University of Southern California Dissertations and Theses

Essays on the economics of non-profit institutions

(USC Thesis Other)

Essays on the economics of non-profit institutions

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content ESSAYS ON THE ECONOMICS OF NON-PROFIT INSTITUTIONS

by

Rahul Nilakantan

A Dissertation Presented to the
FACULTY OF THE USC GRADUATE SCHOOL
UNIVERSITY OF SOUTHERN CALIFORNIA
In Partial Fulfillment of the
Requirements for the Degree
DOCTOR OF PHILOSOPHY
(ECONOMICS)

August 2010

Copyright 2010 Rahul Nilakantan
ii

Acknowledgements
I am grateful foremost to my mother, father, and brother, without whose
constant encouragement and support I would never have survived the PhD
process. Thank you for believing in me, even when I sometimes did not.
My mentors Prof. C. Selvaraj and Prof. Samar K. Datta inspired me to
pursue an academic career in Economics. They put me on the path to my PhD, and
by their personal example, motivated me even when the going was tough. Thank
you for showing me the way and for not allowing me to lose my way.
I am especially grateful to my advisors Prof. Simon Wilkie, Prof. Jeffrey B
Nugent, and Prof. Ricardo Alonso. Every PhD student should be as lucky as I am
to have such supportive advisors. Thank you for guiding me through my PhD and
for helping me to bring it to a successful conclusion.
I am grateful to Ganapathi, Lakshmi, Saraswathi, Shiva, and Venkatesha
for answering all my prayers.
I thank my friends and fellow travelers in the PhD program, who have
shared my joys and sorrows over the past five years, as well as cheerfully served
as a captive audience and sounding board for all my ideas, good or bad.
I thank the Economics Department staff Young Miller, Morgan Ponder,
Shannon Durbin, and Christopher Frias for their continuous support over the last
five years, especially during my job market search.
iii

Table of Contents
Acknowledgements ii
List of Tables v
List of Figures vii
Abbreviations viii
Abstract x
Chapter I: Introduction 1
The Television Industry 2
Agenda 2
Preview of Main Findings 11
Agricultural Cooperatives 26
Agenda 26
Preview of Main Findings 35
Chapter II: The FTA FTA Industry Structure 39
The Simultaneous Setting 40
The PE Case 48
The ME Case 51
Comparing the PE and ME Equilibria 54
The Sequential Setting 62
The PE Case 63
The ME Case 69
Comparing the PE and ME Equilibria 73
Summary of Results 77
Chapter III: The FTA PTV Industry Structure 85
The Simultaneous Setting 86
The PE Case 95
The ME Case 100
Comparing the PE and ME Equilibria 104
The Sequential Setting 114
The PE Case 116
The ME Case 123
iv

Comparing the PE and ME Equilibria 128
Summary of Results 131
Chapter IV: Quality Choice by Agricultural Cooperatives 143
Introduction 143
The Model and Results 154
Perfect Competition 158
Monopoly 159
Industry Structure and Equilibrium Choice of λ 162
Welfare Analysis 164
Conclusion 168
Chapter V: Conclusion 171
The Television Industry 171
Agricultural Cooperatives 182
References 187
Appendix 190
Proofs of Propositions in Chapter II 190
Proofs of Propositions in Chapter III 197
Proofs of Propositions in Chapter IV 205
v

List of Tables
Table 1.1 Market Shares in 2007 5
Table 2.1 The FFSIM PE Equilibrium Actions 49
Table 2.2 Payoffs under FFSIM PE 51
Table 2.3 The FFSIM ME Equilibrium Actions 53
Table 2.4 Broadcaster Profits under FFSIM ME 54
Table 2.5 CS and Welfare under FFSIM ME 54
Table 2.6 The FFSEQ PE Equilibrium Actions 65
Table 2.7 Broadcaster Profits under FFSEQ PE 66
Table 2.8 CS and Welfare under FFSEQ PE 67
Table 2.9 The FFSEQ ME Equilibrium Actions 71
Table 2.10 Broadcaster Profits under FFSEQ ME 72
Table 2.11 CS and Welfare under FFSEQ ME 72
Table 3.1 The FPSIM PE Equilibrium Actions 96
Table 3.2 Payoffs under FPSIM PE 99
Table 3.3 The FPSIM ME Equilibrium Actions 102
Table 3.4 Broadcaster Profits under FPSIM ME 103
Table 3.5 CS and Welfare under FPSIM ME 104
Table 3.6 The FPSEQ PE Equilibrium Actions 118
Table 3.7 Broadcaster Profits under FPSEQ PE 118
vi

