FTC Staff Comment Before the FCC In the Matter of Competition, Rate Deregulation and the Commissions Policies Relating to the Provision of Cable Television Services

COMMISSION AUTHORIZED
BEFORE THE
FEDERAL COMMUNICATIONS COMMISSION
WASHINGTON, D.C. 	20554
ID The Matter or 	)
)
CompetltloD, Rate DerelulatloD aDd the 	)
Commluloa's Policies RelatiDI to the 	)
Proylsloa or Cable TelnlsloD Senlce 	)
MM Docket No. 89-600
CommeDt or the Starr or
the Bureau or EcoDomlcs aDd the SaD FraDcisco ,!eeional Office
or the Federal Trade -Commission
April 20, 1990
• This comment represents the views of the staff of the Bureau of Economics
and the San Francisco Regional Office of the Federal Trade Commission. They
are 	not 	necessarily 	the 	views 	of 	the 	Commission 	or 	any 	individual
Commissioner. Inquiries regarding this comment should be directed to Michael
G. Vita (202-326-3493) of the FTC's Bureau of Economics or John Wiegand (415-
744-7920) of the FTC's San Francisco Regional Office.

Table of Contents
Executive Summary 	III
I. 	In trod uction
II. 	Expertise of the Staff of the Federal Trade Commission 	3
III. 	Restrictions on Entry. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 	4
A. Introduction .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 	4
B. The Likely Competitive Impact of Free Entry into Cable
Markets 	5
C. Alternatives to Cable: Factors Affecting the Likelihood
of Entry 	i • • • • • • • • • • • • •• 	13
D. Conclusion 	13
IV. 	Product Market Definition 	15
A. Conceptual Issues. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 	15
I. Identifying the Relevant Antitrust Market 	16
2. Computing Market Shares in the Relevant Alltitmst
Market. .. .. . . . . . .. . . . .. . . . . . . . . . . . . . . . . . . . .. 	17
3. Applying the Guidelines to Cable Markets: 	An
Illustration 	. . . . . . . . . . . . . .. 	18
4. The Impact of the Regulatory Structure on Product
Market Definition 	21
B. Empirical Approaches to Market Definition. . . . . . . . . .. 	25
C. Conclusion 	28
V. Regulatory Options ... . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 	29
. 	A. Price-Caps Versus Rate-of-Return Regulation: . . . . . . .. 	29
B. Regulation Through Franchise Bidding 	:....... 	33
VI. "Trafficking" in Cable Systems 	!............. 	36
VII. Vertical Integration and Other Vertical Restraints 	43
A. 	Economic 	Theories 	of 	Vertical 	Integration 	and
Control. . .• . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 	43
1. Incentives for Programmers to Adopt Restricted
Distribution 	45
a. Inducement of Marketing and Promotion
Efforts. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 	46
b. Constraining the Exercise of Market Power
by Cable Operators 	48
c. Vertical Restrailrts io Avoid Distortions From
Upstream Market Power 	52
d. Vertical Integration as a Source of Entry
Barriers - , 	56
.."--: i: .'iJiiie~~~ii~i .. ;~ie~' .~i ..
efficiency 	57
ii. Different minimum efficient
scales 	59
iii. When Added Risk Raises the
Cost of Capital 	60
e. Vertical Illlegration to Avoid Rate-of-Return
Regulation 	61
B. Empirical Evidence on Vertical Integration in the Video
Marketplace. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 	62

I. The Salinger SLudy 	63
2. The Klein Sllld.v . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 	65
3. How Should These Findings Be IlILerpreLed? 	68
C. Summary 	; 	71
VIII. 	The Application of the Antitrust Laws to Cable Television 	72
References. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 	77
'- .
ii

Executive Summary
The comment of the staffs of the FTC's Bureau of Economics and the
San Francisco Regional Office addresses a number of i~sues raised by the FCC's
Notice of Inquiry (NOI) on competition in the provision of cable television
services. 	The comment addresses six issues relating to economic efficiency,
competition, and the enforcement of the antitrust laws, and does not discuss
other policy considerations raised in the NOI or in statements submitted by
respondents to the NOI.
First, 	the 	comment 	discusses 	the 	desirability 	of 	permitting 	direct
competition between cable television systems, and the likelihood that such
competition will occur. The FTC staff believes that under most circumstances
such competition should be permitted. Restricting entry into cable markets can
be justified only when the local cable market is a "nonsustainable" natural
monopoly. Moreover, fulfilling this condition does not necessarily ensure that
consumers' interests would be served by insulating an incumbent from the
threat of entry. Though the FTC staff does not claim that entry (or the threat
of entry) would invariably impose an effective competitive constraint upon the
market power of cable systems, the interests of consumers would seldom be
served by erecting statutory or regulatory barriers to entry.
Second, 	the 	comment 	presents 	a 	framework 	for 	examining 	the
_appropriate product market for cable television and its competition. While the
staff cannot-provide a conclusive product market definition for all geographic
markets, we believe that the 1984 Department of Justice Merger Guidelines and
the FTC Merger Statement articulate a useful method for addressing this issue.
The comment provides an illustration of how this method might be applied to
the analysis of cable television market power. The comment also explains how
recent developments in applied economic analysis might be used to quantify
empirically the extent of any existing market power.

Third, the comment discusses available regulatory options if a decision
is made to re-regulate cable systems. There are three general alternatives. The
first 	is 	tradi tional 	lOra te-of -return" 	regula tion; 	the 	second 	is 	"price-ca p
regulation"; and the third is "franchise bidding." 	The potential and actual
problems associated with rate-of -return regulation have been well-documented.
There are theoretical and empirical bases for preferring price cap to rate-of·
return regulation. 	However, the regulatory environment that existed prior to
the enactment of the Cable Act, as well as that which exists today, permits and
encourages the reconfiguration of cable service tiers so as to avoid rate controls
on basic service. 	If it is determined that some form of explicit rate regulation
is 	necessary 	to constrain the exercise of market 	power by cable systems,
effective regulation would seemingly require that regulatory authorities be
empowered either to control the rates charged for all service tiers, or to impose
"minimum quality" controls along with rate controls (e.g .. stipulate that "basic"
service offer certain specified channels). 	Absent such regulatory authority, it
is unclear that cable rate regulation could effectively constrain the exercise of
cable systems' market power should such power be found to exist.
Fourth, the comment discusses "trafficking" (i.e., the sale and rapid
resale) in cable systems. 	The staff believes that "anti trafficking" regulations
are unlikely to help, and may hurt, consumers. 	When a cable system changes
ownership, the new owner may be as efficient, less efficient. or more efficient
than the old owner. 	If equally efficient, the new owner's profit-maximizing
levels of price, output, and quality will not change. If less efficient, the buyer
will unlikely earn a competitive return on the investment during its ownership
tenure, and will likely suffer a capital loss upon resale if its low quality jhigh
price causes a loss of goodwill. 	Transfers of cable systems from more to -less
efficient owners will, therefore, tend to be deterred by profit incentives, and
thus may not require special regulations.
iv

Antitrafficking regulations could, moreover, be counterproductive. To
the extent that such regulations delay the transfer of cable systems from less
to more efficient operators, consumer welfare will be reduced. Antitrafficking
regulations could also deter entry into cable markets if they make it harder to
recoup investment costs through exit (e.g .. by selling to a more efficient buyer)
if entry proves unprofitable.
Fifth, the comment reviews the theoretical and empirical literature on
vertical 	integration and control, especially as 	it 	relates 	to cable systems.
Currently there is considerable debate over the extent to which cable systems
and programmers use vertical controls, as well as over the implications of these
controls 	for 	competition 	and 	consumer 	welfare. 	It 	appears 	that 	some
programmers, whether vertically integrated into cable distribution or not, have
chosen to restrict the carriage of their signals to certain selected distributors.
Similarly, certain cable systems that are integrated into programming appear to
exhibit a preference for their own programs. 	While these policies may harm
particular programmers and particular distributors, it does not necessariJy
follow that competition or consumer welfare is harmed. 	For example, vertical
integration and other vertical controls conceivably may improve efficiency if
they preserve or increase programmers' incentives to increase output. 	Vertical
controls can help accomplish this in some circumstances by (I) facilitating the
efficient marketing of programming services, (2) preventing the erosion of
program value through poor signal quality, (3) preventing the piracy of program
signals, and (4) eliminating "successive mark-up" problems. Vertical integration
and other vertical controls are not necessarily procompetitive, however, and
there are circumstances under which vertical controls could be used to erect
barriers to entry and to evade regulation.
Neither the theoretical literature on vertical integration and control, nor
the 	empirical 	studies 	of 	vertical 	integration 	between 	cable 	systems 	and
v

programmers, yields unequivocal conclusions about the net competitive impact
of vertical controls on programming and cable markets. Therefore. one cannot
infer from available information whether consumers' interests would be served
by creating regulatory or statutory standards that are hostile to the continued
use of vertical controls.
The final section of the comment observes that cable systems and
program suppliers are subject to the provisions of the existing antitrust laws,
and 	that 	both 	the Sherman Act and 	the Clayton 	Act 	have 	been 	used 	to
adjudicate the legality of a variety of practices, including mergers between
competing cable systems, and the misuse of the franchising process to impede
entry by overbuilders.
vi

BEFORE THE
FEDERAL COMMUNICATIONS COMMISSION
WASHINGTON, D.C. 	20554
In The MaUer of 	)
)
)
COlDpetitlon, Rate Dereaulation and the 	)
COlDlDission's Policies Relatlna to the 	)
Provision of Cable Television Service 	)
MM Docket No. 89-600
Comment of the Starr of
the Bureau of Economics and the San Francisco Reaional Office
of the Federal Trade Commission I
April 20, 1990
I. 	Introduction
The staff of the Federal Trade Commission appreciates this opportunity
to submit a comment in response to the Federal Communication Commission's
Notice of Inqui ry 2 concerning competition, rate deregulation and the FCC's
policies relatinog to the provision of cable television service. 	The Notice of
Inquiry (NOI) initiates the FCC's comprehensive inquiry into the effects of the
Cable Communication Policy Act of 1984 3 on the video services marketplace.
Our comment addresses issues relating only to economic efficiency, competition,
and the enforcement of the antitrust laws. 	It does not discuss other policy
considerations raised in the Notice.
This comment addresses six of the issues raised in the NOI. First, Section
III discusses a number of issues relating to the likelihood and desirability of
direct competition between cable television systems. Second, Section IV presents
I This comment represents the views of the staff of the Bureau of Economics
and the San Francisco Regional Office of the Federal Trade Commission. They
are 	not 	necessarily 	the 	views 	of 	the 	Commission 	or 	any 	individual
Commissioner. Inquiries regarding this comment should be directed to Michael
G. Vita (202-326-3493) of the FTC's Bureau of Economics or John Wiegand (415-
744-7920) of the FTC's San Francisco Regional Office.
2 55 Fed. Reg. 1484-86 (Jan. 16, 1990).
347 U.S.c. § § 521 - 559.

a framework for determining the appropriate product market for analyzing
cable 	television 	and 	its 	competition. 	Third, 	Section 	V discusses 	various
regulatory options should a decision be made to re-regulate cable systems.
Fourth, Section VI discusses the issue of "trafficking" in cable systems. Section
VII reviews the theoretical and empirical literature on vertical integration,
especially as it relates to cable. 	Section VIII discusses the applicability of the
antitrust laws to cable television.
2

II. 	Expertise of the Staff of the Federal Trade Commission
The 	FTC 	is 	an 	independent 	regulatory 	agency 	responsible 	ror
maintaining competition and safeguarding the interests of consumcrs. 4 	In
response to requests by federal, state, and local government bodies, the staff of
the FTC often analyzes regulatory or legislative proposals that may affect
competition or the efficiency of the economy. 	In the course of this work, as
well as in antitrust and consumer protection research, nonpublic investigations,
and litigation, the staff applies established principles and recent developments
in economic theory to competition and consumer protecti~n issues, including
efficiency rationales for rate and entry regulation. 5
The FTC staff previously has commented on various issues before the
FCC, including: (I) elimination of the prohibition on common ownership of
cable television systems and national television networks;6 (2) rules relating to
whether cable television systems "must carry" television broadcast signals;7 and
(3) the FCC requirement that broadcast licenses be held for at least three years
before being transferred. 8
• 15 US.C. § § 41 - 59.
5 See, e.g .. Mathios and Rogers, The Impact 0/ State Price and Entry Regulation
on Intrastate Long Distance Telephone Rates, Bureau of Economics Staff Report
to the Federal Trade Commission, November 1988.
6 CT Docket No. 82-434.
7 MM Docket No. 85-349.
8 BC Docket No. 81-897.
3

III. 	Restrictions on Entry
A. Introduction
The 	NOI 	raises a 	number of questions 	about 	the 	extent 	of direct
competition between cable companies, and the likelihood that future direct
competition could limit any market power that might be associated with the
existence of a local cable monopoly.g In this section we discuss the theoretical
literature and 	empirical 	evidence 	that 	is 	available 	to 	help 	answer 	these
questions. 	We also briefly discuss alternative forms of video distribution, and
the barriers to entry that these alternative distributors might face.
Direct competition between cable systems for the same set of subscribers
has been, and remains, a relatively rare phenomenon. 	Of the more than 9,000
cable franchises currently in existence, less than one-half of one percent face
any direct competition for subscribers from other cable companies. 10 There are
some indications that the number of "overbuilds" has recently increased,l1 but
overbuilding nonetheless remains relatively rare.
The paucity of cable overbuilds should not be surprising if, as is widely
believed, cable distribution is a natural monopoly. 	In virtually all discussions
of the properties of cable technology it is assumed that the cost of serving a
given geographic area is lowest when it is served by a single firm. 	The reason
is 	that 	single 	firm 	production 	avoids 	duplicate 	investments 	in 	receiving
equipment (i.e .. the "headend") and in the cable itself. 	Although this is a
g See § I.A.7, and § C, of the NOI. Also see § § IV and VIII of this comment
for additional discussions of the- cable franchising process.
10 Direct competition between cable operators occurs when "two cable
operators acquire overlapping municipal franchises, install cable on the same
streets in some portion of the municipality, and compete to serve the same
households." See Smiley (1986), p. 1. This is known as "overbuilding."
11 The October 31, 1989 issue of Cable TV Franchising lists 78 pending
overbuilds, and 21 actual overbuilds. Cable TV Franchising estimates that there
are 41 percent more overbuilds now than one year ago.
4

reasonable argument,12 only a few econometric studies of cable television costs
exist. IS 	These studies tend to confirm, however, the hypothesized natural
monopoly properties of cable television.
Although cable television may be a natural monopoly, it does not follow
necessarily 	(I) 	that 	incumbent cable 	providers should 	enjoy 	statutory 	or
regulatory protection from entry, or (2) that competition from overbuilding
could never be profitabie and beneficial to consumers. 	As we discuss further
below, while eliminating regulatory entry barriers will not necessarily ensure
competitive behavior in every cable market,14 it seems unlikely that there would
be significant consumer injury from adopting such a policy. We think it likely
that consumers in some cable markets could benefit from a policy of free entry,
and that only in rare instances would consumers be harmed by such a policy.
B. The Likely Competlthe Impact of Free Eotry ioto Cable Markets
Recent 	work in 	the 	theory of natural 	monopoly demonstrates 	that
monopoly is the least-cost industry structure when the firm's cost function has
.
the property of "subadditivity.,,15 Costs arc said to be subadditive at some level
of output if one firm can always produce that output at lower cost than can two
or more firms. 	For example, subadditivity would exist if the total cost of
12 The economics of cable television would not seem to differ all that much
from the economics of electricity and gas distribution, which also are typically
assumed to be natural monopolies. One possible difference is that two different
cable systems could potentially provide significantly differentiated products,
while gas and electric companies could not.
IS See, e.g .• Owen and Greenhalgh (1986).
14 Economic theory, as well as the rarity of overbuilds, suggests that direct
competition is not likely to occur very often. 	There are some circumstances,
however, where two cable systems could survive in equilibrium. 	See Smiley
(1986), and our discussion below, for a more detailed analysis.
15 This theory is discussed in Braeutigam (1989), Baumol et al. (1982), and
Sharkey (1982).
5

serving 50.000 cable subscribers within a given geographic area is lower when
one firm serves all 50,000 subscribers than if two firms each serve 25,000.
It is important to distinguish subadditivity from the concept of "scale
economies" familiar to those active in the field of public utility regulation. 16
Economies of scale imply that long term per unit cost declines as output
increases. The existence of scale economics throughout a given range of output
levels implies that costs are subadditive for that range of output levels. and,
therefore, that the industry is a natural monopoly for that range of output
levels. l1 However. monopoly sometimes may be the least-cost industry structure
even when scale economies do not exist over the entire range of possible output!
levels. In other words, while scale economies imply subadditivity, the converse
does not hold.
When an 	industry 	is a 	natural 	monopoly, it 	has traditionally 	been
assumed that consumers will be better-off if there is only one producer, since
the existence of more than one producer will raise total costs, leading to a
higher price to cover the higher costs. 	The higher price, furthermore, would
cause consumers to purchase fewer units of output than if price were lower.
The 	purpose 	of 	traditional 	public 	utility 	regulation 	(e.g.. 	rate-of -return
regulation) has thus been to require the monopolist to sell more units of output,
at a 	lower price, than it would sell 	if it were free. to choose the 	profit-
maximizing price and output. When carried out successfully, traditional forms
16 The importance of this distinction will become more apparent below.
17 See Sharkey (1982), pp. 86-90. Strictly speaking, this statement is correct
only if the firm produces one good. 	If the firm is a multiproduct natural
monopoly, the conditions for subadditivity become more complex. See Sharkey
(1982), PP. 62-73.
6

