract local subscribers and local advertisers, an operator
must be sensitive to peculiar preference and need of the target group.
They say that keeping the current subscribers (that is, lowering churn
rate) is as important as attracting new ones.
Demographic variables which are expected to influence the demand of cable
television or TV viewing behavior should be included. Previous
research
indicated that age, the number of households with children, education,
household size, income, and demographic diversity make difference in
cable
television subscription. As demographic diversity measure, percent of
population who speak other than English and proportion of white
population
are used.
Younger, higher-income, and higher-educated population is more likely to
subscribe cable. Also households with children and those with larger
family are more likely to do so. Diversity of the population should be
positively related to channel diversity.
2. The model
Following model of channel diversity (total number of channels programmed
in a system) has been constructed and put to a test.
Figure 1
[--- ??? Graphic Goes Here ---]
There are two dependent variables: diversity and hhsub (number of
subscribing households). Aside from the predicted relationship between the
independent variables and dependent variables discussed earlier, there are
several points to make in the model.
The two major dependent variables are diversity and the size of
subscribing households. However, one becomes an independent variable to
each other. As diversity increases there is higher incentive for a
viewer
to subscribe cable as long as a viewer can derive utility from getting
a
diverse menu of television programming. If a viewer derives utility
only
from getting a clear reception of over-the-air signals, which indeed
is
known as one of major reasons for subscribing cable, diversity might
not
matter much for those viewers. However, it is reasonable to assume
that a
viewer perceives the quality of cable television higher when there are
more
channels offered in the service.
On the other hand, as explained earlier, a system operator has higher
incentive to offer more channels when it has more subscribers to serve.
In
the same context, the total number of households in the franchise area
also affects operator's decision on diversity. The total number of
households, regardless of the subscription rate at a given point of time,
reflects potential maximum market size that an operator can expect.
Subscription rate varies constantly in any market and keeping current
subscribers is as important as getting non-subscribers to sign up the
service. The total number of households represents the maximum number of
subscriber an operator can get if the system increases diversity.
Therefore, according to this model the variable 'diversity' (total number
of programming services programmed in a system) is explained by the
number of subscribing households (HHSUB) and all of the independent
variables. For example, the demographic variables such as 'age50' and
'hhsize' affects channel diversity even though in the figure 1 above it
looks like it only affects the number of subscribing households. They
affect channel diversity indirectly through the variable 'hhsub'. In
the
same way, 'hhsub' can be explained by diversity and all other
independent
variables in the model.
Another important variable is channel capacity which is a key independent
variable affecting channel diversity. Channel capacity is given after
the
time of construction. However, at the time of construction channel
capacity is decided based upon the market potential of the area. Expected
revenue should be measured by size of the system and wealth of the
dwellers. Assuming current size of households and median income are highly
correlated with those at the time of construction, HH(total number of
households in the area) and INCOME(median income of the population of
the
area) are expected to influence channel capacity. Current HH and
INCOME
are used as proxies for those at the time of construction.
Price of basic subscription is an independent variable affecting
subscribership. At the same time it is influenced by the number of
channels offered. However, price and channel capacity are not completely
determined by factors specified within the model. Thus, it is hard to
accurately measure the determinants of those two variables. Determining
them is out of the scope of this paper.