Table 3.8 CS and Welfare under FPSEQ PE 118
Table 3.9 The FPSEQ ME Equilibrium Actions 126
Table 3.10 Broadcaster Profits under FPSEQ ME 126
Table 3.11 CS and Welfare under FPSEQ ME 127

vii

List of Figures
Figure 1.1 Typology of Models 11
Figure 2.1 The FFSIM ME Equilibrium 53
Figure 2.2 The FFSEQ PE Equilibrium 66
Figure 2.3 The FFSEQ ME Equilibrium 71
Figure 3.1 The FPSIM ME Equilibrium 103
Figure 3.2 The FPSEQ PE Equilibrium 119
Figure 3.3 The FPSEQ ME Equilibrium 127
Figure 4.1 Welfare Comparison 166

viii

Abbreviations
BBC British Broadcasting Corporation
EC European Commission
EU European Union
FFSEQ Free to Air Free to Air Sequential
FFSIM Free to Air Free to Air Simultaneous
FOC First Order Condition
FPSEQ Free to Air Pay TV Sequential
FPSIM Free to Air Pay TV Simultaneous
FTA Free to Air
HMR Hollander, Monier-Dilhan and Raynal (1999)
ICA International Cooperative Alliance
LHS Left Hand Side
MC Marette and Crespi (2003)
ME Mixed Economy
MENA Middle East and North Africa
PE Private Economy
PPSIM Pay TV Pay TV Simultaneous
PSB Public Service Broadcasting
PTV Pay TV
ix

RHS Right Hand Side
SOC Second Order Condition
SPNE Sub-game Perfect Nash Equilibrium
TV Television
UK United Kingdom

x

Abstract
This dissertation examines the economics of non-profit institutions in the
television industry and agriculture. We analyze two kinds of non-profit
institutions: (1) state owned welfare maximizing television broadcasters, and (2)
agricultural cooperatives. For state owned welfare maximizing television
broadcasters, we focus on competition with privately owned profit maximizing
television broadcasters over advertising levels, viewer subscription fees and
program content. For agricultural cooperatives, we focus on the quality provision
decision when the cooperative has local area monopoly status in a situation where
quality of individual units of output is unobservable and there are capacity
constraints on output grading infrastructure.
We model competition in the television industry under various settings of
industry structure (Free to Air and Pay TV), market structure (private economy
and mixed economy), and broadcaster entry timing (simultaneous and sequential).
We provide theoretical results identifying conditions under which the public
provision of television broadcasting services results in a welfare improvement,
identify the effects of such provision on competing privately owned broadcasters,
and thereby inform the debate in regulatory circles regarding the appropriate
disciplines on state broadcasters as well as the appropriate treatment of loss
making state broadcasters.
xi

Our main results are as follows. Broadcasters’ decisions on advertising
levels and viewer subscription fees are not always more optimal under mixed
economy. Social welfare is always higher under mixed economy except in
situations where the state broadcaster is at a corner solution for advertising level
i.e. 0% advertising level (e.g. BBC, France TV etc.), or 100% advertising level.
The optimal regulatory response is not to privatize the state broadcaster, but
simply to regulate advertising on the state broadcaster i.e. impose a lower (upper)
limit on advertising when there is 0% (100%) advertising level on the state
broadcaster. This is because judicious regulation of advertising levels can deliver a
welfare improvement greater than could be had by privatization.
Private broadcaster profits are mostly lower under mixed economy, thereby
lending credence to their complaints of suffering due to “anti-competitive” actions
of state broadcasters. However, no regulatory action is required unless the state
broadcaster is at a corner solution for advertising level, since lower private
broadcaster profits are consistent with higher welfare under mixed economy. The
state broadcaster usually makes lower profits than its private counterpart, and
sometimes even makes losses while fulfilling its public service broadcasting
mandate. Again, privatization is not the appropriate regulatory response to losses
of the state broadcaster. The appropriate response is either judicious regulation of
its advertising level, or the provision of budgetary support. As long as the
xii