of monopoly regulation distribute the benefits of monopoly production 	to
consumers. 18
Consumers might not benefit from monopoly production, however, even
in a natural monopoly market. if regulatory controls are poorly-designed or
nonexistent. and the monopolist's market power is otherwise not constrained.
If the monopolist is able to keep the gains froin monopoly production for itself
(because its pricing power is unconstrained by regulation or the threat of
competition). then consumers may be worse off under monopoly than if more
than one firm were permitted to engage in production. Even though this would
entail higher production costs. price might be reduced (and output increased)
because there would now be some competition where none existed previously.
Entry (or the threat of entry) could induce an incumbent monopolist to
keep prices lower than it otherwise might. When production is characterized by
scale economies (as defined above), the 	threat of entry can, 	under 	some
circumstances. induce the monopolist to set the market price equal to its average
cost in order to deter entry. 	If it does this. no firm that enters on a smatter
scale can survive, as the entrant's unit cost will always exceed this price. This
is 	an 	example 	of 	what 	is 	known 	as 	a 	sustainable 	natural 	monopoly.
Sustainability means that market demand and cost conditions make available
to the monopolist a price such that (l) the market clears (i.e .. the monopolist
produces all that is demanded at that price). (2) the monopolist (at least) breaks
even, and (3) entry is unprofitable.l~ 	While sustainability does not necessarily
ensure that the monopolist will set a price that results in maximum consumer
18 How well this has succeeded is, of course, the source of considerable
debate. We briefly discuss the problems with rate-of-return regulation in § Y,
below.
l~ See Baumol el al. (1982). pp. 192-3.
7

benefits. it does mean that it can deter inefficient entry.20 	Thus. when the
monopoly is sustainable. there is no compelling rationale for providing the
. 	b'h 	. 	f 	t 	Z1	Incum ent WIt 	statutory protectIOn 	rom en ry.
The case for legal protection from entry is more compelling when costs
are subadditive. but when scale economies do not exist over the entire range of
possible production levels. 	A well-known example occurs when the firm's
average costs have a "U"-shape; i.e .. costs fall until they reach some minimum
point. and then rise. 22 If dema~d intersects this average cost curve to the right
of its minimum point (but not far enough to its right to accommodate two firms
producing at minimum average cost), costs will be subadditive; however, the
monopoly will be unsustainable. 	This monopolist would not be able to charge
a price that both covers its costs and deters entry. 	Inefficient entry might
occur, resulting in higher costs (and perhaps price). 	In'this case, it is possible
that statutory restrictions on entry could benefit consumers.
It would be a mistake. however, to bar entry simply because of the
theoretical possibility of nonsustainability. First. it seems reasonable that those
who 	seek 	to 	bar 	entry 	should 	bear 	the 	burden 	of 	establishing 	that
nonsustainability is likely to be a significant problem. 	As we stated earlier,
20 Cable franchises may sometimes be required to charge a uniform price
throughout a cable franchise area 	even 	if 	the costs of serving different
subscriber groups differ. In such instances, the uniform price means that some
customers will be charged a price that exceeds the cost of serving them. while
others will be charged a price below the cost of serving them. This structure of
prices invites inefficient "creamskimming" entry by firms whose aim is to serve
only the first group of customers. These effQrts to enter might not be attempted
if prices more closely reflected costs.
21 As we discuss below, a natural monopoly will not always be sustainable
against entry. Even when a natural monopoly is sustainable. the firm may have
available to it several different sustainable prices. 	Not all of these prices will
result 	in 	maximum 	consumer 	benefits. 	However. 	when 	a 	monopoly 	is
sustainable, it will generally be true that the price that generates maximum
consumer benefits will be a sustainable price. 	See Baumol et aI., pp. 208 ff.
22 	This example was first presented in Baumol et al. (1982). pp. 29-32,
Sharkey (1982). pp. 88-89, and is discussed in greater detail in Brock and Evans
(1983). pp. 69-76.
8

the limited evidence on cable costs suggests that cable systems experience scale
economies (i.e .. decreasing unit costs) over the observed ranges of output. 	By
itself, this implies sustainability. Moreover, to our knowledge, no studies exist
that suggest that cable systems are subject to nonsustainability.23 	There is,
therefore, no obvious reason why 	one should 	presume 	the existence of a
sustainability problem.
Second, as Baumol et ai. argue,24 the case for entry restrictions outlined
above holds with full force only when there are no sunk costs associated with
entry into a market. 25 Otherwise, as they point out, the nonsustainability of the
natural monopoly may only mean that the incumbent cannot fully exploit the
advantages conferred on it (relative to the entrant) by the existence of the sunk
costs. The incumbent may still be able to set a price that allows it to cover its
costs and deter entry; this price may well be substantially below the price the
monopolist would charge if it enjoyed regulatory protection from entry.
This point can be better explained with the aid of an example taken
from Baumol et ai. (1982, pp. 221·2). 	In figure 	I, ACm and 	ACe are 'the
respective average costs of the incumbent and cntrant, and 0 	is the market
demand curve. The market-clearing price that allows the monopolist to exploit
fully its cost- advantage (relative to the entrant) is Pm' 	This price is not
sustainable. 	However, a somewhat lower price, such as Pe, is sustainable. 	At
this price, the monopolist could still satisfy total market demand and cover its
2S We do not argue that 	the 	nature of cable cost functions 	has 	been
definitively resolved. 	Our understanding would be increased if cable systems
were modelled as multiple output producers (e.g., as producers of basic and pay
services). In a multi-output context, analysis of subadditivity and sustainability
is more complicated. 	It is still true, however, that the issue 	of sustainability
does not arise unless the conditions for subadditivity are satisfied. Empirically
testing for subadditivity is more difficult in a multiproduct setting, but is still
tractable; see Evans and Heckman (1984) for an example.
24 See Ba umol et al. (1982), pp. 221·4.
25 Sunk costs are costs that cannot be recovered if a firm elects to exit a
market.
9

costs; the entrant, however, would not be able to cover its costs. 	This price is
$/Q
pi
Pm
Pe
a Q
ACm
Fig ure 	1
well below the unconstrained monopoly price, P •.
Baumol et al.'s argument would seem to carry considerable weight when
applied 	to 	the 	cable 	television 	industry. 	Entry 	into 	cable 	markets 	is
characterized by substantial sunk costs. 	The cost of the cable, for example, is
sunk;H once installed. an entrant could Dot profitably remove the cable for use
in another geographic market if entry into the first market were uDsuccessfu1. 21
26 In many cases. literally as well as figuratively.
21 By comparison. some of the other equipment (e.g .• the receiving "dish")
might be transferrable.
10

Thus, even 	if 	cable systems 	in 	some 	markets 	are 	nonsustainable 	natural
monopolies, it may make little sense to provide these systems with statutory
protection from entry, since the existence of substantial sunk costs may provide
them with the ability to deter inefficient entry.
Although adopting a policy of free entry would appear to entail few
risks, it does not follow that there necessarily would be a great many instances
wh~re two or more cable franchises could persist in equilibrium. 	Theory, and
the historical infrequency of overbuilds, suggest otherwise.28 	However, there
is some theoretical analysis suggesting that substantial consumer benefits could
be obtained when simultaneous, rather than sequential, entry i'nto a geographic
market can be made possible.2Q 	That is, sustained competition may be more
likely to occur when two firms enter 	the same 	market at the same time
(·simultaneous· entry), than when one firm enters first, with the other following
at a later date (·sequential· entry). 	This may have important implications for
areas that currently do not have cable service.3O Smiley (1986) constructed.a
simulation 	model of cable competition 	in 	which (among other 	things) 	he
compared the total welfare (profits plus consumer surplus) that would be
generated under monopoly and duopoly, the latter with simultaneous entry and
differentiated-products. These comparisons were performed under a variety of
Bit is not clear, of course, whether the historical infrequency of overbuilds
reflects a reluctance on the part of municipalities to grant a second cable
franchise, or the absence of profitable entry opportunities.
~ In Smiley's (1986) model (discuSsed below), sequential entry seldom occurs
because, as one would expect, the incumbent's scale advantages allow it to set
a price that covers its own costs while imposing losses on the entrant. 	Because
the potential entrant knows that it cannot cover its costs, entry does not occur.
30 According to the NOI (paragraph 41), it is estimated that about one-fifth
of all television households are not passed by existing cable systems. 	Smiley's
analysis of the benefits of sequential entry would be applicable principally to
these households, but could, also have relevance in other settings (e.g., as-yet
unbuilt housing developments).
11

assumptions about 	cost Sl and 	demand. 	Smiley 	found 	that 	in 	most 	cases
(including his benchmark case), total welfare and consumer surplus were higher
under duopoly than under monopoly. 	That is, even though costs were higher
under duopoly, consumer welfare was improved relative to monopoly because
of the greater degree of price co~petition.S2
C. Alternatives to Cable: 	Factors Arcectlol the Likelihood of Entry
Sections II.C.2 - II.C.4 of the NOI ask a number of questions about
factors that might affect the competitiveness of alternatives to cable television.
These alternatives comprise, among other things, di'rect broadcast services (DBS),
and multichannel, multipoint distribution systems (MMDS)." 	Because these
technologies have not yet been employed with great frequency, it is somewhat
difficult to assess their viability. 	DDS, in particular, appears to be very much
in its developmental stage, so it is hazardous to speculate about its prospects for
success. 	We also note that DDS technology, as described 	in .the 	NOI (see
paragraphs 27·28), may nave the cost characteristicslhat are typically associated
with natural monopoly: namely, high fixed costs, and low marginal costs (per
subscriber). In the long-run, it is conceivable that local cable television systems
will be simply supplanted by a very small number (perhaps one) of DDS
providers, with '''intermodal'' competition between cable television and DDS
occurring only during a transitional stage. Thus, the ascent of DDS would not
31 In each case, however, Smiley assumed econo~ies of scale. 	A duopoly
equilibrium can occur in Smiley's model because there is complete symmetry
between firms in terms 'of cost and demand, and because neither firm is given
a "first mover" advantage (i.e., the ability to enter the market first and gain the
advantages of scale economies).
S2 Even where total welfare fell under duopoly it is possible that consumer
surplus increased. 	Smiley did not report the breakdown between profits and
consumer surplus for all of his simulations, so it is not possible to tell if this.
occurred.
" 	MMDS is 'sometimes referred to as "wireless cable."
12

necessarily ensure the long-term existence of competition in the distribution of
video signals.
As we observe on pages 20-21, below, the competitive constraint imposed
on a cable system by some of these alternative technologies could well differ in
different geographic settings. 	MMDS, for example, is likely to be a poor
substitute for a cable system when there are physical obstructions (e.g .• hills and
. tall buildings) that would block the former's signal. 	Entry by an MMDS might
also be affected by regulatory factors.M 	Moreover, because MMDS has a
relatively low channel capacity, its attractiveness to consumers as a substitute
for cable will depend upon the Quality (i.e .. number and type of channels)
offered by the incumbent cable system.
Much 	of 	the 	discussion 	about 	the 	likely 	success 	of 	alternative
distribution 	technologies 	has 	centered 	on 	Questions 	regarding 	"access 	to
programming." 	This is discussed in considerable detail in § VII (below).
D. Cooc:lusloo
Restricting entry into cable markets can be justified only under a very
restrictive set of conditions; namely, when ~he local cable market is a natural
monopoly, and when this monopoly is not sustainable. Moreover, these criteria
are best thought of as necessary, but not sufficient, conditions for insulating an
incumbent from the threat of entry. 	When entry requires the creation of sunk
. assets, the incumbent might still be able to deter inefficient entry and cover its
costs even 	when 	the conditions for sustainability are 	not satisfied. 	The
resulting price may be well below that which would prevail if entry were
restricted.
As an empirical matter, it is difficult to say whether entry (or the threat
of entry) would pose an effective constraint on the market power of cable
M We lack data on the impact of regulatory factors on entry by MMDS's.
13

operators, or whether additional constraints, such as price regulation, would be
necessary for this purpose. 35 The principal message of this section is that there
is likely to be little merit in discouraging entry; in some situations, such as a
currently unserved community, encouraging simultaneous entry might result in
lower equilibrium prices than would occur if only a single franchise were
granted.
35 We address these regulatory options in § V, below.
14

IV. 	Product Market Definition
A. Conceptual Issues
Paragraph 	19 of the 	NOt asks for 	information on 	the appropriate
product market for assessing the local market power of cable operators. 	While
we cannot provide a conclusive product market definition applicable to all
geographic markets, we believe that the FTC Merger Sialemenl 38 and the 1984
Department of Justice Merger Guidelines 31 articulate a 	useful method 	for
addressing this issue. Further. recent developments in applied economic analysis
may suggest means by which any existing market power could be formally
Qua n ti fied.
Identifying the possible existence of market pewer under the Guidelines
approach can be thought of (approximately) as a three-stage procedure. 	Step
one consists of identifying a relevant antitrust market. 	Step two consists of
computing market shares in this market. Step three consists of evaluating the
conditions of entry into this market and factors affecting the likelihood of
successful cOllusion. 38 	Generally speaking, unless concentration in a relevant
antitrust market is high. and entry difficult. there will be few opport~nities for
the e~ercise of market power.
38 See FTC Merger Statement § 2.11 June 14, 1982, reprinled in 4 Trade Reg.
Rep. (CCH) para. 13.200.
31 See U.S. Department of Justice Merger Guidelines June 14, 1984, reprinled
in 4 Trade Reg. Rep. (CCH) para. 13,103.
38 Conditions of entry are discussed in § III of these comments; hence, we
do not address this topic further.
IS

1. Identifying the Relevant Antitrust Market
The Guidelines define the relevant product market as the smallest "group
of products such that a hypothetical firm that was the only present and future
seller of those products could profitably impose a 'small but significant and
nontransitory' increase in price."3G The Guidelines generally define a "small but
significant and nontransitory" increase in price as "a price increase of five
percent lasting one year."
The goods or services that make up the relevant product market are
determined by the willingness of consumers to switch to alternative products
and services in response to a price increase..tO 	That is, if an attempt by a
monopolist producer of product A to raise the price of A41 by a small but
significant and nontransitory amount caused so much switching to product B
that the increase in the price of A was unprofitable, the relevant product
market would include at least A and B. It would then be asked if a hypothetical
increase in the prices of A and B would cause so much consumer switching to
product C that the attempted price increase would again be unprofitable. If the
answer is no, A and B make up the relevant product m!lrket. 	Otherwise, C is
added to the market, and the profitability of an attempted price increase of A,
B, and C is evaluated. This procedure is continued until the smallest group of
39 See U.S. Department of Justice Merger Guidelines Issued June 14, 1984,
4 Trade Reg. Rep. (CCH) para. 13,103, § 2.11.
40 See Werden (1983), p. 524-5. 	The issue of supply-side substitution comes
into play when market shares in the relevant market are computed.
41 The Merger Guidelines (§ 	2.0) define the relevant market using the
"prevailing or likely future" price of A as the reference point for assessing the
profitability of a price increase. This price could, however, reflect the exercise
of existing market power by the producer of A. Since the FCC's Objective is to
identify any existing market power held by the producer of A (rather than an
increase in market power that might arise because of a merger), it should use
the competitive price, not the prevailing price, as the benchmark for 	this
procedure.
16

products is found for which a hypothetical monopolist would find it profitable
to establish and maintain a significant and nontransitory price increase.
2. Computing Market Shares in the Relevant Antitrust Market
The ability of the firms in a properly defined market to raise price,
either through 	the market dominance of particular producers, or 	through
collusion, is generally thought to be related 	to market concentration. 	To
measure concentration, the productive capacity of two groups of produ~ers are
typically taken into account: 	those currently producing the products in the
market, and those that could commence (but as yet have not) production of these
products within a "short" (Le., one-year) time period."2 	Firms that would find
it profitable to. produce and sell the relevant products, subsequent to a price
increase, would tend to be assigned market shares that reflect their ability to
convert their production facilities to engage in the production and sale of the
relevant 	products."3 	Such conversio'hs, often 	referred 	to 	as "supply-side"
substitution, are important in assessing market power. Thus, products A and B
might be regarded by consumers as such poor substitutes that they should not
be incorporated into the same product market. 	If, however, the production
facilities used 'for product B could quickly and easily be switched to the
production of A, the producers of B would be assigned market shares (in market
A) to help determine whether product A could be priced supracompetitively.
42 See § 2.21 of the Guidelines.
43 See Werden (1983), p. 519.
17

3. Applying the Guidelines to Cable Markets: 	An Illustration
To better illustrate the Guidelines· method for identifying a relevant
product market, consider the following example. 	Imagine that a city currently
is served by one large cable operator, several over-the-air broadcast stations, and
an MMDS that currently has a small number of subscribers. 	For simplicity,
"initially assume that the cable operator provides only one level (or "tier") of
service. This assumption will be relaxed, below.
To determine 	whether 	the 	service 	provided 	by 	~_re cable operator
constitutes the relevant product market, we might first ask whether a cable
operator 	in 	an 	unregulated 	market, 	if 	permanently 	shielded 	from 	direct
competition from other cable operators, could raise prices above competitive
levels. Consumers' alternatives to cable (in this example) are over-the-air (OT A)
broadcasts, VCR services, and MMDS services. 	The Question then is whether
enough subscribers 	would 	discontinue 	cable service 	(i.e., 	switch 	to 	OT A
broadcasts, 	VCRs, 	or 	tire 	MMDS) 	to 	make 	a 	small 	but 	significant 	and
nontransitory price increase unprofitable. 	If the answer is "no," then cable
services constitute a relevant antitrust product market." If the answer is "yes,"
then one would ask whether a similar increase in the price of both the cable
ser'vice and its next best substitute (the MMDS) would be profitable. Depending
upon the answer to this Question, it would be determined either that cable
service and MMDS service jointly constitute the relevant product market, or that
the product market should be once again expanded t9 include the next closest
substitute, and the profitability of a hypothesized joint price increase assessed.
" We do not address here whether the cable operator could increase rates to
advertisers. 	But the same analysis would apply. Thus, the Question is whether
price is constrained by alternatives to cable faced by advertisers so that the
cable operator would not find it profitable to raise price,. even if it had the only
franchise.
18