Table 1: Variable name, definitions, and source
Variable Description Source
TOTAC Total # of channels carried by the system (all kinds FCC
of channels)
HHSUB The # of households subscribing to cable FCC
HH The # of households within the system area FCC
CH-CAPA Maximum number of channels the system can carry TV&cable factbook
AGE Age of the system principal headend FCC
VI The number of fulltime national programming networks
vertically integrated with the system operator (5%
equity or more) Waterman & Weiss(1995)
MSONU The total number of systems operated by the FCC
system operator
AVGWAGE Average wage of the county where the system is REIS7
located
FRANFEE Franchise fee per subscriber the system pays to the FCC
franchising authority
DENS Density of the housing units measured by the FCC
number of hh passed by the line divided by the line
miles
PRJ-POP Projected population growth rate Rand McNally (1994)
RTLGRW Average retail sales growth rate over the past 5 Survey of buying
years power
ADDRESS % of addressable hh FCC
INCOME Median income of the area FCC
LOCAL The number of local broadcast television signals FCC
carried by the system
VCR VCR ownership rate Nielsen
COMP Existence of a competing multichannel provider FCC
1MC Monthly charge to tier 1 (Basic subscription fee) FCC
URBAN % of population living in the urban environment FCC
AGE65 % of population aged over 65 Census
EDU-COL % of college graduates Census
TONGUE % of population who speak other than English FCC
CHILD % of households with children FCC
WHITE % of white population FCC
HHSIZE Average number of persons living in a household Census
1DTV Number of distant broadcast signals carried in tier 1 FCC
1LTV Number of local broadcast signals carried in tier 1 FCC
1SAT Number of satellite-delivered cable only channels FCC
carried in tier 1
1TTOT Total number of channels programmed in tier 1 FCC
The major hypothesis is that subscriber base will be a strong predictor of
channel diversity holding the channel capacity (and other economic
factors) constant.
Large subscriber base would make it viable for a system to have more
channels programmed. Various factors affecting marginal cost and
marginal
revenue of adding a programmed channel will also determine the channel
diversity. Audience characteristics will also matter.
The number of channels to program in a cable system will vary as a
function of local market condition, system-specific characteristics, and
demographic characteristics. Since relative revenue and costs are
related
to such observable factors, predictions could be made about the
circumstances under which a system is more likely to increase channel
diversity. As capacity increases, there should be a point where
consumers
get less redundant and truly diversified service from the cable system
of
the area.
V. Method of data analysis
The study analyzes FCC cable TV rate survey database collected through
December, 1992 and February, 1993. Since the period is pre-regulation
(of
1992) it is assumed to reflect market solutions without distortions
possibly introduced by regulation. There are 496 cases in the final data
set but this study used only random sampled 293 cases. Excluded
cases
were purposive sampling of top 100 systems, small systems and overbuild
systems. The unit of analysis is a cable system.
Additional information was added to the original FCC database. VCR
penetration (on the county-level) was acquired from Nielsen Research.
Channel capacity was recorded from Television & Cable Factbook (v61).
Population growth projections by county were recorded from Rand McNally
Commercial & Marketing Atlas(1994). Retail growth rate over the past 5
years by county was computed from Survey of buying power(1993, 1992,
1991,
1990, 1989). Additional demographic information such as average
family
size, education, and age group was added from 1990 Population & Housing
Census.
Since cable systems are varied in terms of their profiles, a multivariate
econometric analysis is desirable to account for the differences among
the
local cable systems. Economic factors(including audience factors)
that
might constrain channel diversity of a local system could be fully
explored
through data containing local characteristics.
Two-stage least squares estimate was used for the multiple regression
analysis. This particular type of regression estimate was necessary
since
the two key dependent variables in the model are at once independent
variables to each other. Two-stage least squares estimate purges the
correlated errors introduced by this mutual relationship and correctly
measures two equations determining each dependent variable.
V. Results
After deleting variables which turned out to be irrelevant and specifying
appropriate functional form, the following model was produced to represent
the relationship between subscriber base and channel diversity. In the
following model, broken line represents statistical significance at
p=.1
level. All others are significant at p=.05 level.
Figure 2
[--- ??? Graphic Goes Here ---]
Table 2: The relationship between dependent and independent variables
Dependent var TOTAC LNSUB
TOTAC +
LNSUB +
HH - +
INCOME +
AGE +
CH_CAPA +
RTLGRW +
FRANFEE +
DENS +
LOCAL +
COMP + -
ADDRESS +
VCR +
PRICEBAS -
AGE50 -
HHSIZE -
It turned out that the relationship between channel diversity and the
number of subscribers was curvilinear. Diversity increased as the
subscriber base increases at first and then beyond certain point did not
increase. Diversity does not seem to increase with system size beyond
the
point approximately between subscriber count 50,000 and 100,000.