deadweight losses of budgetary support to the state broadcaster do not exceed the
welfare gains from having an operational state broadcaster, a welfare argument
can be made for the continued existence and operation of loss making state
broadcasters.
With regard to agricultural cooperatives, it is the case that in India, most
agricultural cooperatives have local area monopoly status in their area of operation
(typically a number of villages). One purpose of according local monopoly status
to agricultural cooperatives was to rectify the market power imbalance caused by a
large number of small farmers facing a few large traders. In light of the findings of
the literature on the quantity and quality distorting effects of monopolies, this
dissertation investigates the welfare implications of permitting the unregulated
operation of monopoly cooperatives in agricultural settings where quality is
unobservable and grading capacity is insufficient to grade all available output.
We find that for a given capacity constraint in the grading technology, if the
cost of quality is high enough, or if the quality gap between low and high quality
output is low enough, the co-op member farmer may provide a greater amount of
high quality output at a lower price than the equivalent perfectly competitive
farmer. Similarly, for a given cost of quality and a given quality gap between low
and high quality output, if the grading capacity is low enough, the co-op member
farmer may provide a greater amount of high quality output at a lower price than
xiii

the equivalent perfectly competitive farmer. In both cases, welfare under the
monopoly co-op exceeds that under perfect competition.
The co-op member farmer’s choice of quality is less sensitive to the
reduction in grading capacity than that of the competitive farmer. This is because
the co-op member farmer, acting through the co-op, has two tools at his disposal to
handle the weakened incentive to provide quality – (1) he can reduce quality, and /
or (2) he can distort prices. As a result, his reliance on the first tool i.e. quality
reduction is less than that of the competitive farmer. Therefore, for a sufficiently
low grading capacity, the quality choice of the co-op member farmer will be
higher than that of the competitive farmer. Since farmers in both market structures
provide sub-optimally low levels of quality in the presence of grading capacity
constraints, welfare will be higher under the market structure where the extent of
quality under-provision is less i.e. monopoly. We therefore provide a welfare
argument for the continued existence and operation of monopoly cooperatives in
agricultural settings where quality is unobservable and there are sufficiently tight
constraints on grading capacity.

1

Chapter I
Introduction
This dissertation examines the economics of non-profit institutions in the
television industry and agriculture. We analyze two kinds of non-profit
institutions: (1) state owned television broadcasters, and (2) agricultural
cooperatives. For state owned television broadcasters, we focus on competition
with privately owned television broadcasters over advertising levels, viewer
subscription fees and program content. For agricultural cooperatives, we focus on
the quality provision decision when the cooperative has local area monopoly status
in a situation where quality of individual units of output is unobservable and there
are capacity constraints on output grading infrastructure.
We argue that state owned television broadcasters can be considered non-
profit institutions because they have objectives that may or may not include profit
maximization. We abstract from the case where state owned broadcasters serve
merely as propaganda vehicles to perpetuate corrupt regimes, and focus on a more
benign role for state broadcasters, i.e. to provide public service broadcasting
(PSB). We discuss the meaning of PSB objectives as understood across the world
in later paragraphs.
Cooperatives can also be considered to be non-profit institutions.
According to the International Cooperative Alliance (2009), a cooperative can be
2

defined as “…an autonomous association of persons united voluntarily to meet
their common economic, social, and cultural needs and aspirations through a
jointly-owned and democratically-controlled enterprise.” Like state owned
television broadcasters, cooperatives also have several objectives that may
encompass, but are not limited to profit maximization. The International
Cooperative Alliance (2009) observes the following about cooperative values.
Cooperatives follow a broader set of values than those associated purely
with making a profit. Because co-operatives are owned and democratically-
controlled by their members (individuals or groups and even capital
enterprises) the decisions taken by co-operatives balance the need for
profitability with the needs of their members and the wider interests of the
community.

We first elaborate on our agenda for the television industry, and then for
agricultural cooperatives.
The Television Industry
Agenda
Competition between television broadcasters is an example of platform
competition in a two sided market. Broadcasters act as platforms through which
viewers on one side of the market and advertisers on the other interact. Advertisers
produce goods, and inform customers about their products through advertising.
Viewers receive direct benefits from viewing the programs, and purchase goods
from advertisers after seeing their advertisements.
3

Recent papers
1
on program content and advertising intensity
2
in the
television broadcast industry, with the exception of Kind et. al. (2007) and Pan
(2009), have modeled platform competition in the context of a purely private
economy industry structure (PE), where each broadcaster’s sole objective is to
maximize profits. We argue that the television industry should be modeled as a
mixed economy industry structure (ME), where state owned broadcasters
3
with
possibly non-profit objectives compete with profit maximizing privately owned
broadcasters.
State control of television stations is pervasive - Djankov et. al. (2001) find
in a sample of 97 countries that the state controls 60% of all television stations. In
43 countries (mostly MENA countries), the state monopolizes television stations
with local news programs. However, control of television stations does not ensure
that viewers actually watch the television stations. Even so, state broadcasters
across the world command a significant market share in terms of television
viewers. For example, in the EU alone, state broadcasters in 2007 accounted for an
average of 29.23% of all viewers, with a maximum market share of 71.50% in

1
For example: Anderson and Coate (2005), Gabszewicz et. al. (2004), Nasser et. al. (2007), Peitz and
Valletti (2008), etc.