Suppose that it is determined at the conclusion of this procedure that the
relevant product mar kef consists of both cable and MMDS services. 46 The next
task would be to assess the degree of concentration in this market. 	A number
of different methods are available 	for calculating 	market shares and 	the
corresponding concentration indices. 	One could base these computations on
current subscriber shares, for example, or upon some measure of capacity to
serve subscribers. The measurement base should be chosen so that the resulting
concentration index provides th:~ best possible indicator of potential joint or
unilateral market power. 46 For example, although the current example presumes
that the cable system initially has many more subscribers than the MMDS, it is
hypothetically conceivable that the MMDS could quickly and easily expand its
subscriber base in response to a unilateral price increase by the cable system.
If so, constructing a concentration index from data-on current subscriber shares,
rather than on the capacity to serve subscribers, could exaggerate the market
power of the cable system, and understate the competitive constraint imposed
on the eable system by the MMDS. 47 Conversely, if the MMDS could only add
additional subscribers slowly, or at high cost, then it might be appropriate to
derive- the concentration index using existing subscriber share figures.
Recent 	years 	have 	witnessed 	the 	growth 	of a 	number 	of 	possible
alternatives to cable television. In addition to VCRs and MMDS, there are home
satellite receiving dishes and priyate cable system (Le., satellite master antenna
television systems, or "SMATVs"), all of which are capable, to some degree, of
providing video programming that formerly was obtainable primarily through
41 We emphasize 	that 	this assumption 	is 	made 	purely 	for 	illustrative
purposes; we make no claim as to the actual substitutability of MMDS services
for cable services.
46 See Werden (1983), p. 521, and Dansby and Willig (1979).
47 As a consequence, one might incorrectly ascribe unilateral (or "dominant
firm") market power to the -cable system, when in actuality the exercise of
market power might require the participation of both producers.
19

cable. 4' 	Whether the services produced by these alternative providers should be
considered part of the relevant product market is determined by the willingness
of cable subscribers to drop cable in favor of these substitutes. in response to a
price 	increase. 	This 	will 	depend 	upon 	the 	perceived 	quali ty 4Q of 	these
alternatives, 	and 	their 	prices.50 	The 	ability 	of 	the 	producers 	of 	these
alternatives to increase output to satisfy any increased demand will depend
upon their production technology, the speed with which they can increase
output, and other c~nsiderations, such as regulatory factors. 51
Whether these other services belong in the relevant market, and if so,
whether they can be expected to impose much of a constraint on an incumbent
operator, could vary across geographic markets. 	First, the quality of existing
cable systems is not geographically uniform. Consumers might find that over-
the-air broadcasting, combined with a VCR, provides a very good substitute for
the services provided by a 12 channel system, but a very poor substitute for a
S4 channel system. Also, underlying consumer preferences (and thus consumer
demands) for cable might vary across geographic markets. Third, the existence
of alternative suppliers might differ substantially across markets. For example,
in some regions of the country there are numerous private cable systems
41 This list is not exhaustive. 	As we noted in § III, above, some have
suggested that direct broadcast service ("DBS") television may also become a
source of competition for cable operators in the future. 	DBS systems would
beam programs to homes equipped with rooftop dishes directly from a satellite.
See "Direct-Broadcast TV May Be Getting Off Ground," Wall Street Journal,
February 21, .1990, p. B-1.
4Q It should be emphasized that -quality" as used here includes the suitability
and range of program choices offered by different media; e.g., a VCR can be
used to view movies but cannot provide live sports programming, which would
cause some consumers to consider it a "low quality" alternative to a premium
channel.
50 The prices of 	these alternatives will be determined 	in 	part 	by 	the
technological conditions, and by the prices that these producers face for inputs
(e.g., the cost of labor, equipment, programming, etc.).
SI See § VII of the comment for a discussion of additional factors that may
affect the supply elasticity of these alternative distributors.
20

(SMATVs) that perhaps could be granted municipal franchises,52 and thus be
able to expand geographically and provide a source of competition for the
incumbent cable operators. 53 	In areas where such producers are not already
present, presumably it would take longer for such producers to be able to serve
the market. 	Fourth, topographical and regulatory conditions will vary 	in
different markets. This limits the effectiveness of certain alternative providers.
MMDS, for example, works poorly when there are physical obstructions, such
as hills and tall buildings. 	Regulatory conditions might also affect the ability
of an MMDS to rapidly enter and serve a particular geographic market.
4. The Impact of the Regulatory Structure on Product Market Definition
Up to this point we have discussed the competitive constraints on cable
systems at a fairly high level of generality.54 The FCC is concerned, inter alia,
with a more focused question: identifying the competitive constraints faced by
cable operators on the provision of basic service. Identifying the"se competitive
constraints is not necessarily straightforward, however, since the composition
of basic service, and therefore its competitive alternatives, depends critically
upon the nature of the prevailing regulatory system.
52 SMA TVs generally serve multiple dwelling units, such as apartment and
condominium complexes.
IS Private cable systems may not cross a public right-of-way without a
franchise. 	See 47 U.s.C. § 522 (6)(B) (excluding systems that do not cross the
public right-of-way from the statutory definition of "cable systems"); 47 U.S.C.
§ 541 (b)(l) (requiring all "cable systems" to obtain franchises). 	It should be
noted that a number of states have enacted statutes that impede the ability of
municipalities to grant multiple cable franchises.
54 Section 623(b) of the Cable Act deregulated the price of basic service only
where the cable operator was subject to "effective competition" for basic service.
In its Report and Order (MM Docket No. 84-1296, SO Fed. Reg. 18637 (1985», the
FCC determined 	that effective competition exists where 	three off -the-air
broadcast signals are available in the cable community. 	The FCC recently
issued a NPRM on Effective Competition to determine if this standard should
be revised.
21

Prior to the passage of the Cable Act, the FCC had jurisdiction to
regulate cable TV to the extent that such regulation was "reasonably ancillary
to the ... regulation of televis·ion broadcasting."66 	Under this authority, the
FCC promulgated regulations that permitted all 	franchising authorities to
regulate 	the 	price 	of 	"regular 	subscriber 	service," 	but 	not 	specialized
programming with "per-program" or "per-ehannel" charges. 66 Subsequently, the
FCC defined "regular subscriber service" to include the carriage of broadcast
signals and any local origination or public access channels required under the
franchise. 61 	Premium channels, such 	as 	HBO, 	were 	not 	part of "regular
subscriber service," but rather constituted specialized programming with a per-
channel charge that could not be subjected to price regulation. 58 	Dista·nt
broadcast signals, such as WTBS, WGN and WOR, and advertiser supported
satellite channels, such as CNN, ESPN, USA and MTV, were also determined not
to be "regular subscriber service;" accordingly, the presence of such channels on
a service tier did not bring that tier within the' scope of permissible rate
regulation.5Q
The Cable Act changed the scope of permissible regulation by adding
the requirement that a cable system be without "effective competition" before
it could be subject to rate regulation. eo The scope of regulation, for all practical
55 See United States v. Southeastern Cable Co., 392 U.s. IS7, 178 (1968).
56 47 C.F.R. I 76.3I(a)(4) (1974).
51 Clarification of Cable Television Rules, 52 F.C.C. 2d 1,68 (1975).
5& See Brookhaven Cable TV. Inc. v. Kelly, 573 F.2d 765 (2d Cir. 1978), cert.
denied, 441 U.S. 904 (1979).
5Q See In re Community Cable TV. Inc., 95 F.C.C. 2d 1204, 1216-17 (1983)
(limiting the scope of permissible rate regulation to local broadcast signals as
defined by the FCCs "must-carry" rules).
eo 47 U.S.C. § 623(b).
22

purposes, however, remains the same. 61 	The franchising authority may only
regulate 	tiers 	of 	service 	that 	include 	"retransmission 	of 	local 	television
broadcast signals.,,62
According to the recent GAO Survey oj Cable TV Rales and Services, the
average cable system offers about 30 channels on its lowest-priced basic tier.
Many have inferred from this that local television broadcasts likely constitute
a poor substitute for basic cable service. This inference might. be correct, given
the existing composition of basic service. 	As the FCC has recently observed,63
however, the composition of the basic service tier (or any other tier) is not
fixed; cable operators have considerable discretion in choosing which channels
will be included in the basic service tier, and wh.ich will be incorporated into
higher tiers.64 This choice will be determined by profit considerations, which
will be influenced by prevailing regulatory standards. 	If a more stringent
"effective competition" standard were to be adopted (thus subjecting a greater
number of cable operators to regulation of basic service rates), these cable
operators would have a considerable incentive to reconfigure their tiers so as
to avoid the new regulations.66 That is. they would likely redefine basic service
so that it would consist principally (or _exclusively) of local, over-the-air
broadcast channels. and transfer other non broadcast channels to higher, still-
unregulated tiers, which could be priced to exploit whatever market power cable
has in supplying those market segments. 	It is conceivable that the array of
61 See A.CLU v. FCC, 823 F.2d 1554, 1558-59 (D.C. Clr. 1987), cerl. denied, 485
U.S. 959 (1988).
62 47 U.S.C. § 522 (2).
63 See paragraphs 16-18. and 44, of the NPRM on Effective Competition.
64 The F.C.C. has explicitly noted this possibility in its NOlice oj Proposed
Rulemaking on Reexamination of the Effective Competition Standard for the
Regulation of Cable Television Basic Service Rates ("Effective Competition
NPRM"), MM Docket No. 90-4, January 22, 1990, par. 44.
65 See, e.g .. Leffler (1982).
23

channels offered on the basic service tier would come to resemble closely the
array available over-the-air. 	Thus. while over-the-air broadcasts may 	not
present an effective constraint on the price of basic services as that service
tier is presently structured, it is plausible that these broadcasts would provide
a close substitute for basic services as the latter would come to appear in an
environment where only basic service rates were regulated. 68
It would be incorrect. of course. to infer from this logic that cable
operators face 	"effective competition." 	Depending 	upon 	the 	existence of
effective competitive constraints for the higher level service tiers, perhaps the
best way to view cable operators might be as multiproduct monopolists who face
regulatory constraints only on a subs~t of the services that they produce. If the
cable operators can easily shift services from regulated to unregulated service
tiers (as they apparently can). and face little or no effective competition in the
provision of these higher service tiers. then their market power will be largely
unchecked. 	Thus. as we note in § V, below, effective local regulation of cable
systems' market 	power 	might require 	that 	local 	authorities 	be statutorily
empowered to regulate the content of basic service. or to regulate the prices of
the higher service tiers. 87
68 This is simply a manifestation of a phenomenon that frequently arises
when producers are subject to price controls, but are left free to vary quality
and other dimensions of their service. Other types of price controls are known
to generate similar effects; a well-documented effect of rent control, for
example, is a deterioration of the rental property quality.
87 As we note in footnote 96, below, there may be constitutional barriers to
expanding regulation in this manner.
24

B. Empirical Approaches to Market Definition
The 	market 	definition 	method 	described 	in 	the 	preceding 	section
implicitly embodies the measurement of market power in terms of the "residual
demand elasticity" facing the hypothetical monopolist. sa The residual demand
elasticity incorporates both the demand-side responses of consumers. as well as
the supply-side responses of other firms. to a price increase by the monopolist.
A firm facing a highly elastic residual demand curve has little exploitable
market power. as a small price increase on its part will cause large reductions
in its sales. 69
We are unaware of any published efforts to estimate;directly the residual
demand functions facing cable operators. 	Webb (1983) and Pacey (1985) each
estimate equations that attempt to determine the impact of cable prices on cable
penetration rates (the ratio of subscribers to homes passed). 	However. these
estimates might not contain a great deal of information about the current state
of cable markets. because they are based on data that are relatively outdated. To
and because of econometric difficulties. 71
Recently. 	a 	number 	of 	authors 	have 	discussed 	the 	possibility 	of
empirically 	estimating 	residual 	demand 	functions 	to 	identify 	antitrust
markets.n 	In principle. this technique could be used to answer the question of
whether individual cable franchises 	have significant market power. or 	if
sa See Landes and Posner (1981).
69 A perfectly competitive firm faces an infinitely elastic (Le .• flat) re'sidual
demand curve.
70 The Webb estimates are based on data for 1979; the Pacey study used data
'-. for 1980.
71 The estimates are likely biased and inconsistent (in the statistical sense
of these terms) because the estimation procedures employed did not take account
of the simultaneity between the price of basic service and the number of
subscribers.
72 See Froeb and Werden (1990). Scheff man and Spiller (1987). and Baker
and Bresnahan (1984. 1985. and 1988).
25

instead 	they are subject to significant competitive constraints from 	other
distributors of video signals. While residual demand estimation is certainly not
the 	only 	useful 	empirical 	approach 	to 	identifying 	antitrust 	markets 	and
measuring market power, it is a potentially worthwhile tool that on some
occasions may provide a useful complement to information obtained by other
means.
Although this procedure is complex,13 its basic elements can be fairly
easily understood. Suppose that the costs of a cable operator go up (e.g., because
its program suppliers raised their prices). while the costs of its ostensible rivals
(e.g., a local MMDS) remain unchanged. 74 	If !the cable operator faced a "flat"
.
residual demand curve. it could not pass on any cost increases to subscribers:
any attempt to do so would cause its customers to switch to the services
produced by these rivals, and its sales would faU to zero. 	If. by contrast. the
cost of programming to aU of these different local distributors went uP. then
this cable operator could raise its price without suffering large decreases in
subscriptions, since its rivals also would have raised their prices to reflect their
higher costs. 	If the cable operator instead faced a "steep" rather than flat
residual demand curve. it could raise its price to cover its higher costs. even if
its rivals' costs (and prices) did not change. 	Depending upon how "steep" (Le .•
inelastic) this residual demand curve is, the cable services might qualify as a
separate product market.
This example embodies the essence of econometric estimation of the
residual demand curve. 	The basic approach is to explain the price charged by
a particular firm (or group of firms) in terms of (I) factors that affect only its
13 See Froeb and Werden (1990) for a detailed discussion of the subtleties
associated with the use of this technique.
74 Obviously, the technique that we are describing can be used only when
these rivals were actually present in the market during the sample period. This
empirical method cannot. in other words. measure the impact of competition
from new entrants.
26

costs, and (2) factors that affect its costs and the costs of possible rivals. 	If
variations in a firm's price can be explained by changes in factors that affect
only its costs, then the firm has S9me market power that can be quantified. 	If
variations in the firm's price can be explained only by factors that affect both
its costs and those of its rivals, then the inference of market power is weakened.
This procedure requires two types of data: 	cost data that are unique to the
firm(s) being analyzed, and cost data that pertain to this firm and to possible
rivals.
The residual demand approach is appealing because it captures supply
and demand respo~ses that determine whether a firm (or group of firms) could
exercise market power. However, the approach is not without limitations. Two
potential 	difficulties, 	as 	highlighted 	by 	Froeb 	and 	Werden 	(1990), 	are
extrapolation and nonstationarity. The extrapolation problem refers to the fact
that the market definition task may require "inferences about demand and cost
conditions away from the neighborhood of the prevailing equilibrium and
probably also outside the range of historical experience."l11 The nonstatj~narity
problem arises if demand and cost conditions are subject to major changes, so
that past experience (i.e.. as reflected in the data employed to estimate the
residual demand) might not provide a 	useful guide to future competitive
conditions.
We suspect that the extrapolation problem perhaps may be less an issue
in cable television than in other markets, since the deregulation of basic rates
gave operators a great deal of pricing discretion that had previously not existed.
The price, quantity, and quality changes that occurred in the aftermath of
deregulation may give one a much better idea about the nature of the demand
for cable services than otherwise would be possible.
75 Froeb and Werden (1990), p. 5.
27

Unfortunately, the price and quantity changes that have occurred during
the past few years may reflect not only changes in the regulatory environment,
but changes 	in 	underlying 	(and 	unobservable) 	consumer 	preferences 	and
production technologies. This could contribute to a nonstationarity problem, if,
as seems likely, changes of this nature are likely to continue. 	While these
factors diminish 	the confidence 	in any 	particular demand 	estimate, 	it 	is
doubtful that they eliminate their usefulness entirely. 	Any technique that
attempts to forecast future competitive conditions from historical experience
will be, to some extent, affected by this problem.
C. Conclusion
While we cannot provide 'any empirical generalizations regarding the
magnitude of cable operators' market power, 	we believe 	that the 	method
articulated in the 1984 Department of Justice Merger Guidelines and FTC Merger
Statement could 	focus 	the 	issues and 	provide guidance 	for 	its 	resolution.
Additionally, recently developed empirical techniques for measuring firm-
specific 	market 	power 	may 	provide a 	useful 	complement 	to 	information
gathered by other means. 	Although there are difficulties associated with the
use of these techniques, they may. nonetheless provide some useful insights into
the nature and extent of cable operators' market power.
28