Natural
log of HHSUB(LNSUB) was used in place of HHSUB to make the curvilinear
relationship fit for the linear model. Transforming HHSUB into LNSUB
resulted in a nice linear relationship between the two key dependent
variables.
The variables 'urban' and 'dens' were highly correlated to each other.
Density was used for both equations in place of 'urban'. The
variable,
'urban' was added from other source (Population Census) to FCC survey
data,
and the area for which the proportion of urban dwellers was measured may
not exactly match with cable system area. On the other hand, density
was
measured by the number of households passed in the system area per
mile of
plant. Density measure seemed to be more accurate.
Insignificant variables were removed from the model for the sake of
parsimony. Removing them increased explanatory power of the model
(measured by adjusted R-squares).
Table3: Statistical results of an integrated model determining channel
diversity
1. First equation: Dependent variable.. S7_TOTAC
Multiple R .89329
R Square .79797
Adjusted R Square .79027
Standard Error 5.20116
Analysis of Variance:
DF Sum of Squares Mean Square
Regression 9 25216.672 2801.8524
Residuals 236 6384.280 27.0520
F = 103.57271 Signif F = .0000
------------------ Variables in the Equation ------------------
Variable B SE B Beta T Sig T
LNSUB 3.791428 .537406 .668934 7.055 .0000
CH_CAPA .240043 .033391 .282083 7.189 .0000
FRANFEE .190095 .096449 .072480 1.971 .0499
DENS .032340 .008633 .137848 3.746 .0002
RTLGRW .104406 .053541 .058675 1.950 .0524
ADDRESS .040013 .014675 .098688 2.727 .0069
S4_COMP 3.441420 .911228 .107845 3.777 .0002
LOCAL .539727 .120130 .150367 4.493 .0000
S2_HH -2.96799872E-05 6.9973E-06 -.226090 -4.242 .0000
(Constant) -12.852690 3.331726 -3.858 .0001
2. Second equation: Dependent variable.. LNSUB
Multiple R .86923
R Square .75556
Adjusted R Square .74407
Standard Error 1.00484
Analysis of Variance:
DF Sum of Squares Mean Square
Regression 11 730.31233 66.392030
Residuals 234 236.26949 1.009699
F = 65.75430 Signif F = .0000
------------------ Variables in the Equation ------------------
Variable B SE B Beta T Sig T
S7_TOTAC .087871 .010894 .498042 8.066 .0000
S2_AGEHE .041216 .008752 .149806 4.710 .0000
S2_HH 9.29543000E-06 9.0972E-07 .401334 10.218 .0000
S4_COMP -.517076 .182013 -.091841 -2.841 .0049
VCR .011968 .005479 .073299 2.184 .0299
LOCAL -.025625 .027514 -.040463 -.931 .3526
INCOME 1.46940909E-05 7.0526E-06 .085218 2.083 .0383
S7_1MC -.047883 .013895 -.113469 -3.446 .0007
AGE50 -.028581 .011574 -.103151 -2.469 .0143
CHILD .009133 .007274 .043440 1.255 .2106
HHSIZE -.480997 .275503 -.070220 -1.746 .0821
(Constant) 5.756657 1.147460 5.017 .0000
The model overall predicted 79% of the variation in the dependent variable
TOTAC and 74% of LNSUB. Overall, the explanatory power of the model seems
satisfactory. As expected, channel diversity of a cable television system
is determined by factors affecting cost of running a system and revenue
potential of the system area. Characteristics of local system area
determines about 80% variation of channel diversity. In contrast,
ownership related characteristics of the system -i.e. if the system is
owned by MSO or not and how large the MSO is- was not a significant
predictor of channel diversity. These variables were dropped from final
model. MSO's bargaining power vis-a-vis programming networks measured
by
the number of cable systems an MSO owns did not affect channel
diversity
level of local cable systems.