2
Advertising intensity is a measure of the amount of advertisements carried per broadcast.

3
State owned broadcasters may also be referred to as state broadcasters, public broadcasters, or public
service broadcasters.
4

Denmark
4
. Table 1.1 displays the names and market shares of state broadcasters in
several mature television markets.
State owned broadcasters may have several objectives that may or may not
include profit maximization. We abstract from the case where state owned
broadcasters serve merely as propaganda vehicles to perpetuate corrupt regimes,
and focus on a more benign role for state broadcasters, i.e. to provide PSB
services.
Heap (2005) states that PSB objectives are similar the world over, whether
fulfilled by state broadcasters acting as dedicated public service broadcasters, or
by private broadcasters fulfilling mandatory PSB requirements. He provides an
example of PSB objectives within the EU
5
, as laid down by the 1994 Prague
resolution and the 1996 EU Parliament resolution: (1) to aid informed citizenship,
(2) to support democratic values, (3) to offer a wide range of quality in all program
genres, and (4) to promote social cohesion and the vitality of national cultures.

4
Data on the market share of state broadcasters was obtained from the MAVISE database as on 07/04/09.

5
Within the UK, Block et. al. (2001) report that the Broadcast Research Unit of the British Film Institute
suggested in 1986 the following guidelines for public service broadcasting: (1) geographic universality, (2)
universality of appeal, (3) space for minorities, (4) national identity, (5) distance from vested interests, (6)
universality of payment, (7) promotion of competition in good programming, and (8) guidelines to liberate
program makers. See Block et. al. (2001) for a discussion of these objectives.

5

Table 1.1 Market Shares
6
in 2007

Although Norris et. al. (2003, pg.48) find cases where dedicated public
service broadcasters seem to be more interested in increasing audience shares than
in fulfilling their PSB mandates, most state broadcasters adhere to some or all of
their PSB objectives. The BBC is an extreme example of adherence to PSB
objectives to the exclusion of all other objectives. For example, the BBC’s “public

6
Source: Eurovision Audiovisual Observatory (2009).
6

purposes” are (1) sustaining citizenship through news, information and analysis of
current events and ideas; (2) promoting education and learning, both formal and
informal; (3) stimulating creativity and cultural excellence; (4) reflecting the UK’s
nations, regions and communities; and (5) bringing the world to the UK and UK to
the world. According to an internal review of its royal charter, the BBC exists to
deliver public value through these purposes, and for no other reason (British
Broadcasting Corporation, 2005).
The mixed economy literature in one sided market
7
settings has long
recognized the fact that when state owned enterprises with objectives that differ
from profit maximization compete with profit maximizing private enterprises, they
may have incentives to undertake “anti-competitive” actions that hurt their private
enterprise counterparts. Further, even if the objective of the state owned enterprise
is welfare maximization, its actions may reduce welfare in equilibrium
(Sappington and Sidak, 2003). It is important therefore to see whether such
concerns hold good in two sided market settings such as the television broadcast
industry.
An example of “anti-competitive” behavior of public service broadcasters
comes from private broadcaster complaints to the European Commission. A
complaint from Germany highlighted the failure of public service broadcasters to

7
The one sided market setting here refers to markets where buyers and sellers interact directly i.e. without
intermediaries who may charge for the privilege of bringing buyers and sellers together.
7

charge “market prices” for granting private broadcasters access to their
transmission facilities, thus creating an uneven playing field for private
broadcasters (European Commission, 2007).
State broadcasters have different sources of funding than private
broadcasters. State broadcasters may receive budgetary support from the
Government in addition to revenues from advertising, thus loosening the
constraints of the profit and loss calculus. This has led to complaints from
privately owned competitors of an uneven playing field, with state broadcasters
having a competitive advantage on account of being able to tap sources of funds
unavailable to their private competitors.
Complaints by private broadcasters to the European Commission against
state aid to public service broadcasters typically focus on the excessive levels of
such aid. Private broadcasters claim that these aid levels are more than sufficient
to provide the given level of public service broadcasting, and thus serve to
subsidize the commercial activities of public service broadcasters.
In Germany for example, a complaint highlighted the tendency of public
service broadcasters to outbid private broadcasters on lucrative assets like sports
telecast rights, which have no relation to public service broadcasting obligations
(European Commission, 2007). This has also been a source of concern to other
public service broadcasters in situations where there is more than one public
8