V. Regulatory Options
Section II.D of the NOI requests a discussion of regulatory options should
it be determined that market forces are generally inadequate for the purpose of
limiting cable operators' market power. 	There arc three general options that
might be considered for this purpose. 	The first is traditional "rate-of -return"
regulation; the second is "price cap regulation"; and the third is "franchise
bidding." 	The potential and actual problems associated with rate-of -return
regulation have been well-documented, and numerous observers have argued
that price cap regulations and franchise 	bidding schemes hold out greater
promise as efficient regulatory mechanisms. 	In this section, we briefly review
the emerging theoretical and empirical literature that addresses this question
and attempt to assess its implications for any attempt to re-regulate cable
television.
A. Price-Caps Versus Rate-of-Return Regulation
Public utilities have traditionally been subject to rate-of-return (or "cost-
oC-service") regulation. Th.e goal oC this regulation is to exploit the advantages
oC natural monopoly, while requiring the monopolist to pass on the benefits of
these technological advantages (e.g .. scale economies) to consumers of the
product. 	Rate-oC-return regulation requires the regulator to estimate the cost
and demand conditions facing the monopolist, and to set a price (or a set of
prices) that just permits the firm to cover its costs, including a competitive
return on capita1.78
The potential difficulties associated with this .form of regutation are
many and well-documented." First, the regulator requires an enormous ;;mount
of information to carry out its task and generally must rely upon the regulated
78 Usually referred to as a "fair" return in most regulations.
" 	See Brennan (1989), and the sources cited therein.
29

firm to supply a substantial portion of this information. This regulatory process
is administratively costly and subject to manipulation on 	the 	part of the
regulated entity. Second, regulatory errors in setting the allowed rate-of -return
can lead to systematic biases in capital investment by the regulated firm. Third,
rate-of -return regulation provides the "firm with little incentive to reduce cost,
as any reductions are fully rebated to consumers. thus leaving the firm no better
off than if it had continued to operate at higher cost.7I Fourth, if the regulated
entity also sells in unregulated markets, the opportunity for profitable, yet
socially inefficient, cross-subsidization is created. 7Q Cross-subsidization occurs
if the firm attempts to build the costs associated with the unregulated product
into the rate base for its regulated product.
The existence of these problems has induced a search for alternative
regulatory mechanisms that might offer a promise of superior performance.
Price-cap regulation 	is one such alternative. 	Price-cap regulation 	can 	be
characterized as follows: 80 (1) the regulator sets a price ceiling, but, in contrast
to rate-of -return regulation, the firm has discretion to set its prices below this
ceiling; (2) th~ price ceiling is periodically adjusted" automatically by a factor
that is exogenous to the firm (e.J., an adjustment to reflect overall inflation);
and (3) over longer intervals, the ceiling and automatic adjustment factor are
subject to review and possible revision.
Proponents of price-cap regulation have made a number of claims on its
behalf. 	First. because firms are allowed to retain a portion of their cost
reductions. they may have a greater incentive to reduce costs than under rate-
"'- "
71 See Mathios and Rogers (1989), and the references cited therein.
7Q See Brennan (1987). 	Before cable rate deregulation, only the price of
basic service was regulated; "extended basic" service and premium channels were
not regulated.
80 See Acton and Vogelsang (1989), p. 370.
30

of-return regulation.81 Consumers might benefit from this action if price cap
regulation causes some portion of anticipated cost decreases to be passed on to
consumers through a reduction in the cap. 	The fact that price can be reduced
without prior regulatory review provides an added inducement (relative to rate-
of-return regulation) to cut price when costs fall.82 The administrative costs of
price-cap regulation are likely to be considerably lower than those associated
with rate-of-return regulation, since there is less regulatory oversight. Further,
the regulator's informational burden is probably lower under a price-cap regime.
Price-cap regulation may also make it easier to develop schemes that provide the
firm with an incendve to reveal cost and demand information truthfully to the
regulator. sa
Though 	powerful, 	the 	theoretical 	evidence 	in 	favor 	of 	price-cap
regulation is not unequivocal. 	Brennan (1989) and Schmalensee (1989) ha ve
argued on theoretical grounds that the supposed benefits of price-cap regulation
will not materialize under some conditions. The available empirical evidence,
however, suggests that consumers are better off when t>rice-cap regulation is
used instead of rate-of -return regulation. 	The only major empirical study of
comparative regulatory regimes done to this point is that of Mathios and
Rogers,84 who compared AT&Ts rates for intrastate, long-distance telephone
service in states that allow AT&T pricing flexibility with its rates in states that
use rate-of -return regulation. 	Their findings suggest that AT&T's daytime,
81 See Cabral and Riordan (1989).
82 Not only 	is regulatory, review of price reductions costly and 	time-
consuming, it also affords less 'efficient rivals an opportunity to prevent or
delay the price cut. if, as is typically the case, they have standing to participate
in these proceedings. 	Haring and K werel (1987) reported that AT&T's rivals
opposed virtually every price reduction proposed by AT&T since its 1982
divestiture, and some of these rivals may have been less efficient than AT&T.
sa See Sibley (1989).
84 See Mathios and Rogers (1988, 1989).
31

evening, nighttime, and weekend rates were significantly lower in states that
allowed pricing f1ex.ibility than in states that used rate-of-return regulation.
If consideration is to be given to re-imposing rate regulation on cable
operators, 	we 	believe 	that 	price-cap 	regulation 	should 	be 	given 	serious
consideration 	as 	an 	alternative 	to 	rate-of-return 	regulation. 	The 	lower
information 	requirements 	of 	price-cap 	regulation 	may' be 	particularly
advantageous in this setting, given that the regulatory responsibilities will often
be carried out by local authorities who may lack the resources to analyze or
otherwise handle large quantities of information.
lAlthough price-cap regulation may outperform rate-of-return regulation
under a variety of circumstances, there arc problems, as Brennan (1989) has
observed, that even perfect price-cap regulation cannot adequately address.
One of these is the manipulation of product quality. As numerous authors have
demonstrated,81i the level of quality emerging under monopoly will differ from
that associated with competition. 	We discussed in § III of this comment 86 how
the regulatory framework that existed prior' to the enactment of the Cable Act,
as well as that which exists today, permits and encourages the redefinition of
service tiers so as to avoid regulatory rate controls. 	If it is determined that
some form of explicit rate regulation is necessary to constrain the market power
of cable operators, effective regulation 	would seemingly 	require 	that the
regu'latory authority be empowered either to control the rates charged for all
service tiers, or to i~pose "minimum quality" controls along with rate controls
(e.g., stipulate that "basic" service offer certain specified channels).81 Without
iii See, for e~ampLe, Leffler (1982), and Besanko, Donnenfeld, and White
(1987, .1988).
86 Also see NPRM on Effective Competition, par. 44.
81 We recognize that the law is unsettled as to whether the FCC is authorized
to require cable operators to provide certain services as part of their basic
service package. 	See NPRM on the Effective Competition Standard, par. 44.
Constitutional issues may also arise here. 	See footnote 96, below.
32

such 	regulatory 	authority, 	it 	is 	unclear 	that 	rate 	regulation 	could 	ever
effectively constrain the exercise of market power.
B. Reaulatlon Throuah Franchise Siddioa
As an alternative to detailed regulation of prices. some observers have
proposed that natural monopolies be regulated through a franchise bidding
process." 	Under this scheme. the government does not regulate in detail the
monopoly. 	Instead. it grants a monopoly franchise to the firm that promises
in advance to charge (for a given service quality) the lowest price throughout
the franchise period. 	The potential problems with this approach have been
widely discussed in the literature.8~ Much of this criticism has focused on the
possibility of "post-contractual opportunism." Once granted the franchise, the
monopolist will have an incentive to reduce quality. and to exaggerate costs to
provide a pretext for deviating from the low price that it had promised to
charge. 	There is also the possibility that the government will engage in post-
contractual opportunism; it might. for example, later demand services valued
by small but well-organized groups that raise the franchisee's costs by more
than the associated benefits to all subscribers.~
The performance of franchise bidding schemes in cable television has
been analyzed in several recently published papers.~l The authors conclude that
the franchise bidding process. though imperfect. seems to perform better than
its' critics predict. 	Zupan (l989a. p. 439). who studied cable rates prior to
deregulation. concludes that "ex ante competition for franchise awards as well
II The seminal article on this is Demsetz (1968).
8~ See Williamson (1976). who saw cable television as the archetypical.
example of a market where such a scheme would not work.
~ 	Examples 	of 	this 	are 	"institutional 	networks" 	and 	"community
programming." 	See Zupan (1989a. pp. 404-7) ..
91 See Zupan (1989a. 1989b. and 1989c), and Prager (1989).
33

as the informal and formal rate control mechanisms possessed by most local
franchisors appear to be quite successful at preventing monopoly pricing."
Zupan (I 989c, p. 481.) found that "with regard to basic tier prices and the prices
charged per channel of basic service, the terms of trade are statistically no more
favorable to operators in renewal contracts than in initial franchise a wards."
Similarly, Prager (1989, p. 126)g2 finds "a notable absence of the types of
problems that Williamson gS might lead us to expect."
A present problem with franchise bidding (when compared with the
periods studied by Zupan and Prager) is that the 1984 Cable Act may have
made it easier for cable operators to engage in post-contractual opportunism.
. 	. ..Zupan (I 989a, p. 439) concludes that the "institutional mechanisms that so
effectively restrain guileful operator conduct are cross-city and within-contract
reputational considerations and cities' 	monopsony 	power." 	The 	threat of
franchise nonrenewal also seems likely to have played a role in constraining a
franchisee's 	conduct. 	At 	the 	time of expiration, 	the 	operator 	may 	have
substantial market-specific physical and human capital whose value may in part
be lost if its franchise were not renewed. Q4 The 1984 Cable Act, however, may
make it more difficult for local governments to threaten nonrenewal. Stction
626(c)( I) limits the criteria that the government may use in decidin$ to not
renew an operator's franchise. 	This deCision may not be based on the prices
charged by the operator, nor on "the mix, quality, or level of cable services or
other services provided over the system."gS 	These are obviously the service
variables of greatest interest to cable subscribers. 	The fact that cities cannot
g2 Prager studied the franchise bidding process in Massachusetts over the
period 1973-81.
Q3 See Williamson (1976).
Q4 See Zupan (l989c), p. 477. 	The incumbent might be able to recover part
of the cost of its physical capital if it can sell it to the new franchisee.
gil 47 U.S.C. § 626(c)( I).
34

VI. 	"TraCflckina" In Cable Systems
Paragraph 47 of the NOI seeks comment on whether direct regulation
of the sale of cable systems is necessary; i.e .• whether "trafficking" in cable
systems is undesirable and should thus be regulated. 	In the present context.
"trafficking" refers to the purchase and rapid resale of cable systems.
The FCCs interest in the possible detrimental effects of trafficking is
longstanding. 	Previously, concern focused principally on the sale of broadcast·
licenses and permits for the purpose of making "a profit rather than for the
purpose of rendering a public service."~1 The chief objection to the unrestricted
exchange of these assets was that frequent changes in station ownership would
create 	"uncertainty 	on 	the 	part 	of 	station 	personnel 	and 	disruption 	in
operational continuity (and thus) cause programming deterioration incompatible
with programming in the public interest."QI The FCC's apprehension about the
deleterious effects of trafficking 	was grounded partly on 	the 	belief 	that
"'speculation in licenses' would lead to increased station prices. allowing 'only
wealthy individuals or businesses' to purchase broadcast properties. and possibly
leading the new ownership to decrease the quality of programming and increase
the quantity of commercials in an effort to recover its investment."gQ
In 	questioning 	the 	appropriateness 	or 	necessity 	of 	antitrafficking
regulation. the NOI observes that some parties have alleged that trafficking in
cable systems is responsible for much of the rise in basic rates that has occurred
~1 See Paul Crosely. Jr., 11 F.C.C. 3, 23 (1945).
98 Notice 01 Proposed Rulemtlking, FCC 60-1466, 25 Fed Reg. 12898, Docket
No. 13864, December 1960. 	This proceeding led to the adoption of the "three
year rule," which was repealed in 	1982. 	The "three-year rule" required a
broadcast application for assignment or transfer filed prior to th.e completion
of a three-year holding period to be designated for a hearing, unless there were
certain. specified extenuating circumstances, or a meritorious waiver request.
gQ Notice 01 Proposed Ru/emaking, In The Matter of Amendment of Section
73.3597 of the Commission's Rules (Applications for Voluntary Assignments or
Transfers of Control), BC Docket No. 81-897, (December 30. 1981).
36

in recent years. lOO 	Accordingly, the NOI solicits comments on what effects
regulation of trafficking would have on the market price of cable systems, basic
cable rates, and innovation or technological change by cable systems.
In evaluating whether the unrestricted sale of cable systems would likely
adversely affect consumers, we assume current system owners operate their
systems in ways that maximize their market value. lOl We also assume initially
that the buyer of a cable system is no more efficient than the seller. 	In many
cases this will not be true; as will be discussed later, the motivation for many
acquisitions is likely to be the gains from improvements by more efficient
operators.
Assuming that cable system owners seek 	to maximize profits, 	it is
difficult to see how "trafficking" in and of itself could affect the rates charged
to cable subscribers. The standard analysis of profit maximization calls for the
owner of a firm to choose prices and outputs such that any subsequent change
adds more to costs that it docs to revenues.- This is the familiar concept in
which the firm equates marginal revenue and marginal cost. Marginal revenue
is determined by the demand for cable services. 	The demand facing a given
cable operator is determined by such factors as: 	(i) the quality of the cable
system (e.g., number and quality of channels); (ii) the extent of competition
from otber sources (e.g., over-the-air broadcasters); and (iii) the income, tastes,
100 See. e.g .• the response of the state of Hawaii to the NOI. 	According to a
recent survey by the General Accounting Office (August 1989), the average
nominal (i.c., not adjusted for inflation) charge for basic service increased by
26 percent between December 1986 and October 1988. 	The real (inflation-
adjusted) change was smaller (about IS percent). This increase in basic service
rates was accompanied by a 21 percent increase in the average number of basic
channels offered (Le., from an average of 26.6 stations to 32.1 channels). 	The
United States Telephone Association has suggested that basic service rates
increased by 68 percent between 1986 and 1989 (see submission of USTA, "Fifty
State Summary of Cable Service Fees for 1986 and 1989", Appendix 6, 1990).
For reasons to be discussed, we are doubtful that this price increase can be
attributed to "trafficking."
101 That is, we assume that each owner seeks to maximize the discounted
present value of the system's profits.
37

and preferences of the population. Cost is determined by production technology
and factor prices faced by the cable system. 	The fact that a cable system
changes hands would not by itself be expected to alter marginal revenue or
marginal cost, so it is difficult to see how simply changing ownership would
change price or output. 	To do either would involve a sacrifice of profits.
This conclusion is not altered by adopting more realistic assumptions
about cable markets. Cable operators make decisions concerning not just price
and output, but also over the quality of service provided. 102 	In the short-run,
with fixed cable capacity, decisions over product quality (e.g., program choices,
quality of signal,· maintenance of system) are made so as to earn the highest
possible return; in the longer-run, choices on system capacity and other quality-
related investments also will be made to maximize returns. 	As in the simpler
analysis, however, it is hard to see why these decisions would change simply
from a change in ownership - or if these decisions do change, why they would
change in ways that would make consumers worse-off. 	The determinants of
demand and cost do not change when the system changes hands; hence, neither
will the profit-maximizing price, quantity, and quality. If the previous owner
was 	maximizing 	returns, 	any 	departure 	from 	this 	price-quality-quantity
combination would again involve a reduction in profits. lOS
The arguments above take on particular force if it is true that the selling
prices of cable systems have been increasing. 	It is the desire to earn capital
gains from 	the 	purchase and resale (at a 	higher price) of a system 	that
presumably motivates a "trafficker." 	If a buyer pays an inflated price for a
system. and subsequently operates it in a way that fails to maximize current
102 See Leffler (1982), and Besanko, Donnenfeld, and White (1988).
lOS We are not asserting that managers could not, from time to time, make
errors that reduce the value of the firm; rather, we contend that systematic
overpricing or underinvestment are unlikely to occur.
38

profits, it becomes doubtful that the buyer will earn even a competitive return
on the investment during its time of ownership.
Such behavior, moreover, tends to reduce the value of the system to any
subsequent purchaser if, as seems likelY,the system's market value incorporates
"goodwill" or "brand' name" capital (Le., the discounted returns to high quality
production). 	For example, if a system develops a reputation for poor quality
(subsequent to its acquisition by a "trafficker"), the demand facing this system
will fall. 	This decreases the price that the current (and future) owner can
charge for (say) basic service, which consequently decreases present and future
profitsYu Unless prospective purchasers of this system are somehow unaware
of this decline in system quality, the price that they will pay for the system
subsequent to the quality decline will fall. 	The "trafficker" will, under these
circumstances, suffer a capital loss on its sale of the system.
We have ignored so far the possibility that sales of systems represent the
transf~r of assets from less to more efficient operators. 	It seems plausib'le that
many buyers of cable systems will pay a premium for an existing system because
they can provide the same service at lower cost, or better service at the same
cost, than can the incumbent. The purchaser will keep some of the returns from
this superior efficiency (hence its willingness to pay a premium for the asset),
but in general will also pass some of these gains on to consumers. 106 Preventing
or restricting sales of cable systems would delay such efficient transfers and
thus hurt consumers. 	Quality-adjusted cable rates could be higher, not lower,
as a result of restrictions on cable system sales.
104 A subsequent purchaser could improve quality. However, this purchaser
would probably have to set lower prices for some period of time to lure
disenchanted former subscribers back to the system. This postpones the buyer's
ability to raise price to compensate for the higher cost of higher quality service,
and thus reduces the present value of the system. 	See Shapiro (1983).
106 A profit-maximizing firm will generally reduce price in response to cost
reductions.
39