Whether the system is vertically integrated with a national programming
network did not make difference in the diversity at the local system
level.
It might have bigger influence in which network gets to be selected in
the system menu, but it did not influence the level of diversity. An
integrated system will be more likely to include affiliated network with
all other things equal. Whether a system is integrated with a
programming
network is highly correlated with the size of a MSO (r=.75, p=.00).
The
more systems an MSO manages, the more likely the MSO is affiliated
with a
programming network. The efficiency coming from a MSO managing a
multiple
number of systems did not improve the level of diversity at the local
system level. VI and MSONU are MSO-level characteristics rather than a
local system characteristics. Local system's level of diversity seems
to
be determined by local market and system-specific characteristics
rather
than on who owns and manages the system.
Demographic variables (except AGE50) in general did not significantly
affect the level of diversity. Diversity seems to be determined solely
by
economic condition of the system area, but not by the qualitative
characteristics of the audience the system serves.
Future growth potential of the system area was represented by projected
population growth and average retail growth rate in the original
model.
Projected population growth rate was dropped since it was not a
significant
factor, but retail growth rate was included in the final equation. Retail
sales growth rate had a positive relationship to the total number of
channels carried. If local advertising becomes a more common
industry-wide
practice, the influence of it on channel diversity might increase in the
future. As the contribution of local advertising revenue to the cable
system's total revenue gets larger, the influence of local advertising
market potential on diversity will increase.
TOTAC and LNSUB had mutually positive influence to each other. In
addition, both factors were the strongest predictors for each other. The
influence of subscriber base on channel diversity is statistically
significant (p-value of t-statistic: .00). A system has more programmed
channels (higher diversity) when it serves bigger number of
subscribers.
With approximately 3.79% increase of the number of subscribers, one
more
channel is added to the system. It means that as we move from a small
system to a larger system more and more subscribers are needed to
increase
channel diversity. After beyond a certain level of subscriber base,
the
increase of channel diversity becomes stagnant. Since this is a
simultaneous equation model, the variables such as system age, income, age
of the population, and price of basic-tier which seemingly affect
subscribership also exert influence on total number of channels programmed
indirectly through increasing or decreasing subscribership.
Also, the influence of TOTAC on subscriber size is statistically
significant and positive. The relationship holds with all other factors,
especially the total number of households in the area, held constant.
However, the coefficient is not so large. Approximately 11- channel
increase brings 1% increase in the subscribership.
Among the factors classified as market characteristics above, franchise
fee per subscriber, housing density over the plant, existence of
multichannel competitor, number of local broadcast signals, retail growth
rate, and total number of households in addition to subscriber base
were
significant factors predicting diversity. Franchise fee had positive
influence contrary to the initial expectation. It proved that franchise
fee does not add to unnecessary burden to the system operators so as
to
decrease channel diversity. Higher franchise fee might reflect higher
profit potential of the franchise area estimated at the time of
franchising
agreement (or recent renewal). As expected, higher housing density was
translated into higher channel diversity. The number of local
broadcast
signal had positive influence on channel diversity. An increase of
less
than two broadcast signals increase channel diversity by one. This
positive impact of the number of local broadcast signals on cable diversity
seems to come from two reasons. A cable operator might have an incentive
to increase channel diversity if its competition (in this case,
over-the-air broadcast signals available) is stronger and thus higher
quality service is necessary to get people to subscribe to cable. Also,
it
is likely that a cable system's channel line-up includes more broadcast
channels as there are more local broadcast channel available in the
area.
Unfortunately, we cannot separate out the two types of effect with the
data
at hand.
Contrary to the expectation, total number of households in the system area
had negative influence on the channel capacity. The coefficient is small
enough to be ignored. The number of households and the number of
subscribing households are highly correlated to each other. (r=.95 with
p=.00) There might be a multicollinearity problem in the equation.