service broadcaster. For example, Channel 4, a commercially funded public
service broadcaster in the UK, has complained that the Government’s generous
support to BBC (another public service broadcaster) in the form of license fees
makes it difficult for Channel 4 to compete against the BBC! (Channel 4, 2008).
As we shall show later, it is not always the case that competing with a state
broadcaster proves detrimental to the private broadcaster. In certain cases, private
broadcasters would actually make greater profits when competing with a state
broadcaster than if they were competing with another private broadcaster!
Commercially funded public broadcasters that make losses face pressure to
improve their “bottom lines” or be privatized. For example, in the case of Channel
4 of the UK, anticipated future losses have led to suggestions of merger with a
private broadcaster (Shaw & Wake, 2009). Privatization pressures may also be
driven not just by profit and loss considerations, but also by a belief in the
efficiency of markets. We argue that in most cases, privatization of even loss
making state broadcasters would probably reduce social welfare rather than
increase it.
Since (1) state broadcasters’ objectives may vary from profit maximization,
and (2) state broadcasters command a significant market share in terms of viewers,
it becomes necessary to look at program content and advertising intensity in a ME
television broadcasting industry. This dissertation attempts such an analysis,
9

treating the state broadcaster as a social welfare maximizer. We treat the choice of
program content and advertising intensity as endogenous, and contrast the
equilibrium in a ME with that in a PE industry structure, paying particular
attention to the equilibrium and optimal choices of program content, advertising
intensity, viewer subscription fees, social welfare, and broadcaster profits. The
main questions this dissertation asks are:
1. Do broadcasters make more optimal ad space and viewer subscription
fee decisions under ME i.e. do prices do a better job of internalizing the
externalities inherent in the FTA FTA and FTA PTV market structures
when one broadcaster is state owned, and has welfare maximization as
its objective?
2. Is social welfare enhanced by the presence and operation of state
broadcasters?
3. Do private broadcasters make lower profits under ME due to the “anti-
competitive” actions of state broadcasters? If so, what regulatory action
is called for?
We answer these questions under various settings of industry structure
(FTA FTA and FTA PTV), market structure (PE and ME), and entry timing
(simultaneous and sequential). In Chapter II, we analyze the FTA FTA industry
structure. The FTA FTA industry structure is modeled as one where both
10

broadcasters use the FTA broadcast technology. The FTA broadcast technology is
one where broadcasters do not charge viewers for the privilege of watching the
telecast. We first analyze competition in a simultaneous setting (called FFSIM
hereafter) where both broadcasters enter the industry simultaneously, and then in a
sequential setting (called FFSEQ hereafter) where there is phased entry of
broadcasters into the industry. In Chapter III, we analyze the FTA PTV industry
structure. The FTA PTV industry structure is modeled as one where one
broadcaster uses the FTA broadcast technology and the other broadcaster uses the
PTV broadcast technology. The PTV broadcast technology is one where
broadcasters charge viewers for the privilege of watching the telecast. We first
analyze competition in a simultaneous setting (called FPSIM hereafter) where both
broadcasters enter the industry simultaneously, and then in a sequential setting
(called FPSEQ hereafter) where there is phased entry of broadcasters into the
industry. Figure 1.1 displays the typology of models considered in Chapters II and
III.
This dissertation therefore extends the literature on program content and
advertising intensity in the television broadcast industry by analyzing the mostly
ignored mixed economy setting, where a welfare maximizing state broadcaster
competes against a profit maximizing private broadcaster, under alternate industry
structures (FTA FTA and FTA PTV) and alternate entry regimes (simultaneous
11

and sequential). We provide theoretical results identifying conditions under which
the public provision of TV broadcasting services results in a welfare improvement,
identify the effects of such provision on competing privately owned broadcasters,
and thereby inform the debate in regulatory circles regarding the appropriate
disciplines on state broadcasters as well as the appropriate treatment of loss
making state broadcasters.
Figure 1.1: Typology of Models