For similar reasons, 	restrictions on "trafficking" could 	increase the
riskiness of cable system investments and accordingly reduce the incentives to
make these investments. The risk of an investment is reduced when the investor
is able to sell the asset to another party (e.g., should the venture not fulfill its
original profit expectations). Because antitrafficking rules lock franchisees into
an economic arrangement for a fixed time period, they inhibit the ability of
cable operators to recoup any part of the costs of an unsuccessful project until
this time period expires.
It is easy to see how antitrafficking rules might deter new investment.
First, investors will be less likely to enter previously unserved markets. Making
it harder to exit a market in the event of unprofitable entry (perhaps by selling
to a more efficient aCQuirer) will reduce the incentives to enter in the first
place. Similarly, "overbuild" competition (i.e., entry into an area with existing
cable service) will be deterred. 	By making exit harder, an antitrafficking rule
can discourage entry.
Upgrades of existing systems might also be deterred. 	For example, a
firm that increases its channel capacity might find, ex post, this investment
unprofitable because (for example) it is a poor marketer. 	By preventing this
firm from selling its system to a more efficient marketer, anti trafficking rules
could prevent recoupment of investment costs, and thus deter the firm from
making the investment.
Overall, it is difficult to see how restricting "trafficking" would improve
consumer welfare. Those who have attributed increases in cable service prices
to trafficking may be confusing cause and effect. 	In the aftermath of cable
deregulation, it is possible that cable operators that faced little potential or
actual competition were able to exercise market power that had previously been
40

constrained by price regulation. 106 The capitalized value of any newly-created
monopoly returns would have been reflected in an increase in the value of cable
systems. as would the increase in value arising from the elimination of any
inefficiencies that might have been induced by rate-of-return regulation (see
§ IV of our comments for a brief discussion of these).lo7 Moreover. it is clear
from the General Accounting Office Surv~y of Cab/~ Te/eyision Rates (1989) that
the quality ofbasic service rose after the passage of the Cable Act. 108 	Other
things held equal. this quality increase would likely be reflected in a higher
average price for basic cable service. 	Thus. to the extent that deregulation
induced an increase in the number of system exchanges. as investors pursued
possible monopoly or efficiency gains. 1OQ one might' observe (I) higher ca ble
rates. (2) higher system values. and (3) more system exchanges. 	The direction
of causation. however. would be from (l) to (3). rather than the opposite. as
some may have suggested.
Allowing 	the sale of cable systems 	without 	regard 	to duration 	of
ownership would be consisteiu with existing FCC policies with regard to the
transfer of other operating licenses. As noted earlier. in 1982 the FCC abolished
its so-called -three year rule.- 	This rule required that a broadcast application
for assignment or transfer filed prior to the completion of a three-year holding
period be desig.nated for a hearing. In abolishing the rule. the FCC determined
106 The possible natural monopoly characteristics of cable markets. as well
as the reluctance of local authorities to grant additional franchises. might
contribute to a lack of entry in cable markets.
lOT Zupan (1989. p. 409) notes that the stock market value of cable companies
rose by 16 percent in the two months following the enactment of the Cable Act.
During that same period. the Standard'" Poor Index rose 8 percent. and the
Dow Jones average increased 6 percent.
108 According to the GAO Suryey. the average number of channels received
per subscriber on the lowest-priced basic service tier rose from 23.8 in 1986 to
30.1 in 1988. See GAO Survey. table 111.5.
IOQ We do not know if there has actually been an increase in the number of
cable system transactions since enactment of the Cable Act.
41

that 	"allowing 	the 	marketplace 	to 	operate 	as 	to 	transactions 	involving
[broadcast) stations will undoubtedly lead to better broadcasting in the public
interest. NllO 	Increased 	reliance 	on 	market 	forces 	to 	allocate 	resources 	is
consistent with other FCC programming initiatives lll and has ample support in
law. 1l2
In conclusion, we do not believe that there is a compelling economic
rationale for restricting the sale of cable systems. 	Such restrictions would be
unlikely to help, and could in fact harm, cable customers. 	Accordingly, we do
not believe that the public interest would be served by the creation of statutory
or regulatory restrictions on the purchase and resale of cable systems.
110 See Notice of Proposed Ru/emaking, In the Matter of Amendment of
Section 	73.3597 	of 	the 	Commission's 	Rules 	(Applications 	for 	Voluntary
Assignments or Transfers of Control), BC Docket No. 81-897, at 5-6 (December
30, 1981).
111 See. e.g .. Deregulation of Radio, 84 F.C.C. 2d 968, reconsid. denied. 87
F.C.C. 2d 797 (1981).
112 In FCC v. WNCN Listeners Guild, 450 U.S. 582 (1981), the Court held that
the FCC's decision to rely on market forces for promotion of radio format
diversity is not inconsistent with the statutory requirement that licensing be
based upon "public interest, convenience, and necessity."-
42

VII. Vertical Integration and Other Vertical Restraints
Section 	III 	of 	the 	NOI 	poses Questions on 	the 	effects 	of 	vertical
int~gration on competition in the different segments of the video services
marketplace. 	Interest in vertical integration between program suppliers and
cable systems stems from fears that such vertical relationships may result in
anticompetitive "foreclosure." 	There is concern that vertical integration may
alter a firm's incentives to deal with firms that are not integrated with it, and
that this change in incentives may reduce, rather than improve, consumers'
welfare. us 	There are, however, both anticompetitive 	and 	procompetitive
reasons for vertical integration and other forms of vertical control. 	What
follows is a review of the theoretical literature relevant to an understanding of
the competitive effects of vertical controls, as well as an overview of the
empirical 	studies 	that 	have 	attempted 	to 	assess 	the 	effects 	of 	vertical
integration between program suppliers and cable systems.
A. Economic Theories or Vertical Integration and Control
Although the NOI inquires specifically about vertical integration, it is
important to emphasize that vertical integration is but one of many for~s of
vertical control. Whether its motives are pro- or anticompetitive, a firm may be
able to select from an array of vertical controls that can, in principle, yield
similar 	profits. 	In 	fact, 	there 	is 	a 	substantial 	theoretical 	literature
us This concern was raised by a n~mber of respondents to the NOr. See, e.g ..
Statement of Michael E. Cortese, Chairman of the National Private Cable
Association and President of Western Cable Communications, Inc., Before the
Federal Communications Commission Los Angeles Field Hearing, February 12,
1990; Statement of Milton Katz, Chairman of the 	Board & 	CEO, Malrite
Communications Group, Before the Federal Communications Commission Los
Angeles 	Field 	Hearing, 	February 	12, 	1990; 	Testimony 	of 	George 	Ring,
Chairman, Cross Country Telecommunications, and Member, Wireless Cable
Association, Before the Federal Communications Commission Los Angeles Field
Hearing, February 12, 1990; and Statement of Michael Pandzik, President,
National Cable Television Cooperative, Before the Federal Communications
Commission Los Angeles Field Hearing, February 12, 1990.
43

demonstrating that when certain carefully specified conditions are fulfilled, a
variety of different vertical controls are equivalent, in that each yields (I) the
identical profits to the firm imposing the control, and (2) the identical price and
output 	to consumers. 114 	In 	reality. 	however, 	it 	seems 	unlikely 	that 	the
conditions for equivalence will be met. 	The particular control chosen in any
given instance will be determined by the costs of writing, monitoring, and
enforcing different types of contracts, and by the particular characteristics of
production and distribution. Contracting costs can be substantial, and can vary
across industries, across firms within an industry, and over time, depending
upon a variety of circumstances. us
Accordingly, 	firms 	cannot always costlessly substitute one 	vertical
control for 	another. 	Controls that are equivalent 	in 	a 	world 	of 	perfect
information 	and 	costless 	enforcement 	may 	differ 	substantially 	if 	these
assumptions are relaxed. 	To determine which controls are best suited to their
114 See Blair and Kaserman, 1983, ch. 4. The principal alternatives to vertical
integration are output royalties, sales revenue royalties, and lump-sum entry
fees paid by downstream firms to upstream firms. 	The equivalency of these
alternatives has been theoretically established under the assumptions that (I) the
downstream industry is perfectly competitive, (2) downstream 	firms 	have
identical, constant-returns to scale production technologies, and (3) downstream
demand is constant and known by all parties with certainty. If these conditions
are not fulfilled, the different vertical controls will generally not yield the
same results.
Additionally, 	these alternatives 	generally 	will 	not 	be 	equivalent once
transaction costs are taken into account. Successful implementation ofa royalty
scheme, for example, requires that the upstream firm be able to monitor the
output or revenue of the downstream firm (who will have an incentive to evade
the royalty payments by understating sales or output). The lump-sum payment
(whereby the upstream firm charges a lump-sum fee to downstream firms for
the right to buy the former's product) requires considerable information, since
the profit-maximizing Tee will vary across geographic markets if there is
geographic variation in downstream demand. Also, if the downstream average
cost curve is "U-shaped", rather than flat (i.e.. the downstream production
technology docs not exhibit constant returns to scale), then the lump-sum
contractual alternative will fail to yield profits to the upstream firm equivalent
to the profits from integration. 	See Blair and Kaserman, p. 75.
I1S Analyzing the determinants of these costs, and their implications for the
vertical structure 	of an 	industry, 	is the 	focus of the "transactions costs"
literature. 	See, for example, Klein, Crawford, and Alchian (1978) and Klein
and Murphy (198-8).
44

economic environment, firms may engage in experimentation and adaptation.
Experimentation may be particularly important when shifts in demand and
technology cause the economic environment to evolve rapidly.
1. Incentives for Programmers to Adopt Restricted Distribution
The possible economic incentives for upstream firms (e.g., "programmers")
to impose restraints upon a downstream firm (e.g .• a given distributor) are
:~umerous. The most general view of the purpose of vertical restraints is that
they 	are 	means 	by 	which 	a 	"principal-agent" problem 	is 	solved. 1l6 	The
distributor is the agent of the programmer (the principal), and may ha ve
incentives that differ from those of the programmer. 1l7 One purpose of vertical
restraints is to align these otherwise incompatible incentives.
There arc a number of reasons why these incentives may differ, and
therefore why an upstream firm may wish to integrate vertically, or adopt other
vertical controls. These differences in incentives may relate to (1) the efficient
marketing and promotion of the product by distributors, (2) constraining the
exercise of market power by distributors, (3) avoiding certain market distortions
arising from upstream market power. (4) increasing barriers to entry, and (5)
evading rate-of-return regulation. 	These motives arc not mutually exclusive,
and all may be present to some degree in the video marketplace.
11e See Ippolito (1988), p. 13. and the references cited therein.
111 As Ippolito (p. 13) notes. the incentives will "depend on the information
available to both the dealer and the manufacturer, the ease with which the
manufacturer can control various actions of the dealer, and other market
factors, such as the relative risk aversion and the comparative efficiencies of
the parties in providing particular services related to selling the product and
determining its quality,"
45

a. Inducement of Marketing andPromOlion Efforts
In many settings upstream firms will find it efficient to arrange for
distributors to market and promote the product or service. In the case of cable
television programming. it appears that some programmers find it desirable to
have their distributors (e.g .• cable systems) engage in advertising through local
media. and to enga'ge in tightly focused marketing efforts (e.g .• direct mail and
telephone 	solicitation 	of 	nonsubscribers). 	It 	might 	be -efficient 	for 	the
distributor, rather than the programmer, to invest in these efforts because of the.
distributor's greater familiarity with the characteristics of consumers in a
particular geographic area. If it is infeasible or inefficient for the programmer
to finance all of this activhy itself, thus requiring the commitment of resources
by distributors, the danger arises that some distributors of the program might
attempt to "free-ride" on the promotional efforts of others. 	Unless this free-
riding can be constrained, distributors will have a disincentive to invest in
promotion, since some of the attendant increase in den;tand will be captured by
rival distributors. UI
Under conditions such as those described above, it may be 	in 	the
programmer's 	interest 	to 	adopt 	a 	restrictive 	distribution 	policy.Ug 	The
programmer 	might, 	for 	example, 	grant 	its 	distributors 	"exclusive 	sales
territories," which would permit one designated distributor to solicit potential
customers in one particular area (and only in that area). 	Other potential
carriers of the program (such as overbuilders and MMDS) would not be allowed
to distribute the program to customers in that area. 	Adoption of such a policy
would not necessarily involve vertical integration with the distributor; exclusive
UI For example, a SMA TV operation (i.e., a private cable system that serves
(say) a multi-dwelling complex) that is located within the area served by a
franchised cable operation might be able to engage in this sort of free-riding.
Ug When challenged under the antitrust laws, nonprice vertical restraints are
analyzed under the Rule of Reason. Conti/rental TY .. Inc. v. GTE Sylvania. Inc ..
433 U.S. 36 (1977).
46

sales territories conceivably could 	be granted 	through explicit or 	implicit
contracts. Vertical integration might become attractive, however, if it is costly
to enforce such contracts (e.g., distributors may continue to try to invade each
other's territories).120 Ownership of the downstream entity might sometimes be
the least-cost way of preventing such activity.121
A related incentive for vertical control can arise when a distributor's
failure to provide high-quality distribution services diminishes the value of a
upstream firm's product. For example, a poor quality cable distributor (such as
one whose signal is frequently interrupted, whose system maintenance is poor,
or who provides a poor signal) will have fewer subscribers than a high Quality
operator, and, in turn, any given program supplier will have fewer potential
subscribers and lower profitability.
Video programmers might also have an incentive to restrict distribution
if they are concerned about the inability of certain types of distributors to
provide signal security. 	Certain distribution technologies, such as multipoint
distribution services (MDS), distribute their signals 'locally using microwa ve
transmission. 	Even if these signals an; "scrambled" to prevent unauthorized
reception, 	it 	is 	sometimes 	possible 	for 	persons 	to 	obtain 	de-scrambling
equipment from suppliers that have managed to duplicate the de-scrambling
technology. If so, it may be possible for persons to receive certain video signals
without compensating the suppliers of the programming. This phenomenon, if
it 	occurs 	on 	a 	sufficiently 	large scale, 	might 	reduce 	program 	suppliers'
120 Vertical integratio'n'might also become attractive if distributors fear
post-contractual opportunism by upstream firms (e.g., a programmer might
license other distributors after the 	incumbent distributor 	has 	made sunk
promotional expenditures). 	See Klein, Crawford, and Alchian (1978).
121 Creation of an exclusive sales territory also provides the distributor with
a local monopoly in the sales of that product. 	The upstream firm will wish to
constrain the exercise of any market power associated with this monopoly;
vertical integratior:t may be the most efficient method for this. 	This issue is
described in greater detail below.
47

incentives 	to 	permit 	their 	signal 	to 	be 	distributed 	by 	firms 	using 	these
technologies.
In many situations, the upstream firm can arrange adequate marketing
and promotion efforts by combining some contractual vertical restraint with the
threat of dealer termination for inadequate performance. 122 	This approach
would 	be 	feasible 	when 	there 	are 	numerous 	possible 	distributors 	in 	any
particular 	geographic 	market, 	since 	the 	upstream 	firm 	can 	easily 	find
replacements for 	the inefficient distributor. 	It might not be credible to
threaten a distributor with termination, however, when alternative distributors
are not present (as may often be the case with cable distribution), since there'
may be no other way of reaching a large number of potential customers. In such
an instance, the least-cost method for ensuring that the distributor provides the
desired level of quality might be to vertically integrate program production
with distribution. 	The extent to which integration between cable systems and
programmers has been motivated by a desire to ensure adequate marketing and
promotion is not known, however.
b. Constraining the Exercise of Market Power by Cable Operators
An upstream firm that faces a downward-sloping demand cutve will
have an incentive to impose controls on its distributors if the latter have market
power in the resale of the former's products. 	When firms at each stage of
production can raise price above marginal cost, a "double marginalization"
problem may exist. "Double marginalization· reduces the combined profits of
the distributor and the producer, as well as total output.. The problem is
potentially 	correctable 	through 	vertical 	integration 	and 	other 	vertical
controis.1U
122 See Klein and Murphy (1988).
1U Analysis of this issue was first presented in Spengler (1950).
48

The problem is easiest to illustrate by assuming a monopolist at each of
the two stages of production. 	Each monopolist will set marginal revenue
(instead 	of 	price) 	equal 	to 	marginal 	cost 	(hence 	the 	term 	"double
marginalization"). 	The upstream firm sets a monopoly price by equating its
marginal revenue to its marginal cost. 	The downstream 	firm also sets a
monopoly retail price by equating its marginal revenue to marginal cost; the
latter's marginal cost will reflect the upstream monopoly price charged to it. In
this case, total output will be smaller, and price charged to ultimate consumers
higher, than if both monopolists were under the control of a single entity.
The upstream firm has an obvious incentive to constrain the exercise of
downstream market power (and the downstream monopolist has an incentive to
constrain the exercise of upstream market power). 	When a distributor exploits
market power, it raises the price to consumers above the price that would
prevail under competition; the quantity sold to consumers falls. 	This rise in
price in turn reduces the (derived) demand facing the producer, and therefore
the price that the upstream firm can charge for its product, clearly making the
upstream firm worse-off.
If these firms merge, the integrated producer will no longer "charge
itself" 	a 	monopoly 	upstream 	price 	when 	it 	transfers 	its 	product 	to 	its
downstream subsidiary; rather, it will internally price the good at marginal cost.
It 	continues 	to 	set 	a 	monopoly 	price 	for 	the 	final 	good; 	but 	since 	the
downstream subsidiary's costs are now lower because an input price has fallen,
its profits after integration will be greater than the combined pre-integration
profits of the two firms operating independently of one another. 124 	When the
firms do not operate at cross-purposes, total profits will rise. 	Consumers are
also apt to benefit from the integration of the two monopolists. If the upstream
price falls, downstream marginal cost will also fall. The new profit-maximizing
124 See Blair and Kaserman (1983), p. 31.
49

downstream monopoly price will, therefore, be lower, and downstream output
larger, as a consequence of the vertical integration.
"Double marginalization" issues could arise when a producer deals with
more than one distributor. For example, an upstream firm may sell to a number
of distributors, each of which has some market power in its own geographic
market. 	This market power could arise from a number of sources, such as
government entry regulations, scale economies (i.e., natural monopoly), or even
the producer's own decision to grant exclusive sales territories (see above).
This situation might be applicable to the relationship between program
suppliers and cable systems. 	Cable operators may have market power in their
franchise areas owing to some combination of the three factors listed above. If
so, a program supplier will have an incentive to adopt measures that constrain
each system's ability to raise consumer prices. 	For example, if the program
supplier cannot authorize additional distributors (e.g., because of downstream
entry barriers) or chooses not to do so (e.g., because potential free-riding among
distributors makes it desirable to grant exclusive sales territories), it may try to
limit the distributo~'s exercise of market power through other means. 	One
contractual possibility is to impose maximum resale price maintenance on the
distributor; i.e., contractually forbid the distributor from charging subscribers
a 	price that 	exceeds some specified 	level.125 	There 	arc 	several 	possible
drawbacks to this approach, however. 	First, program suppliers may be legally
constrained from imposing maximum resale prices. 	Historically, maximum
resale price maintenance (RPM) has been accorded the same status as minimum
resale price maintenance: 	per se illegality.us 	Thus, firms may perceive that
125 See Blair and Kaserman, pp. 35-6. Another contractual alternative is the
use of performance standards, such as minimum output quotas. 	.
126 See Kiefer-Stewart Co. v. Joseph E. Seagram and Sons. Inc .. 340 U.S. 211
(195 I)
50