But
each represents separate theoretical concept. The number of
subscribing
households measures the effect of current subscriber size and that of
total
households measures potential market size. It is possible that since the
effect of HHSUB on channel diversity tapers off the effect of further
increase in HHSUB with no change in diversity level is picked up by HH,
resulting in the negative coefficient.
Among the factors classified as system characteristics in the above
discussion, channel capacity and addressability made difference in
diversity. As mentioned before, the ownership related variables(VI and
MSONU) did not make difference in diversity. An increase of channel
capacity by 4 brings an addition of one programmed channel. Increase of
the proportion of addressable subscribers had positive influence on
diversity. This is so because operators might add PPV, interactive, or any
other newer kind of services. PPV can cater to narrow interests of the
audience and can be a good additional revenue source other than
monthly
subscription fee.
Factors affecting subscription level are system age, income, price of
basic service, proportion of old people in the area, vcr ownership, and
the
existence of competitive multichannel programming provider in addition to
the number of households. The effect of the first four variables was
just
as expected. System age and income level increased subscribership.
Price
and proportion of residents aged 50 and older had negative effect on
subscribership. The effect of other multichannel service in the area and
vcr penetration rate was expected as ambiguous. It was found that the
effect of competition from other multichannel service was negative, and
that of vcr ownership rate, positive on cable subscription. The
positive
influence of vcr ownership rate might reflect the overall appetite of
system area viewers for more media use.
Probably the diversity offered by the system is only one of many quality
factors considered by viewers at the moment of decision to subscribe.
Density was originally included in the equation determining
subscribership
to account for the differences between rural and urban markets, but
eventually dropped from the equation since it was insignificant. Overall
quality of operating service provided to subscribers such as
responsive
service personnel, prompt phone answering, and convenient billing,
which
was not included in the model, might explain some of the unexplained
variation.
The relationship between system size and channel diversity was positive
and mutually reinforcing. The larger the system is, the higher the
diversity level is. Higher diversity, in turn, results in higher
subscription. However, the effect of subscriber base tapered off as we
move to a larger sized systems. The effect of channel capacity was
also
strong and positive, which holds true after holding all other factors
constant.
VII. Discussion and future research
This study revealed that the same economic forces that affect content
diversity of newspapers exist for determining cable television channel
diversity. A system serving larger market in terms of subscriber base
offers higher quality service, in this case higher channel diversity, to
its subscribers. The same force works in newspaper industry; Big city
newspapers have more diverse sections than suburban papers. The same is
for international trade of motion pictures; Countries with larger
domestic
market produce and export films with more diverse subjects (Wildman &
Siwek, 1988). The media service provider with higher quality (in this
case
higher diversity), in turn, generates even bigger audience, which gives
the service provider further incentive to invest in the quality of its
service in anticipation of increased audience, popularity, and revenue.
Channel capacity puts apparent limit to the channel diversity. But
systems with larger subscriber base offered higher level of diversity even
after controlling for the channel capacity and many other variables.
Moreover, the effect of subscriber base was at least as strong as that of
channel capacity in determining channel diversity .
Technological advance which abolishes physical limit will not bring about
drastic increase in diversity; at least by itself, as is often
expected.
Economic constraints on diversity as revealed in this study will
remain the
same in the '500-channel world'. Media managers and policy makers should
bear this simple principle in mind when they prepare for the future
media
world. Regulations of multichannel media service should take the
differences among local systems and its market characteristics into consi
deration since different systems face different economic environment.
For
television audience located in disadvantaged areas to have equal
access to
the multitude of information as those living in large urban areas,
special
attention from the regulators is called for.
It seems that how we divide areas served by a multichannel service
provider matters much in determining channel diversity as it is defined in
this study. However, since this study did not extend to exploring the
composition of programming line-up in the systems, whether having a
large
service area is always better or not in terms of other aspects of
diversity
cannot be answered. Traditional policy goals such as localism might not
be better served by dividing the nation into a few large system areas.