Preview of Main Findings
Anderson and Coate (2005) recognized that an FTA television broadcast
has public good properties in that viewers derive direct benefits from watching the
broadcast, but cannot be excluded from watching the broadcast. Further,
advertisers derive indirect benefit from broadcasts by contacting viewers through
advertisements, but impose negative externalities on viewers on account of
PE
ME
FTA
FTA
FTA
PTV
FTA
FTA
FTA
PTV
PE
ME
Ch II
FFSIM
Ch III
FPSIM
Ch II
FFSIM
Ch III
FPSIM
Ch II
FFSEQ
Ch II
FFSEQ
Ch III
FPSEQ
Ch III
FPSEQ
Simultaneous Sequential
12

subjecting them to advertising. As a result, there are two sources of market
imperfections under FTA broadcasting.
Not surprisingly, Anderson and Coate (2005) found that the decisions of
private broadcasters under FTA technology regime in PE with respect to ad
intensity are sub-optimal, since they have no incentive to equate the marginal
social cost of advertising (the disutility or nuisance to viewers from ads) with the
marginal social benefits of advertising (gains from trade between viewers and
advertisers
8
). The tradeoff relevant to the profit maximizing private broadcaster is
the revenue gains from carrying more ads versus the loss of audience market share
due to carrying more ads. They find that when the viewer nuisance per ad is low
(high), there is social under (over) provision of advertisements in a PE.
In all the models under consideration in this dissertation, we find sub-
optimal charging of viewer subscription fee and sub-optimal provision of ad space
under both PE and ME. The general pattern of sub-optimality observed is over-
charging (over-provision) of viewer subscription fee (ad space) when the marginal
social benefit of advertising is less than the marginal social cost of advertising, and
vice versa. Just as in Anderson and Coate (2005), the reason for this sub-
optimality is that the tradeoff that private broadcasters face is different from the
tradeoff confronting the social planner.

8
Advertisers in their model produce goods that are bought by viewers of advertisements.
13

The first question this dissertation asks is whether prices do a better job of
internalizing the twin market imperfections of the FTA broadcast technology when
one broadcaster is state owned, and has welfare maximization as its objective. We
find that the answer is “yes, but not always”. Under FFSIM, it is always true that
both broadcasters make more optimal ad space choices under ME, while it is true
only under certain conditions on the marginal social benefits and costs of
advertising under FFSEQ. Under FPSIM, both broadcasters make more optimal ad
space and viewer subscription fee choices under ME only under certain conditions
on the marginal social benefits and costs of advertising, while it is always the case
that broadcasters’ decisions are more optimal under ME for FPSEQ.
The two main components of social welfare are welfare from advertising
i.e. the gains from trade between advertisers and viewers, and the welfare from
program content i.e. viewer transport costs from watching non-preferred programs
in a Hotelling framework. As a welfare maximizer, the state broadcaster pays
attention to both the welfare from advertising and the welfare from program
content when making its ad space decisions.
If there is social under / over provision of advertising, the state broadcaster
can improve welfare from advertising by increasing / decreasing its own ad space,
thus mitigating the extent of under / over provision of advertising. Since there is
strategic complementarity in the choice of ad space and viewer subscription fee,
14

this usually results in more optimal ad space / viewer subscription fee decisions by
the private broadcaster under ME. However, increased welfare from advertising
may come at the cost of decreased welfare from program content.
In our framework, welfare from program content is maximized when there
is content differentiation and both broadcasters share the audience equally, since
this configuration minimizes viewer transport costs from watching non-preferred
programs. More optimal (in the sense of welfare from advertising) ad space /
viewer subscription fee decisions by broadcasters under ME may result in lower
welfare from program content on account of skewing the audience share even
further away from equality under ME. In order to increase total welfare, it may
therefore be necessary to trade welfare from advertising against welfare from
program content, resulting in possibly less optimal (in the sense of welfare from
advertising) ad space / viewer subscription decisions under ME in spite of having
greater total welfare under ME.
Further, it is not always the case that more optimal ad space decisions by
the state broadcaster will result in more optimal ad space / viewer subscription fee
decisions by the private broadcaster. In the sequential setting, we find that the
private broadcaster has a second mover advantage due to strategic
complementarity in the choice of ad space / viewer subscription fee. Thus, the well
known result of Gal-Or (1985) regarding second mover advantage for identical
15

firms competing in prices, even if they produce differentiated products, extends to
this particular two sided markets setting. Here, advertising space / viewer
subscription fee serves as the indirect / direct price viewers pay in order to watch
the program. This strategic complementarity weakens the ability of the state
broadcaster to induce more optimal (in the sense of welfare from advertising) ad
space / viewer subscription fee decisions from the private broadcaster.
These two forces i.e. (1) the potential tradeoff between welfare from
program content and welfare from advertising, and (2) the second mover
advantage of the private broadcaster, sometimes result in more optimal (from the
point of view of welfare from advertising) ad space / viewer subscription fee
decisions under ME, but sometimes also result in less optimal (from the point of
view of welfare from advertising) decisions under ME.
In the spirit of the present dissertation, Pan (2009) also compares
competition between television broadcaster in the ME and PE settings, but in a
situation where the state owned broadcaster maximizes viewer surplus with a
(rather restrictive) binding break even constraint, and the broadcasting technology
is PTV. The equilibrium in that model involves under provision of advertisements
by both broadcasters under ME and PE. This is because neither broadcaster cares
about the welfare of advertisers, nor can they capture all the rent from advertisers
even if they wanted to. In contrast to the models in this dissertation, both
16