there are legal constraints on their ability to write and enforce a maximum
RPM agreement.
Second, 	even 	without 	legal 	constraints, 	entering 	and 	monitoring
compliance with the maximum RPM contract could be costly if there are a
large number of distributors. Third, RPM is not a good substitute for vertical
integration when there is uncertainty and risk-aversion. 	For example, the
inability of a distributor to raise price to cover unexpected increases in retail
costs forces the distributor to bear all of the risk of retail cost variations. u7
Yet another contractual alternative to vertical integration is the use of
a "franchise fee" (sometimes referred to as a "two-part tariff") by the upstream
firm. 128 Under this method, the price charged by the upstream firm consists of
two parts: a flat fee for the right to buy the product, plus a per-unit price. 12Q
The basic approach is to set ~he per-unit price equal to upstream marginal cost,
which induces the downstream firm to choose the same output level that would
be chosen under vertical integration. 	The flat fee is then set at a level that
expropriates the increase in downstream profits that are generated by the lower
per-unit price. 	Under some circumstances, the two-part tariff can completely
solve the upstream firm's proble~. 	These conditions are fairly restrictive,
however. Tirole (1988, p. 177) concluded that "[f]ranchise fees, in general, will
not suffice to realize the vertically integrated profit."I30
127 See Tirole (1988, p. 177).
128 See Tirole (1988, pp. 176-86).
In In other words, the total payments from the retailer to the upstream firm
equal F + p.q, where F is the flat fee, p. is the wholesale price, and q is the
number of units of output purchased by the retailer. 	.
130 As we noted earlier, a franchise fee scheme requires a great deal of
information if the upstream firm is to replicate the profits that would be
available through vertical integration. For example, the profit-maximizing fee
will 	vary 	across 	geographic 	markets 	if 	there 	is 	geographic 	variation 	in
downstream demand; the upstream firm would require information on the
degree of this variation to set the correct fees. 	Second, if the downstream
(continued ... )
51

c. Vertical Restraints to Avoid Distortions From Upstream Market Power
When an upstream firm has market power,131 it will raise its price above
the competitive price and will reduce its output below the competitive level. If
this firm's output constitutes an input into the production of some "final" good,
the producers of this final good will attempt, to the extent permitted 	by
technology, to substitute other, competitively-priced inputs for the input that
is priced supracompetitively. 	The downstream firms' input choices become
"distorted" by the existence of a supracompetitive price for the monopolized
input. In such instances, the producer of the input can increase its profits if it
can eliminate this distortion while continuing to exercise market power. 132
Vertical 	integration 	is one 	means 	by 	which 	the distortion 	can 	be
eliminated. 	When the monopolist acquires a downstream user, it can increase
the combined profits of the two entities by eliminating this distortion; i.e., by
transferring the input to its downstream subsidiary at a competitive, rather than
supracompetitive price. The vertically integrated monopolist captures for itself
the increase in profit that is created by the .now-efficient choice of inputs.
Further, as long as there are other downstream firms making inefficient input
130(••• continued)
average 	cost 	curve 	is 	"U-shaped", 	rather 	than 	flat, 	then 	the 	lump-sum
contractual alternative fails to yield equivalent profits to the upstream firm.
See Blair and Kaserman, p. 75. 	Third, if the downstream demand curve is
subject to random variation, the franchise fee transfers all of the associated risk
from this variation to the distributor. 	If the distributors are risk-averse, this
requires that the upstream firm adjust the flat fee and the wholesale price to
redistribute the risk. These computations become complex if dealers differ in
their degree of risk aversion. 	See Tirole, p. 177.
151 We define "market power· in terms of the elasticity of the residual
demand curve facing the firm. 	The lower the value of this elasticity, the
greater the degree of the firm's market power. This residual demand elasticity
will be jointly determined by (I) the firm's 	market share, (2) the (price)
elasticity of the market demand function, and (3) the (price) elasticity of the
fringe supply. 	See Landes and Posner (1981).
152 See Schmalensee (1973); Warren-Boulton (1974); Mallela and 	Nahata
(1980); Westfield (1981); Quirmbach (1986); Also see Blair and Kaserman, ch. 4.
52

choices (because the monopolist continues to charge them a monopoly price),
there is a profit incentive for the upstream monopolist to acquire them as
well. l33
Whether programmers are using vertical restraints to resolve a problem
arising from input distortion is uncertain. 	The answer depends upon whether
(1) programmers have market power, and (2) cable operators can substitute
among different programming sources. Unless condition (I) holds, there can be
no monopoly price distortion, and thus no distortion-induced 	incentive to
integrate. 	If condition (2) does not 	hold 	(Le., 	if cable operators cannot
substitute less expensive programming), there is again no distortion-induced
incentive to integrate. IS.
Whether condition (I) holds is an empirical question on which we have
little information. 	We do not address it further. 	Concerning condition (2), it
may be plausible to characterize cable systems as using a variable proportions
production technology. 	For 'example, cable operators may offer subscribers
different "packages" (or "tiers") of service that are made up of different
collections of programming services. l36 	The lowest tier might consist of local
broadcast stations and one or more "superstations." 	The next tier ("expanded
basic") might consist of basic service plus channels like TNT, USA Network,
ESPN, and American Movie Classics. 	Higher tiers would have additional
channels (e.g., HBO). 	Clearly, an operator's choice of programs to include in
each tier may be determined in part by the price set by the programmer.
133 See Schmalensee (1973). 	Whether the complete monopolization of the
downstream industry would likely raise or lower the price of the final good is
discussed below.
134 	If 	downstream 	production 	is 	characterized 	by 	so-called 	"fixed
proportions," then any upstream market power can be fully exploited simply
by setting a monopoly price upstream. 	In this setting, incentives to vertically
integrate arise from some other source.
IS5 For a description of recent innovations in this area, see "New Fees Alter
'Basic' Idea of Cable TV," Wall Street Journal, January 23, 1990, p. Bl.
S3

Increases in the price of a program may give operators an incentive to shift it
into a higher-priced tier (replacing it in the lower tier with a less expensive
program) where it will have fewer subscribers, or perhaps to drop it altogether.
It may thus be plausible to apply the "variable proportions" model at least
to the downstream portion of the video marketplace. 	If there is market power
upstream, an incentive to integrate vertically (or impose other vertical controls)
may 	be 	present. 	The 	question 	would 	then 	arise 	as 	to 	the 	competitive
implications from adopting such controls. 	The answer depends upon (I) the
competitiveness of the downstream market (prior to the imposition of the
vertical 	restraints). 	(2) 	the 	elasticity 	of 	demand 	facing 	the 	downstream
producer,l36 and (3) the extent to which cable operators can engage in the sort
of substitution described above. 	If the downstream market is competitive
(before integration), its subsequent monopolization through complete vertical
integration would help consumers by eliminating the monopoly input-distortion
(Le .• since the program would be transferred at marginal cost, not a monopoly
price), but would hurt them by transforming a competitive market into a
monopoly market. 1S1 It is not possible to generalize about which effect is likely
to aominate. 1S1 If, by contrast, the downstream sector was already a monopoly
~efore the integration occurred, then (as discussed earlier) there is a greater
likelihood that the merger will benefit consumers.
136 If 	the 	upstream 	firm 	buys all 	of 	the 	downstream 	firms, 	thereby
monopolizing the downstream industry, the markup of final good price over
marginal cost will depend upon the elasticity of final good demand.
IS1 Complete vertical integration in this context means that the upstream
monopolist acquires all of the previously independent downstream firms.
lSi See Quirmbacb (1986).
54

Waterman (1987) analyzes the incentives to integrate vertically when the
upstream 	production 	is 	characterized 	by 	increasing 	returns 	to 	scale. 139
Waterman 	argues 	that 	the 	costs 	of 	programming 	(or, 	more 	generally,
"information products") are largely fixed; Le., once a firm assembles a package
of programs, 	the additional 	cost of 	making 	it available 	to an additional
sUbscriber is small. 	If true, this would mean that the marginal cost (per
subscriber) of programming is below average cost (per subscriber), implying that
the 	average 	cost 	of 	programming declines 	as 	the 	numbe~; of subscribers
increases.
For a programmer to cover its costs, the average revenue (per subscribe!')
must be at least as high as average total cost. 	Under Waterman's analysis, a
nonintegrated downstream firm will treat this average total cost as the relevant
upstream price when making its input choices; an integrated firm, by contrast, .
will treat the marginal cost as the relevant wholesale price when making its
input decisions. 	This, in turn, should give the integrated system an incentive
either to reduce its resale price or "to "favor" its own network. 	Such beha vior
could be both profit-maximizing and welfare-enhancing.
There are models other than those discussed above that analyze vertical
mergers under alternative market structures (e.g., oligopoly at both stages of
production) 	and 	alternative 	specifications 	of 	downstream 	substitution
possibilities. l40 	Though the detailed analysis is different, all of these models
yield a similar result: 	the effect ot vertical integration on final price cannot
be generalized; sometimes price will rise, sometimes it will fall.
The theories discussed above identify a number of situations where
1'0 In contrast to the "variable proportions" model analyzed above, Waterman
does not assume the existence of upstream monopoly, but rather assumes a
monopolistically competitive upstream market. 	The downstream market, by
contrast, is assumed to consist of a collection of spatially separated firms, each
ofwhich may have market power in some local market.
140 See Waterson (1982); Salinger (I988a); and Abiru (I988).
ss

vertical controls could potentially increase efficiency and improve consumer
welfare. Whether, and to what extent, these theories apply to the cable industry
is not known. 	In the ensuing section, we discuss how. vertical integration may
be used for anticompetitive purposes.
d. Vertical Integration as a Source of Entry Barriers
The 1984 Department of Justice Merger Guidelines discuss whether a
vertical merger might have anticompetitive effe~~s.141 The discussion focuses
on a product market in which, but for ease of entry, conditions are favorable
to the exercise of market power, and asks whether a vertical merger might
reduce entry so that market power could be exercised. The Guidelines'identify
three necessary (but not sufficient) conditions for this problem to exist. 	First.
the market where the market power would be exercised (the ·primary· market)
must be sufficiently conducive to anticompetitive behavior that the impact of
vertical integration in reducing entry would allow such behavior to occur.
Second, the degree of vertical integration subsequent to the merger must be so
extensive that an entrant into the primary market would also have to enter the
other market (the ·secondary· market). 	If substantial unintegrated capacity
remains in the secondary' market after the vertical merger, it is less likely that
the merger will facilitate an anticompetitive outcome. Third, the requirement
that a firm enter both the primary and secondary markets (rather than just the
primary) must make entry into the primary market significantly more difficult
and therefore less likely to occur. 142
141 See § 4.21. Also see Blair and Kaserman. PI'. 42-4. and Fisher and Sciacca
(1984), especially Pl'. 22-27.
142 We noted before that there may be contractual alternatives to vertical
integration. 	The same may be true here. 	For example, the effects of vertical
integration might occur through an arrangement whereby an upstream firm
grants a distributor an exclusive territory in exchange for exclusive distribution
(i.e .• the distributor carries only the upstream firm's product).
56

How might the necessity of vertically integrated entry create an entry
barrier? 	Economists have identified conditions under which this could occur.
Fisher and Sciacca (1984) provide an extensive discussion. 143 The possibilities
of greatest relevance to the cable television context are (1) when potential
entrants arc less efficient at the one stage of production than the other; (2)
when the minimum efficient scales (MES) of production at the different stages
are such that the addition of a fully integrated MES entrant adds more capacity
to the market than can be profitably absorbed;!" and (3) when the risks
associated with integrated entry raise capital costs. We discuss these possibilities
below.
i. Differential rates of efficiency
Let us suppose that the upstream ("primary") market is favorable to
noncompetitive behavior among incumbents, and therefore that the incumbents
arc 	already 	protected 	from 	entry. 	Suppose 	further 	that 	entry 	into 	the
downstream ("secondary") market is easy. 	Can the existing producers in the
primary market make entry even less likely by integrating into the secondary
markets?
By assumption, vertical integration is unlikely to make entry into the
primary 	market 	more difficult, since entry 	into the secondary 	market 	is
presumed easy. 	If a potential entrant into the primary mark~t could have
overcome whatever entry barriers existed before the incumbents integrated into
the primary market, then presumably the entrant would not be any more
deterred by a requirement that it also enter into the secondary market. 	This
Certainly would be true if the potential entrant into the primary market could
. 	143 See Fisher and Sciacca, pp. 22-27 for a detailed discussion; also see
§ 4.212 of the Merger Guidelines.
1" That is, the post-entry price would be below the average cost of the 	.
entrant, which will deter entry.
57

--- .
operate in the secondary market as efficiently as the incumbent 	primary
producers. 	However, if the potential entrant into the primary market was less
efficient than the incumbents, then perhaps primary entry ·could be deterred
if the incumbents' price was increased by no more than the difference in
efficiency in the secondary market. Note. however. that the incumbents' price
may not be increased if the potential entrant. even if not itself an efficient
producer 	in 	the 	secondary 	market. 	could 	contract 	for 	(or 	make 	other
arrangements) for entry into the secondary market by a producer as efficient
as the incumbents. 	Primary market entry would then be deterred only if
making these alternative arrangements were 	more costly than 	the vertical
integration by the incumbents. 	This seems somewhat unlikely, but it remains
an empirical Question.
Fisher and Sciacca also discuss whether vertical integration could reduce
two-stage entry when "high" barriers to entry exist at each stage. They consider
a situation where the "least disadvantaged" potential entrant into one stage is
different 	f.rom 	the "least disadvantaged" entrant into the other. 	Whether
vertical integration by incu.mbents reduces the likelihood that these firms would
enter is an empirical issue: 	is it more difficult for the potential entrants to
replicate what the incumbents have done. which is to combine "economic agents
with comparative advantages at [one stage] with agents with comparative
advantages 	at 	[another]?d45 	It 	is 	difficult 	to ~eneralize. a 	priori, 	how
burdensome this coordination problem is likely to be; the answer will depend
upon the particular circumstances of any given case. l46
141 Fisher and Sciacca. p. 24.
146 It is tempting to conclude that this coordination problem is likely always
to be insurmountable. since firms cannot contract for a joint venture or merger
until the· firms are first created, and in turn these firms will not be created
unless the efficiencies of vertical integration can be exploited. 	This view
overlooks the possibility that entry often occurs through diversification by
existing firms. 	Thus. one can conceive of situations where two exi~ting firms
(continued ... )
58

ii. Differellt millimum efficiem scales
The Merger Guideiines 147 discuss whether substantial scale economies at
one stage (e.g., the primary market) could, if combined with the necessity of
two-stage entry, create an entry barrier. The idea is that by requiring two-stage
entry the incumbents in the primary market could force a potential entrant to
(1) add more capacity to the market at the one stage than can profitably be
absorbed (i.e., thus forcing price below the entrant's average cost), or (2) enter
one stage at an inefficiently small scale l4 ' (thus incurring higher average costs
than the incumbent producers). 	Either could allow the incumbents to raise
price and profits without attracting entry which would have occurred absent.
the integration by the incumbent producers.
Fisher and Sciacca illustrate this possibility with an example. 14g Suppose
that the primary and secondary markets are unintegrated, and that the primary
market can accommodate exactly three minimum efficient scale (MES) single-
plant firms, and the secondary market can accommodate exactly 12.5 MES
single-plant firms. 	A competitive equilibrium would exist when upstream
output equalled three times the .upstream MES, and downstream output 12.5
times downstream MES. There could be at most three MES upstream firms, and
12 downstream MES fi~ms, in this eQuilibrium. l50
146(.•• continued)
form a joint venture, with one diversifying into manufacturing, the other into
distribution. 	.
147 See § 4.212. 	Also see Fisher and Sciacca, pp. 24-5.
148 Alternatively, the entrant could add a MES plant, but operate it at an
inefficiently low level of output.
14g See Fisher and Sciacca (1984), pp. 24-5. 	The situation described can be
considered a special case of a more general problem in which an incumbent
seeks to deter entry through investments in excess capacity. See Gilbert (1986)
for a discussion of these models.
150 Each downstream firm would produce slightly more than the MES output.
59

Although the upstream stage could accommodate up to three firms, there
need not be three firms in equilibrium. 	If long-run costs are flat after MES is
reached, one could have, for example, two firms, each producing 1.5 the MES
output. 	As Fisher and Sciacca point out, however, these two firms could not
successfully raise price above cost, as the upstream market could accommoda te
a third MES entrant.
The situation might differ, however, if there were complete vertical.
integration. 	Suppose that there are two upstream firms, each owning six
downstream plants, and that each refuses to purchase from de novo upstream
entrants, 	thus requiring an entrant 	to 	build one 	upstream 	plant and 	one
downstream plant. 	Given the specific market conditions outlined above, the
addition of this new downstream plant would add more capacity to the market
than could be profitably absorbed if the market price equalled minimum
average cost. Unless the entrant expected above-competitive pricing to prevail
after entry, it might be deterred from entering. 	This would make conditions
more favorable for successful collusion among the incumbents.
iii. Wh~n Added Risk Raises the Cost of Capital
The Merger Guidelines l51 observe that because two-stage entry requires
a larger investment than one-stage entry, vertical integration could increase
entry barriers. 	The conditions under which this could occur are carefully
defined. 	The Guidelines note that increasing the magnitude of the investment
necessary for entry will not necessarily act as an entry barrier. 	A problem is
likely to arise only if there is an asymmetry in capital costs between incumbents
and potential entrants. Such asymmetries could arise because the assets required
for entry are long-lived and ·sunk" (Le., have little or no value should entry
151 See § 4.212 of the Guidelines; also see Fisher and Sciacca, pp. 25-6.
60

prove unsuccessful).lU Of course, an entrant attempting only single"-stage entry
would also have to make these sunk cost investments. 163 To the extent that sunk
cost investments discourage entry, this will be true whether or not incumbents
are vertically integrated. The relevant Question is whether the need to enter as
a vertically integrated producer differentially increases the risk of entry. Fisher
and Sciacca note 1" 	that there exists little evidence (and that which exists is
mixed) whether capital requirements create entry barriers. 	Little empirical
research has been conducted on the extent to which vertical integration raises
capital-market entry barriers.
e. Vertical Integration to Avoid Rate-of-Return Regulation
An incentive to integrate vertically may exist when a firm possessing
market power is subject to rate-of -return regulation. 	Should cable systems
become subject to comprehensive rate-of -return regulation, the effectiveness
of this regulation may be adversely affected by extensive vertical integration.
Under effective rate-of-return regulation, the regula tor permi ts "the fi rm
to 	charge 	a 	price 	that 	covers 	its 	variab'e 	costs 	and 	provides 	it 	with 	a
"competitive" return on its investment. A drawback to rate-of-return regulation
is that the regulator is often dependent upon the firm for cost information.
This may allow the firm to exaggerate its reported costs and thus receive a
higher price (rate-of -return) than if the regulator k"new the firm's true costs.
112 See Baumol, Panzar, and Willig (1982), pp. 296-301, for a theoretical
exposition 	of how 	sunk cost 	investments create capital 	cost differentials
between incumbents and potential entrants.
us The Guidelines argue that a capital cost barrier could arise if suppliers
of capi"tal believe that a firm likely to be an efficient entrant at one stage does
not possess the skills and knowledge necessary for 	entry at another, and
accordingly demand compensation in the form of a risk premium. 	Fisher and
Sciacca argue persuasively (p. 24, n. 52) that 	this 	is really an efficiency
differential as described above.
1" Fisher and Sciacca, p. 26.
61