The
goal of achieving diversity in a mass medium and that of serving unique
local needs may not necessarily go together. The issue of economic
viability vis-a-vis social desirability is an important subject to ponder
upon. Study of the determinants of system programming composition,
that is
the study of what determines the selection of particular set of networks
in a system, should be followed as a future research.
The notion of competition should be reconsidered. According to the result
of this study, competition in the same system area might not be better for
channel diversity since competing multichannel service providers are bound
to divide up the subscriber base of the given area. Implication of this
result on cable-telco competition should be studied further.
Overall, demographic variables were not so strong predictors of channel
diversity. Probably demographic variables exert stronger influence
over
which network to be selected rather than how many network to be
carried.
It should be also the subject for future research.
Appendix1: Description of key variables
Number of valid observations (listwise) = 240.00
Valid
Variable Mean Std Dev Minimum Maximum N Label
AGE1217 8.35 1.80 2.73 14.50 293 % age 12-17
AGE50 29.92 7.66 11.78 54.33 293 % age 50+
CH_CAPA 39.82 13.69 12 120 265 CHANNEL CAPACITY (FAC
CHILD 36.29 9.49 1.5283 88.4868 291
DENS .02 .01 .00 .11 292 DENSITY(LINE MILES /
EDU_HS 71.08 11.71 29.73 97.16 293 % highsch graduates
HHSIZE 2.55 .30 .53 3.72 293 Avg hh size
INCOME 27250.51 11450.98 9117 104523 291
POPULATI 11560.34 13552.29 117 69577 291
PRJ_POP 3.57 4.80 -9.40 19.80 290 PROJECTED POP GROWTH
RTLGRW 6.39 6.50 -13.48 38.11 291 AVG RTL GRWTH % OVER
s2_AGEHE 12.07 7.56 1 41 288 Age of principal head
S2_HHSUB 20891.27 49606.61 23 379461 293 # hh subscribing
S2_MSONU 330.50 445.32 2 1200 272 # sys in the MSO
S2_PARTM .93 .26 0 1 293 MSO or not
S7_TOBCT 9.25 2.97 1.00 19.00 293 Tot # bcst chans carr
S7_TOCBL 17.51 7.06 .00 39.00 293 Tot # cbl chn carried
S7_TOTAC 34.18 11.70 11 76 290
SIGNAL 10.15 5.26 1.00 26.00 290 OTA SIGNALS IN THE AD
TONGUE 8.09 13.55 .0000 89.9344 291
URBAN 35.60 40.48 .0000 100.0000 291
VCR 73.37 12.40 .00 100.0 289 VCR % BY COUNTY
WHITE 88.30 17.61 7.1754 100.0000 291
Appendix 2: Correlation matrix
Variables 2 3 4 5 6 7 8 9 10
1.AGE1217 -.31* -.23* -.14* -.17* .21* .12 -.35* .55* -.15*
2.AGE50 -.26* .39* -.14* -.38* .07 -.08 -.60* -.15*
3.TOTAC .39* .68* .08 -.44* .35* .05 .43*
4.HHSUB .30* -.03 -.32* .32* .13* .51*
5.CH_CAPA -.01 -.23 .23 .05 .32*
6. CHILD -.00 .01 .35* .13*
7. DENS -.22 -.02 -.24*
8. EDU_HS -.20* .54*
9. HHSIZE .17*
10. INCOME
11. POPULATI
12. PRJ_POP
13. RTLGRW
14. AGEHE
15. PARTM
16. SIGNAL
17. TONGUE
18. URBAN
19. VCR
20. WHITE
11 12 13 14 15 16 17 18 19 20
1.AGE1217 -.04 -.03 -.09 .12* .05 -.16* .07 -.06 -.10 -.15*
2.AGE50 -.26* -.27* -.12* .03 -.08 -.09 -.17* -.26* -.09 .22*
3.TOTAC .44* .11 .11 .03 .15* .29* .14* .52* .21* -.05
4.HHSUB .16* .01 -.07 -.09 .08 .35 .13* .42* .18 -.01