advertisers in Pan (2009) choose identical levels of ad space under ME and PE.
Given that the broadcasting technology is PTV, the equilibrium involves both
broadcasters maximizing joint surplus of broadcaster and viewer, and then using
subscription fees to allocate the surplus between broadcaster and viewer. As a
result, both broadcasters set the same ad space, but different subscription fees
under ME and PE.
Kind et. al. (2007) consider competition between television broadcasters
under ME and PE in an oligopoly setting. This is in contrast to the recent literature
(including the present dissertation), which has analyzed competition between
television broadcasters in a duopoly setting. The broadcast technology they
consider is FTA, and the state broadcaster is a welfare maximizer. Under PE, they
find that ad space is over provided when programs are poor substitutes, but under
provided if programs are good substitutes.
The intuition behind this result is that broadcasters exploit market power
when programs are poor substitutes by carrying more ad space even though this
reduces viewer surplus. When programs are poor substitutes however, the state
broadcaster reduces its ad space, thus mitigating the problem of over provision of
ad space under ME. Unfortunately, Kind et. al. (2007) do not analyze the welfare
and profit implications of the mitigation of the over provision of advertising under
ME.
17

With regard to program content, Gal-Or and Dukes (2003) consider the
case of a differentiated products market with Nash bargaining on advertising
prices between stations and broadcasters. The broadcasting technology is FTA.
They find that broadcasters have incentive to minimize content differentiation,
which in turn leads producers to advertise less. With less product information
available to consumers, producers gain higher margins on product sales, which
permits broadcasters to command higher prices for advertisements.
Gabszewicz et. al. (2004) consider the case of a homogeneous products
market with advertisers acting as price takers on the advertising market. The
broadcasting technology is FTA. They find that when viewer aversion to
advertisements is low, broadcasters tend to differentiate content, and vice versa.
Differentiating content allows broadcasters to have niche markets, carry more
advertising and hence make greater revenues. However, the ability to carry more
advertising is inversely proportional to the strength of viewer aversion to
advertisements. When aversion is high enough, broadcasters prefer content
duplication to retain market share since they can carry only limited amounts of
advertising.
Peitz and Valletti (2008) consider the case of a quality differentiated
products market with advertisers acting as price takers on the advertising market.
They consider both FTA and PTV broadcasting technologies. Under the FTA
18

technology regime, they find that the greater the viewer aversion to
advertisements, the more broadcasters tend to differentiate program content. This
is because content duplication would result in ruinous Bertrand competition in the
sense of lowering the number of advertisements (and hence revenues) in an
attempt to boost market share. Content differentiation however makes viewers less
sensitive to differences in advertising levels across broadcasters, permitting
broadcasters to carry more ads and thus make greater revenues.
Pan (2009) finds maximal content differentiation under PE. Under ME, Pan
finds intermediate content differentiation, with the state broadcaster close to the
mid point of the Hotelling line, and the private broadcaster close to one end. The
state broadcaster tends to move closer to the private broadcaster to “squeeze” its
profit and therefore increase viewer surplus, while the private broadcaster tends to
retreat in order to soften the competition.
In the models under consideration in this dissertation, we permit only two
content choices: maximal content differentiation, and perfect content duplication.
By an analogous argument to that in Peitz and Valletti (2008), perfect content
duplication would lead to a “race to the bottom” in terms of viewer subscription
fees and advertising levels, and thus zero viewer subscription and advertising
revenue under all industry structures (FTA FTA and FTA PTV), all entry
situations (simultaneous and sequential), and all market structures (PE and ME).
19