By vertically integrating, the regulated firm may be better able to pursue
this strategy. 	The reasoning is straightforward. 	If the firm purchases inputs
on external markets, the regulator may be able to observe the prices actually
paid or at which purchases could be made. If so, the firm will find it difficult
to misrepresent 	the 	price paid for 	inputs. 	If 	the 	inputs are 	transferred
internally, the firm may be able to inflate the transfer prices. This inflates the
firm's rate base, and thus the price that it will be permitted to charge to earn
its allowed rate-of _return. 1SS
B. Empirical Evidence on Vertical Inteeutlon In the Video Marketpll~ce
.
Although vertical 	integration 	between 	program suppliers and cable
systems is currently the focus of a 	heated policy debate, it has 	not been
subjected to extensive empirical analysis. 	We are aware of only two such
studies: Klein (1989), and Salinger (1988b). 	Salinger attempted to determine
whether vertically integrated firms are less likely to carry the programs of
. nonintegrated suppliers'. 	Klein also attempted to address this issue, as well as
the question of whether integrated firms are unwilling to sell their programs to
alternative video distributors (e.g., MMDS and SMA TV). These papers provide
valuable 	information 	on 	behavioral 	difrerences 	between 	integrated 	and
nonintegrated cable systems. 	Below we summarize these studies.
155 See Blair and Kaserman (1983), pp. 110-14. 	Along these lines, the FTC
has challenged vertical acquisitions when one of the parties to the transaction
was subject to rate-of-return regulation. 	For example, on June 25, 1986, the
Commission accepted a consent order with Occidental Petroleum Corporation
(a producer and marketer of oil, gas, and coal) and MidCon Corporation (an
owner and operator of natural gas pipelines), FTC Okt. No. C-3191. 	In the
accompanying complaint, the Commission alleged that Occidental's acquisition
of the stock or assets of MidCon violated Section 7 of the Clayton Act, as
amended (15 U.S.c. § 45), and Section 5 of the Federal Trade Commission Act,
as amended (15 U.S.C. § 18).
62

J. The Salinger Study
Salinger sought to determine whether vertical integration affects (I) the
decision to carry a particular premium pay service (HBO, Cinemax, Showtime,
and 	The 	Movie 	Channel), (2) 	the 	price 	of 	premium 	services charged 	to
consumers, (3) the price of basic service, and (4) the number of premium
services offered. Salinger used data (drawn from the 1987 Television and Cable
Factbook) on a sample of 217 integrated and nonintegrated cable franchises.
The franchises included in the analysis were all owned by the twenty largest
multi-system operators (MSOs). Controlling for a number of other factors that
might explain prices and program offerings,166 Slllinger first estimated "probit"
.
equations for each of the four premium movie services. 167 These equations
estimate 	the 	probability 	that 	a 	particular cable franchise 	would 	carry 	a
particular movie channel. 	Included as explanatory factors are a dummy
variable that indicated whether the franchise was owned by Time, Inc. (the
parent of HBO and Cinemax, and 80 percent owner of American TV and
Communications Corp. (ATC), an MSO), and a dummy variable that indicates
whether the franchise was owned by Viacom (the owner of Showtime and The
Movie Channel). The results indicated that, compared with the other franchises
in the sample, Viacom cable franchises were less likely to carry HBO and
Cine max. UI Similarly, ATC franchises were more likely to carry Cinemax, and
lie The explanatory variables in Salinger's equations are number of homes
passed, ranking of size of the television market in which the cable franchise
competes, age of system, and the ratio of homes passed to total miles of cable.
157 Probit estimation is a statistical technique frequently used when the
dependent variable is discrete (rather than continuous). 	In Salinger's model,
the dependent variable takes on a value of "I" (when the cable system carries
a particular programming service), or "0" (when the system does not carry the
service).
lSI Viacom systems are not more likely to carry The Movie Channel than
nonintegrated systems.
63

less likely to carry The Movie Channel (but not Showtime), than the other cable
franchises in the sample. lSO
Although 	he controlled for 	a 	number of potential determinants of
programming choices. Salinger did not control for the channel capacity of the
franchise in the reported regressions. l60 The Klein study (discussed below)
found that channel capacity was significantly related to the number of premium
movie services offered by a cable 'franchise. l61
In his price regressions. l62 Salinger obtained no evidence that vertical
integration affected the price of pay services if looked at alone. 	He did find,
however. that ATC charged a lower price for basic and "extended basic" services
than did nonintegrated franchises. Viacom, by contrast, charged a higher price
for basic services, but was less likely to have an additional charge for higher
levels (i.e., "extended") of basic service.
Salinger's 	results 	suggest 	that 	A TC 	and 	Viacom 	franchises 	were
somewhat less likely than other cable franchises to carry at least three of the
four movie services, but somewhat more likely to' carry at least two services. l63
Salinger believes that if the "number of services· regressions had been run using
the full sample (i.e.. including the Viacom observations that were omitted due
159 All of the A TC franchises in the sample carried HBO, and all of the
Viacom franchises used in the estimation procedure carried Showtime. Salinger
excluded several Viacom observations that did not carry Showtime because of
missing data. 	See Salinger, p. 17.
160 Salinger ran regressions that include channel capacity, but does not
report these results. He argues that channel capacity is an endogenous variable,
and thus should not be included as an explanatory variable. See Salinger, p. 17,
note 22.
161 According to the 1988 Television and Cable Faclbook, about 37 percent of
all cable franchises had a capacity of fewer than 30 channels. 	Only 7 percent
had more than S4 channels.
162 These equations had essentially the same specification as the probit
equations.
163 This latter result was significant at the 86 percent significance level.
64

to missing data), the results would likely have shown that ATe and Viacom
have no greater propensity to carry at least 	two movie services than did
nonintegrated franchises.
2. The Klein Study
Klein attempted to provide a comprehensive overview of the causes and
consequenccsof vertical integration in the cable industry. 	Like Salinger, hc
analyzed the behavior of firms that both produce programming for cable
tclevision and op~rate their own cable systems. 	Klein asked 	whether (I)
vertically 	intcgrated 	cable 	systems 	showed 	a 	preference 	for 	their 	own
programming (were less likely to distribute programs produced by others), and
(2) vertically integrated programers showed a preference. for their own cable
systems (were less likely to sell programs to others). 	Klein had less empirical
evidence bearing on the second issue than on the first. Accordingly, we review
only his analysis of the first issue.
Klein tested for the first preference by examining the programming
choices of a ran<;lom sample of 400 cable franchises. 	Klein first examined
whether the cable franchises having an ownership interest in a particular
program 	supplier 1e4 	were 	more 	likely 	to 	carry 	that 	service 	than 	were
unaffiliated cable systems. 	Unlike Salinger, Klein did not use a multivariate
analysis to control for other determinants of programming choices; instead, he
simply computed ·carriage rates· us, 	for each programming service (i.e., he
computed one carriage rate for affiliated systems, one for unaffiliated systems).
To tbe extcnt that omitted factors are import~nt determinants of programming
choices, and are correlated with vertical integration, Klein's findings will be
164 Klein examined ·carriage rates· for 28 different programming services
(20 ·basic· networks, 8 ·premiuin· networks).
165 The ·carriage rate· for a programming servicc is simply the percentage
of cable systems offering that service.
65

biased in an unknown direction. 	Like Salinger, however, Klein found that
vertically integrated systems were more likely to carry their own programs than
were nonintegrated systems. 	The average carriage rate for affiliated systems
was about 15 percent higher than for unaffiliated systems. l66 	That is, on
average, about 88 percent of integrated systems carried the programming service
with which they were integrated; these same programming services were carried
by about 73 percent of unaffiliated systems.
Klein next analyzed the programming choices of the four most vertically
integr.ated MSOS 181 (TCI, ATC, Viacom, and Cablevision). The objective was to
determine if these MSOs avoided carrying the programming services in which
they did not have ownership interests. He first determined which programming
services were unaffiliated with at least one of these four MSOs. 	For example,
neither TCI, ATC. nor Cablevision had an interest in Music Television (MTV).
Next, 	for 	each 	of 	the 	programming services satisfying 	this 	criterion, 	he
computed the carriage rate for these unaffiliated MSOs. 	In the case of MTV,
for example, 98 percent of the cable franchises owned by TCI, A TC, and
Cablevision carried MTV. 	Klein then computed. for each of the selected
programming services, the carriage rate on- completely nonintegrated cable
systems (i.e., cable systems that 	were unaffiliated 	with any programming
service). 	MTV, 	for 	example, 	was 	carried 	on 	about 	93 	percent 	of 	these
completely nonintegrated systems. lea
These computations 	were performed for 24 different programming
services. Overall, Klein found that the rate at which vertically integrated MSOs
188 There is only about a 2 percent chance that this difference could have
been caused by sampling error. 	See Klein, p. 37, note 41.
181 The extent of vertical integration was measured by the number of
networks in which the MSO had an equity interest.
188 As before, this analysis did not control for any other explanatory factors
(e.g., channel capacity of the system). 	-
66
--- .

carried an 	unaffiliated 	programming service exceeded 	the 	rate 	at 	which
completely nonintegrated systems carried those same programming services. l69
The average carriage rate differential was about 5 percent. l70
As 	we observed earlier, Klein did 	not perform a 	full 	multivariate
analysis of the decision to carry a particular network. 	However, as Klein
said,l71 his results may have somewhat greater general applicability than do the
results of studies confined to carriage rates for the four pay movie channels.
Klein also performed a series of regressions that can-be compared with
the regressions presented in Salinger's Table S. 	Salinger tested for whether
vertical integration affects the number of pay services offered to subscribers,
and found 	that ATC and Viacom systems were somewhat less likely than
nonintegrated systems to carry at least three pay channels. 	Klein performed
three sets of regressions in which the dependent variables were (i) total number
of 	pay 	and 	basic 	programming 	services 	carried, 	(ii) 	number 	of 	basic
programming services carried, and (iii) number of premium programming
services carried. Controlling for channel capacity, Klein examined whether the
degree of vertical integration (measured by the number of programming services
in 	which 	the 	system 	had 	an 	equity 	interest)l72 affected 	the 	number 	of
programming services carried. 	[n a-Il three equations Klein found a strong,
leG [n 20 of the 24 	networks analyzed, 	the 	vertically 	integrated (but
nonaffiliated) 	MSOs 	had 	higher 	carriage 	rates 	than 	did 	the 	completely
nonintegrated cable systems. The exceptions were Nick at Nite, The Nashville
Network, Video Hits· I. and The Movie Channel.
170 There is only a I percent chance that this differenc; ts--due to sampling
error. 	See Klein, p. 42, note 44. 	.
171 See Klein, p. 45.
112 [n the equation explaining premium channel carriage~ the explanatory
variable was the number of affiliated premium programming services. 	In the
basic channel equation, the explanatory variable was the number of affiliated
basic programming services.
67

statistically significant positive relationship between vertical integration and
the number of programming services carried.
This 	result 	is not 	necessarily 	inconsistent 	with Salinger's 	findings.
Salinger asked a very specific question: 	how do the programming choices of
Viacom and A TC diffe.r from non- Viacom, non-A TC systems with respect to the
carriage of four particular premium movie channels? Salinger did not explore
other issues, such as (1) whether there were differences among Viacom, ATC,
other vertically integrated MSOS,l73 and completely nonintegrated MSOs with
respect to the carriage of the four premium movie services, and (2) whether
there were differences among Viacom, ATC, other vertically integrated MSOs,
and completely nonintegrated MSOs with respect to other program services (e.g.,
Disney). It appears, for example, that A TC systems may be more likely to carry
Cinemax, and less likely to carry The Movie Channel,· than non-A TC systems;
it is not yet known, however, whether vertically integrated systems as a group
tend to carry. more premium programming services, especially when the carriage
of all premium programming services (and not only pay movie channels) is
taken into account.
J. How Should These Findings Be Interpreted?
Both Sali.nger and Klein find that vertical integration pia ys a role in
affecting 	program choices. 	One 	finding common 	to 	both 	studies 	is 	that
integrated ,cable 	systems 	have 	a 	greater 	propensity 	to 	carry 	their 	own
programming 	services 	relative 	to 	unaffiliated 	systems. 	This 	finding 	is
consistent 	with 	both 	pro- 	and 	anticompetitive 	motivations 	for 	vertical
integration. 	Whatever the motivation for vertical integration -- whether to
reduce transaction costs, obtain a lower transfer price for pay programming, or
173 Many of the non-ATC, non- Viacom systems in Salinger's sample almost
certainly exhibited some degree of vertical integration, as his sample consisted
entirely of systems owned by the 20 largest MSOs.
68

facilitate anticompetitive behavior -- a preference for one's own program should
be observed. 	As Salinger states, "[t)he conclusion that can be drawn with the
most confidence from [my) results is that vertical integration does matter."IH
However, while acknowledging that it is uncertain whether vertical integration
is, on balance, 	pro- or anticompetitive, Salinger does 	regard 	the 	reduced
likelihood that a vertically integrated system will carry at least three premium
movie services as a detrimental effect of vertical integration.
We believe that Salinger's results, when viewed alone, and more so when
viewed in conjunction with Klein's findings, shed light on the issue of vertical
integration. ATC appears to have charged lower prices for basic and extended
basic services than did nonintegrated systems. The prices of the different pay
services did not depend upon the extent of vertical integration. There is some
indication that ATC and Viacom systems were 'more likely than nonintegrated
systems to offer at least two pay services; we note, however, that Salinger has
Questioned whether this finding would hold if a larger sample had been used.
On the negative side, Salinger's paper suggests that Viacom charged a
higher price for basic service than did nonintegrated systems. l75 It also appears
that A TC and Viacom may have been less likely than nonintegrated systems to
carry at least three of the four primary pay movie channels. 176 It is not clear,
however, that. this latter result is unambiguously detrimental to consumers.
First, this may only reflect the benefits of lower internal transfer prices, as
discussed previously.l71 Second, unless a system has unused channels, the fact
that it is not showing all of the four programming services studied may mean
174 Salinger, p. 23.
175 Viacom was less likely, however, to impose an additional charge for
extended basic service.
176 It might be useful to know if this result would have changed had the
analysis controlled for channel capacity.
171 See the discussion of Waterman (1987), above.
69

that 	some 	other 	programming service 	has 	been 	substituted 	for 	it. 	This
substitution may simply reflect differences in consumer tastes across geographic
markets, although we have no data to suggest this is true.
For example, consumers in the ATC and Viacom markets may not have
preferred to have all four programming services offered if it meant that
something of higher value was displaced. It is conceivable that such consumers
might (for example) have preferred American Movie Classics l78 (which mainly
offers movies from the 1930s, 1940s. and 1950s) and the Disney Channel l79
(which offers "family" entertainment) instead of (say) Cinemax and The Movie
Channel. Klein's data are consistent with this possibility, although as we noted
earlier, Klein did nOt perform a full multivariate analysis of carriage decisions.
Almost half of the A TC and Viacom systems in his sample carried American
Movie 	Classics. 	as 	compared 	with 	about 	one-fifth 	of 	the 	completely
nonintegrated systems; 97% of the TCI, ATC, Viacom, and Cablevision systems
carried Disney, as compared 	with about 93% of the totally 	nonintegrated
systems. ISO 	More generally, Klein's regression analysis 18l suggests that higher
levels 	of 	vertical 	integration 	(as 	measured 	by 	the 	number 	of 	premium
programming services in which a system has an ownership interest) exhibit a
positive and statistically significant relationship to the number of premium
stations offered on the system.
178 AMC is owned by Cablevision, TCI, and United.
179 Disney is not affiliated with any cable systems.
180 See Klein, table 7, p. 41.
181 See Klein's table II. appendix.
70

C. Summary
The theoretical literature provides a number of circumstances under
which vertical integration or contr.ol can promote competition and consumer
welfare. 	There also are some circumstances under which vertical integration
can harm consumers and competition. This literature does not yield definitive
conclusions about the 	impact of 	vertical 	integration or controls on 	video
markets. The competitive implications of vertical integration and other vertical
controls 	in 	any 	particular 	instance 	likely 	depend 	upon 	the 	economic
circumstances specific to that instance. The limited empirical evidence on the
effects of vertical integration in the video marketplace suggests that cable
franchises' 	choice 	of 	programming 	is 	affected 	by 	the 	extent 	of 	vertical
integration. 	Some 	of 	this 	evidence 	suggests 	positive 	aspects 	to 	vertical
integration; some suggests negative aspects. 	While one cannot say. on the basis
of the existing literature, that vertical integration in the video marketplace is
unambiguously procompetitive, we also cannot say, drawing on this literature,
that 	vertical 	integration 	in 	the 	video 	marketplace 	is 	unambiguousl y
anticompetitive.
7i