5.CH_CAPA .33* -.00 -.00 -.06 .04 .22* .08 .40* .14* -.02
6. CHILD -.02 .08 .08 -.04 .08 .04 .17* .02 .06 -.10
7. DENS -.22* .07 .09 -.15* -.18* -.20* -.19* -.39* -.12* .08
8. EDU_HS .24* .13* .07 -.07 -.00 .20* -.09 .32* .29* .27*
9. HHSIZE .11* .18* -.02 -.09 .10 .12* .34* .15* .07 -.32*
10. INCOME .36* .12* .04 -.13* .02 .41* .06 .33* .30* .19*
11. POPULATI .15* .02 -.11 -.00 .30* .17* .64* .14* -.20*
12. PRJ_POP .36* -.01 -.00 .17* .19* .09 .18* -.05
13. RTLGRW .00 .03 .02 -.05 -.08 .12* .08
14. AGEHE .09 -.12* -.08 -.03 -.05 .13
15. PARTM .01 .09 .09 -.05 -.11
16. SIGNAL .19* .31* .13* -.12*
17. TONGUE .25* -.03 -.50*
18. URBAN .08 -.14*
19. VCR .10
20. WHITE
Reference
Baer, Walter S. (1973). Cable television: A handbook for decision making.
Santa Monica,
CA: Rand.
Baer, Walter S. & Park, Rolla Edward. (1972). Financial projections for the
Dayton
Metropolitan area. In Johnson, Leland L. et al. Cable communications in
the Dayton Miami
Valley: Basic report. Santa Monica, CA: Rand.
Barnes, Beth Ellyn. (1990). Electronic media audience behavior in the
multichannel
environment: Patterns of demographic homogeneity and time spent viewing.
Evanston, IL:
Unpublished PH.D. Dissertation.
Chipty, Tasneem (1993). Horizontal integration for bargaining power: Evidence
from the
cable television industry.
Chipty, Tasneem (1994a). Vertical integration and market foreclosure: Evidence
from the
cable television industry.
Chipty, Tasneem (1994b). Vertical integration, market foreclosure, and
efficiency:
Evidence from the cable television industry.
Dennis, Roger. (Unknown). Evanston Il and cable television. Evanston, IL: Center
for Urban
Affairs, Northwestern University.
De jong, Allard Sicco & Bates, Benjamin J. (1991). Channel diversity in cable
television.
Journal of broadcasting & Electronic media, 35 (2), pp159-166.
Dertouzos, James N. & Thorpe, Kenneth E. (1982). Newspaper groups: Economics of
scale,
tax laws and merger incentives. Santa Monica, CA: Rand.
Dertuozos, James N. & Wildman, Steven S. (1990). Competitive effects of
broadcast signals
on cable. Attachment to comments of NCTA in FCC filing: MMDocket
Nos. 89-600 and 90-4.
Dertuozos, James N. & Wildman, Steven S. (1993). Regulatory benchmarks for cable
rates:
Review of the FCC methodology. June 21, 1993.
Dominick, Joseph, Sherman, Barry L., & Copeland, Gary (1990).
Broadcasting/cable and
beyond: An introduction to modern electronic media. NY, NY:
McGraw-Hill.
Eastman, Susan Tyler. (1989). Cable system programming. In Eastman, S. T., Head,
Sydney
W., & Klein, Lewis (Eds.) Broadcast/cable programming: Strategies and
practices. Belmont,
CA: Wadsworth.
Hazlett, Thomas, W. (1994). Regulating cable television rates: An economic
analysis. A
working paper. Institute of governmental affairs. University of
California, Davis.
Lee, Varnell Mcafee (1993). The growth and development of basic cable services
and a case
study of TNN. Dissertation. Ohio University.