VIII. 	The Application of the Antitrust Laws to Cable Television
Paragraph 22(e) of the NOI asks whether the Cable Act should 	be
amended to "establish structural rules, or to encourage antitrust action, to
prevent mergers among cable systems competing in the same local market." The
NOI further asks whether "there is a need to amend the Cable Act to discourage
or forbid agreements (with or without franchise authority permission) that
enable competing cable systems to divide cable service areas and thus avoid
competing against each other in localities where two or more systems are
authorized." 	Currently, the Cable Act is silent regarding antitrust policy; it
neither exempts system operators from 	the antitrust laws nor incorporates
specific standards. regarding competition.
The antitrust laws (Le., the Sherman Act and the Clayton Act) and the
Federal Trade Commission Act reach all trade or commerce in or affecting
interstate commerce. 182 	In fact, the scope of these statutes extends to the full
constitutional power of Congress under the Commerce Clause. 18 :! 	Moreover,
federal regulation of a particular industry does not exempt the industry from
the antitrust laws except to the extent that there is a "clear repugnancy" between
the regulation and the antitrust laws. l ... There appears to be no such conflict
between 	the antitrust 	laws and 	either 	the Cable 	Act or 	local 	regulation.
Therefore, courts have applied the full force and effect of the antitrust laws to
the cable industry.186
182 IS U.S.C. § I (Sherman Act); IS U.S.C. § 12 (Clayton Act); IS U.S.C. §
45(a)( l) (FTC Act); McClain v. Real Estate Board. Inc., 444 U.S. 232 (1980) (in or
affecting commerce standard).
lU Hospital Building Co. v. Trustees of Re."C Hospital. 425 U.S. 738, 743 n. 2
(1976); United States v. South-Eastern Underwriters Ass'n., 322 U.S. 533, 558-59
(1944).
184 Georgia v. Pennsylvania Railroad, 324 U.S. 439, 457 (1945).
186 See, e.g .. Central Telecommunications. Inc. v. TCI Cablevision. Itrc.. 800 F.2d
711,'726 (8th Cir. 1986), cert. denied. 480 U.S. 910 (1987) (holding that the
Sherman Act applied to a cable system subject to local rate regulation).
72

The first specific issue raised 	by paragraph 22(e) concerns mergers
between cable systems competing in the samc local market. 	Section 7 of the
Clayton Act l86 currently prohibits all acquisitions of stock, ·share capital or
assets that may substantially lessen competition, or tend to create a monopoly.
This prohibition is broad in scope, and includes aU mergers without regard to
their legal form. 181
We do not believe that the cable industry possesses any characteristics
that 	would 	necessitate 	the 	enactment 	of 	unique 	statutory 	standards 	for
adjudicating the legality of mergers between competing cable systems. 	The
statutory provisions· of Section 7 of 	the Clayton 	Act, 	and 	the 	principles
articulated in the 1984 Department of Justice Merger Guidelines, are sufficiently
general to address any competitive issues that are likely to arise in connection
with a merger between competing cable companies. The Clayton Act has in fact
already 	been 	applied 	to 	acquisitions 	of 	local 	cable 	systems 	by 	their
competitors. 111 	We also note that vertical controls (see § VII of the comment)
are subject to scrutiny under the antitrust laws.l~
The 	NOI 	also 	invites comment on 	the question 	of 	multiple 	cable
franchises agreeing to divide the franchised territory among themselves, thereby
avoiding direct competition. Section I of the Sherman Act currently prohibits
all 	contracts, 	combinations, 	and 	conspiracies 	in 	restraint 	of 	trade 	or
commerce. UIO The Supreme 	Court 	has 	long 	held 	that 	agreements 	among
lee IS U.S.C § 18.
181 United States v. Philadelphia National Bank. 374 U.S. 321 (1963).
188 Cable Holdings. Inc. v. Home Video. Inc .. S72 F. Supp. 482 (N.D. Ga. 1983).
1~ See, e.g .. Continental TV. Inc .. v. GTE-Sylvania. Inc .. 433 U.S. 36 (1977), and
Monsanto v. Spray-Rite Service Corp .. 465 U.S. 752 (1984).
190 15 U.s.C. § 1.
73

competitors to divide a territory to avoid competition is illegal per se. llH 	In
Affiliated Capital Corp. v. City of Housto",192 the court specifically held that a
division of a franchise territory by cable operators was a sufficient basis for
finding a violation of Section 1 of the Sherman Act.
The Supreme Court applies the' per se rule to a division of a market by
competitors because it views the division as "conclusively presumed to be
unreasonable and therefore illegal."lga 	The rule serves to reduce uncertainty
about the legality of the proscribed conduct and reduces enforcement costs by
"avoiding an incredibly complicated and prolonged economic investigation ...
to determine ... whether a particular restraint has been unreasonable."l~
There 	are 	other 	manifestations 	of 	anticompetitive 	conduct 	not
specifically identified in the NOI that have been previously addressed by the
courts. 	For example, 'situations have arisen where the franchising system was
held to have been used to facilitate anticompetitive behavior. 	Although local
governmental units have a recognized interest in regulating access to the public
right-of-way,105 	some 	franchising 	authorities, 	together 	with 	incumbent
operators, may have used the cable franchising process to restrict competition
in violation of the Sherman 	Act. 101 	The history of cable franchising 	in
101 United States v. Addyston Pipe & Steel Co., 175 U.S. 211 (1899); Ford Motor
Co. v. Webster's Auto Sales. Inc .. 361 F.2d 874 (1st Cir. 1966).
102 735 F. 2d IS55 (5th Cir. 1984) (en bane), cert. denied 474 U.S. 1053 (1986).
1"1 See Northern Pacific Ry. Co. v. United States. 356 U.S. 1,5 (1958).
104 [d.
IN See. e.g .. Video 	International Production. 	Inc. 	v. Warner 	Annex Cable
Communications. Inc .. 858 F.2d 1075, 1081 (5th Cir. 1988), cert. denied, 109 S.Ct.
1955 (1989).
101 	Although 	franchising 	authorities 	are 	legally 	restrained 	by 	the
prohibitions of the antitrust laws, such authorities, as units of local government,
may not be sued for damages. 	15 U.S.C. § 35 (a). 	Injunctive and declaratory
relief, however, may be obtained from units of local government.
74

Sacramento, California may be illustrative. 197 There, the city granted the initial
franchise to an affiliate of a large multiple system operator. Another firm later
applied for a franchise to "overbuild" the city, but its franchise application was
denied. 	The potential overbuilder then brought suit against the city and the
incumbent ·cable operator, alleging, inter alia, a violation of Section 1 of the
Sherman Act.u~, 	The jury found, and the judge agreed, that the city had
conspired with the incumbent operator to exclude the overbuilder in exchange
for increased cash payments and provision of free cable services to the city
government. l99
Not only may the franchise power be employed to exclude overbuildcrs,
it may also be used against SMATVs, despite the fact that SMATVs arc outside
the scope of the franchising process. 200 	This possibility is illustrated by the
experience of Dallas, Texas. There, the city granted an initial franchise to an
affiliate of a large multiple system operator. 201 	The franchise agreement
provided for the largest franchise fee permitted under federal law (5 percent)
and stated that "no CATV system shall ... be allowed to operate within the City
without a CATV franchise." 	Pursuant to this provision, the city initiated a
campaign against SMA TV operators within its territory.
191 For a more expansive factual account, see Pacific West Cable Co. v. City
0/ Sacramento. 672 F. Supp. 1322 (E.D. Cal. 1987).
191 Id. at 1325.
199 	[d. at 1328. 	The court adopted this finding of fact from the jury's
special verdict despite its decision to dismiss the plaintiff's antitrust claims
against the city as immune under the state action doctrine. Id. at 1-325. See also
Pre/erred Communications. Inc.v. City 0/ Los Angeles. 754 F.2d 1396, 1411-15 (9th
Cir. 1985) (interpreting California statutes to be sufficiently specific so as to
constitute a "clearly articulated and affirmatively expressed" state policy to
displace competition in the cable television industry), a/I'd on other grounds.
474 U.S. 979 (1986).
200 47 U.S.c. § § 541-42.
201 For a more expansive factual account, see Video luternational Production,
858 F.2d at 1077-80.
75

This campaign was apparently abandoned when it became clear' to the
city and the incumbent franchisee that the city lacked franchise authority over
cable companies that operated without use of the public right-of-way. 	At this
point, however, the city and the franchised cable operator together interpreted
,the city's zoning ordinances in a manner that would have eliminated 	the
SMA TV operators. 	The largest SMA TV operator in Dallas then brought suit
against the franchised operator and the city, alleging, inter alia, a violation of
§ I of the Sherman Act. 202 The Court of Appeals affirmed the jury's finding
that the city had violated the Sherman Act by conspiring with the franchised
operator to eliminate SMATVs in' exchange for the cash payments mandated
under the franchise agreement. 20S
As Section III of this comment explained, restrictions on entry into cable
markets 	are 	not 	necessarily 	anticompetitive; 	in 	theory, 	one 	can 	identify
circumstances under which a franchising authority might serve consumers'
interests 	by 	protecting 	an 	incumbent 	cable' system 	from 	competition.
Nonetheless, 	we 	suspect 	that 	only 	rarely 	will 	such 	circumstances 	arise.
Therefore, any modification to the franchising process should be carefully
designed so that it does not result in cable franchising authorities receiving
blanket immunity from the antitrust laws.
202 {d. at 1080-81.
203 {d. at 1085-86.
76

References
M 	Abiru, 	"Vertical 	Integration, 	Variable 	Proportions 	and 	Successive
Oligopolies," Journal of Industrial Economics 36 (1988),315-25.
J. 	Acton 	and 	I. 	Vogelsang, 	"introduction 	to 	Symposium 	on 	Price-Cap
Regulation," RAND Journal of Economics 20 (1989), 369-72.
J. Baker and T. Bresnahan, "Estimating the Elasticity of Demand Facing a Single
Firm: Evidence on Three Brewing Firms," Stanford Research Paper No. 54,
Stanford: Stanford University. Department of Economics. 1984.
J. Baker and T. Bresnahan. "The Gains From Merger or Collusion in Product-
Differentiated Industries," Journal of Industrial Economics 33 (1985), 427-44.
J. Baker and T. Bresnahan. "Estimating the Demand Curve Facing a Single
Firm," International Journal of Industrial Organization 2 (1988), 282-300.
W. Baumol, J. Panzar. and R. Willig. Contestable Markets and the The'ory of
Industry Structure. 	New York: 	Harcourt Brace Jovanovich. 1982. 	'
D. Besanko, S. Donnenfeld, and L. White. "Monopoly and Quality Distortion:
Effects and Remedies." Quarterly Journal of Economics 102 (1987). 741-67.
D. Besanko. S. Donnenfeld, and L. White. "The Multiproduct Firm, Quality
Choice, and Regulation," Journal of Industrial Economics 36 (1988), 411-29.
R. Blair and D. Kaserman. Economics of Vertical I,ttegration and Control. .New
York: 	Academic Press. 1983.
R. Bork, The Antitrust Paradox. 	New York: 	Basic Books, 1978.
R. Braeutigam, "Optimal Policies for Natural Monopolies," in R. Schmalensee
and R. Willig. Handbook of Industrial Organization. II. fl. 	Amsterdam: 	Elsevier,
1989.
T. Brennan, "Cross-Subsidization and Discrimination by Regulated Monopolists,"
U.S. Department of Justice Economic Analysis Group Working Paper No. EAG
87-2, 1987.
T. Brennan. "Regulation by Capping Prices." Journal of Regulatory Economics I
(1989). 133-47.
W. Brock and D. Evans, "Creamskimming" in D. Evans ed., Breaking up Bell:
Essays in Industrial Organization and Regulation. 	New York: 	North-Holland.
1983.
L. Cabral and M 	Riordan, "Incentives for Cost Reduction Under Price Cap
Regulation." Journal of Regultuory Economics I (1989),93-102.
R. Dansby and R. Willig. "Industry Performance Gradient Indexes." American
Economic Review 69 (J 979), 249-60.
H. Demsetz. "Why Regulate Public Utilities?," Journal of Law and Economics II
(1968), 55-65.
77

D. Evans and J. Heckman, "A Test for Subadditivity of the Cost Function with
an Applicat·ion to the Bell System," American Economic Review 74 (1984), 615-
23.
A. 	Fisher 	and 	R. 	Sciacca, 	"An 	Economic 	Analysis 	of 	Vertical 	Merger
Enforcement Policy,'" Research in Law and Economics 6 (1984), 1-134.
L. Froeb and G. Werden, "Market Delineation and the Merger Guidelines: 	The
Role of 	Residual 	Demand 	Estimation," draft, U.S. 	Department of Justice
Economic Analysis Group, February 1990.
R. Gilbert, "Pre-emptive Competition," in J. Stiglitz and F. Mathewson, eds.,
New Developments in the Analysis 0/ Market Structure. 	Cambridge: 	MIT Press,
1986.
J. Haring and E. Kwerel, "Competition Policy in the Post-Equal Access Market,"
Federal Communications Commission, opp Working Paper No. 22, 1987,
P. Ippolito, Resale Price Maintenance (Economic Evidence From Litigation J, Bureau
of Economics Staff Report to th.c Federal Trade Commission, April 1988.
B. Klein, "The Competitive Consequences of Vertical Integration in the Cable
Industry," June 1989.
B. Klein, R. Crawford, and A. Alchian, "Vertical Integration, Appropriable
Rents, and the Competitive Contracting Process," Journal 0/ Law and Economics
21 (1978),297-326.
B. 	Klein 	and 	K. 	Murphy, 	"Vertical 	Restraints 	as 	Contract 	Enforcement
Mechanisms," Journal 0/ Law and Economips 31 (1988),265-98.
W. Landes and R. Posner, "Market Power in Antitrust Cases," Harvard Law
Review 94 (1981), 937-83.
K. Leffler, "Ambiguous Changes in Product Quality," American Economic Review
72 (1982), 956-67.
P. 	MaUeia 	and 	B. 	Nahata, 	"Theory 	of 	Vertical 	Control 	With 	Variable
Proportions," Journal 0/ Political Economy 88 (1980), 1009-25.
A. Mathios and R. Rogers, The Impact 0/ State Price and Entry Regulation on
Intrastate Long Distance Telephone Rates, Bureau of Economics Staff Report to
the Federal Trade Commission, November 1988.
A. Mathios and R. Rogers, "The Impact of Alternative Forms of State Regulation
of AT&T on Direct-Dial, Long-Distance Telephone Rates," RAND Journal 0/
Economics 20 (1989), 437-53.
H. McFarland, "Evaluating q as an 	Alternative to the Rate of Return 	in
Measuring Profitability," Review 0/ Economics and Statistics 70 (1988), 614-22.
B. Owen and P. Greenhalgh, "Competitive Considerations in Cable Television
Franchising," Contemporary Policy Issues 4 (1986), 69-79.
P. Pacey, "Cable Television in a Less Regulated Market," Journal o/Industrial
Economics 34 (1985), 81-91.
78

R. Posner, Antitrust Law (An Economic Perspective). 	Chicago: 	University of
Chicago Press, 1976.
R. Prager, "Franchise Bidding for 	Natural Monopoly: 	The Case of Cable
Television in Massachusetts," Journal of Regulatory Economics I (1989), 115-32.
H. Quirmbach, "The Path of Price Changes in Vertical Integration," Journal of
Political Economy 94 (1986), 1110-1119.
M. Salinger, "Vertical Mergers and Market Foreclosure," Quarterly Journal of
Economics 103 (1988a), 345-56.
M. Salinger, "A Test of Successive Monopoly and Foreclosure Effects: 	Vertical
Integration Between Cable Systems and Pay Services," draft, September 1988b.
D. Scheffman and P. Spiller, "Geographic Market Definition Under the U.S.
Department of Justice Merger Guidelines," Journal of Law and Economics 30
(1987), 123-48.
-R. Schmalensee, "A Note on the Theory of Vertical Integration," Journal of
Political Economy 81 (1973), 442-49.
C. Shapiro, "Premiums for High Quality Products as Returns to Reputations,"
Quarterly Journal of Economics 98 (1983), 659-79.
W. Sharkey, The Theory of Natural Monopoly. New York: Cambridge University
Press, 1982.
D. Sibley, "Asymmetric Information, Incentives, and Price-Cap Regulation,"
RAND Journal of Economics 20 (1989), 392·404.
A. 	Smiley, 	"Direct 	Competition 	Among 	Cable 	Television 	Systems,"- 	U.S.
Department of Justice Economic Analysis Group Working Paper No. EAG 86-
9, June 5, 1986.
J. Spengler, "Vertical Integration and Anti-trust Policy," Journal of Political
Economy 58 (1950), 347-52.
J. Tirole, The Theory of Industrial Organization. 	Cambridge: 	MIT Press, 1988.
U.S. GAO, Report to the Chairman. Subcommittee on Telecommunications and
Finance. Committee on Energy and Commerce. House of Representatives. National
Survey of Cable Television Rates and Prices, August 1989.
U.S. GAO, Resources, Community, and Economic Development Division, Follow-
up Survey of Cable Television Rates and Services, 1990.
C. von Weiszacker, "A Welfare Analysis of Barriers to Entry," Bell Journal of
Economics 11 (l980a), 399-420.
C. von Weiszacker, Barriers to Entry (A Theoretical Treatment). Berlin: Springer
Verlag, 1980b.
F. Warren-Boulton, "Vertical Control With Variable Proportions," JOllmal of
Political Economy 82 (1974), 783-802.
79

D. Waterman, "Vertical Integration in Information Industries," MRG Working
Paper #M8740, November 1987.
M 	Waterson, 	"Vertical 	Integration, 	Variable 	Proportions, 	and 	Oligopoly,"
Economic Journal 92 (1982), 129-44.
G. Webb, The Economics 01 Cable Television. 	Lexington, MA: 	Lexington, 1983.
G. 	Werden, 	"Market 	Delineation 	and 	the 	Justice 	Department's 	Merger
Guidelines," Duke Law Journal (1983), 514-79. 	.
F. Westfield, "Vertical Integration: Does Product Price Rise or Fall?" American
Economic Review 71 (1981), 334-46.
O. Williamson, "Franchise Bidding for Natural Monopoly: In General, and with
Respect to CATV," Bell Journal 01 Economics 7 (1976), 73-104.
M 	Zupan, "The Efficacy of Franchise Bidding Schemes in the Case of Cable
Television: 	Some Systematic Evidence," Journal 01 Law and 	Economics 32
(I 989a), 40 I-56.
M Zupan, "Non price Concessions and the Effect of Franchise Bidding Schemes
on Cable Company Costs," Applied Economics 21 (I989b), 305-23.
M. 	Zupan, 	"Cable 	Franchise 	Renewals: 	Do 	Incumbent 	Firms 	Behave
Opportunistically?" RAND Journal 01 Economics 20 (l989c), 473-82.
80