Levin, Harvey J. (1971). Program duplication, diversity, and effective viewer
choices:
Some empirical findings. American Economic Review, 61 (2), pp
81-88.
Myers reports: 9th annual survey of cable operator Executives on basic
networks. (1994).
Parsippany, NJ: Myers Communications Inc.
Nadel, Mark S. (1985). Comment: Multichannel video competition. In Noam, E.
(Ed.) Video
Media Competition. NY, NY: Columbia University Press.
"New network listings update" Cablevision Aug 8, 1994, p57-58.
Noam, Eli M. (1985). Economies of scale in cable television. In Noam, E. (Ed.)
Video
Media Competition. NY, NY: Columbia University Press.
Owen, bruce M. & Wildman Steven S. (1985). Program competition, diversity, and
multichannel bundling in the new video industry. In Noam, E. (Ed.) Video
Media
Competition. NY, NY: Columbia University Press.
Pacey, Patricia L. (1985). Cable television in a less regulated market.
Journal of
Industrial Economics, 34(1), pp81-91.
Park, Rolla Edward. (1971). Prospects for cable in the 100 largest television
markets.
Santa Monica, CA: Rand.
Rubinovitz, Robert. (1991). Market power and price increases for basic cable
service since
deregulation. Washington, DC: Economic Analysis Group, Antitrust
division, US department
of Justice.
Setzer, Florence & Levy, Jonathan (1991). Broadcast television in a multichannel
marketplace. OPP working paper series. Washington, DC: Office of Plans
and Policy, FCC.
Sloan Commission. (1971). On the cable: The television of abundance. NY, NY:
McGraw-Hill.
Smith, William. (1989). Freeze frame. Marketing & Media Decisions. 24(4),April,
p26
Solomon, Harvey (1989). Refranchising: Cities fight back. Channels, 9(4),
p46-49.
Thorpe, Kenneth. (1985). Cable television, market power, and regulation. Santa
monica,
CA: Rand corp.
Vita, Michael, G. & Wiegand, John, P. (1993). Must-carry regulations for cable
television
systems: An economic policy analysis. Journal of Broadcasting &
Electronic Media, 37(1),
1-19.
Vogel, Harold, L. (1990). Entertainment industry economics: A guide for
financial
analysis. Cambridge, UK: Cambridge University Press.
Waterman, David & Weiss, Andrew A. (1993). Vertical integration, program
access, and the
1992 cable television act. Paper presented at ICA, Sydney,
Australia, July 11-15, 1994
Webb, G. Kent. (1983). The economics of cable television. Lexington, MA:
Lexington books.
Wildman, Steven S. (1994). The economics of audiencemaking. In Ettema, James S.
&
Whitney, D. Charles (Eds.) Audiencemaking: How the media create the
audience. Thousand
Oaks, CA: Sage.
Wildman, Steven S. & Siwek, S. E. (1988). International trade in films and
television
programs Cambridge, MA: Ballinger publishing.
Woodard, Jr., Charles, C. (1974). Cable television: Acquisition and operation
of CATV
systems. NY, NY: McGraw-Hill.
Zupan, Mark A. (1989). The efficiency of franchise bidding schemes in the case
of cable
television: Some systematic evidence. Journal of law and economics,
32(2), p401-456.
Zupan, Mark A. (1989). Non-price concessions and the effect of franchise
bidding schemes
on cable company costs. Applied economics, 21, pp305-323.
Student Paper
An Integrated Model for Determining Channel
Diversity
in a Multichannel Environment
Running head: Channel diversity
April 1, 1995
Submitted to Media Management & Economics Division
AEJMC, 1995 Annual convention
By Eun-mee Kim
Department of Communication Studies, Northwestern University
Rm 12, Harris Hall
1881 Sheridan Rd
Evanston, IL 60208
(Tel) 708-332-1854
E-mail: [log in to unmask]
|