Further, it would lead to a reduction in social welfare on account of lost
opportunities for gains from trade between advertisers and viewers, as well as
increased viewer transport costs. Therefore, both broadcasters in all situations
prefer content differentiation to content duplication in equilibrium.
The second question this dissertation asks is if social welfare is enhanced
by the operation of state broadcasters. We find that the answer is “yes, but not
always”. In FFSIM, we find that welfare under ME is always higher than total
welfare under PE. For FFSEQ, we find that welfare is higher under ME except in a
small parameter range characterized by high marginal social benefit of advertising
and low marginal social cost of advertising. For FPSIM, we find that in equilibria
where the choice of ad space is the same or more optimal under ME i.e. when the
marginal social cost of advertising is high, we have welfare improvements under
ME, and vice versa. For FPSEQ, we find that welfare is the same or higher under
ME whenever the marginal social benefit of advertising is high enough relative to
the marginal social cost, and vice versa.
In sequential settings, reduced welfare under certain conditions on the
marginal social benefits and marginal social costs of advertising under ME can be
explained in terms of the second mover advantage of the private broadcaster
weakening the ability of the state broadcaster to call forth more optimal (in terms
of welfare from advertising) ad space / viewer subscription fee decisions from
20

both broadcasters, and thus resolving the tradeoff between welfare from
advertising and welfare from program content in the direction of lower overall
welfare. In simultaneous settings, lower welfare under ME is observed only in the
FTA PTV industry structure, where one of the multiple equilibria results in less
optimal ad space / viewer subscription fee choices by both broadcasters under
certain conditions on marginal social benefits and costs of advertising.
Despite the fact that social welfare is not always higher under ME, the
situations where social welfare is lower under ME are easily identified, since these
are the situation where there is a corner solution for the choice of ad space by the
state broadcaster i.e. 100% ad space or 0% ad space
9
. The appropriate response is
then the regulation of ad space on the state broadcaster, not the privatization of the
state broadcaster. This is because privatization of the state broadcaster may give a
welfare improvement that is smaller than could be realized by regulation of ad
space on the state broadcaster.
Recall that there were two market imperfections in the FTA FTA industry
structure i.e. the public good nature of the FTA broadcast and the fact that
advertisements impose a negative externality on viewers. The extent of these
imperfections is partially mitigated under the FTA PTV industry structure, since
viewers of channel using the PTV technology can be made to pay for the privilege

9
There are state broadcasters with 0% ad space e.g. BBC, but no state broadcasters with 100% ad space.

21

of watching the telecast, and do not have to view any advertisements
10
. When
comparing the FTA FTA and the FTA PTV equilibria for PE in the simultaneous
setting, we find that even though both market imperfections are alleviated to a
degree under FTA PTV, social welfare is lower under FTA PTV than under FTA
FTA when the marginal social benefit of advertising exceeds the marginal social
cost, and vice versa. This is an interesting application of the general theory of
second best. As Lipsey and Lancaster (1957, pg.12) observe, “… it is not true that
a situation in which more, but not all, of the optimum conditions are fulfilled (e.g.
FTA PTV) is necessarily, or is even likely to be, superior to a situation in which
fewer are fulfilled (e.g. FTA FTA).”
Welfare is lower under FTA PTV when the marginal social benefit of
advertising exceeds the marginal social cost because this is the situation where
optimality demands that every producer be allowed to advertise – since
programming is advertisement free under PTV broadcast technology and the
channel using the FTA broadcast technology carries the same amount of
advertising under both FTA PTV and FTA FTA, it is no wonder that welfare is
lower under FTA PTV. This is a perverse case where partially alleviating both
market imperfections through the use of PTV broadcast technology actually
lowered welfare!

10
There are however PTV broadcasters that do carry advertisements. We do not consider such broadcasters
in this dissertation.
22

Similarly, welfare is higher under FTA PTV when the marginal social
benefit of advertising is lower than the marginal social cost because this is the
situation where optimality demands that no producer be allowed to advertise –
since programming is advertisement free under PTV broadcast technology and the
channel using the FTA broadcast technology carries the same amount of
advertising under both FTA PTV and FTA FTA, it is no wonder that welfare is
higher under FTA PTV.
In a PPSIM setting, Pan (2009) finds a welfare improvement under ME.
However, unlike in the present model, the source of the welfare improvement is
not from more optimal ad space choices, since Pan finds that both broadcasters
have identical ad space choices under ME and PE. The source of welfare gains is
from more optimal program content decisions under ME, since broadcasters no
longer maximize content differentiation like they do under PE.
The third question this dissertation asks is whether the private broadcaster
makes lower profits under ME. This question is motivated by complaints of
private broadcasters against “anti-competitive” practices of state broadcasters. We
find that the answer is “yes, but not always”. For FFSIM, we find that the private
broadcaster actually makes greater profits under ME when the marginal social
benefit of advertising exceeds the marginal social cost! For FFSEQ, the private
broadcaster always makes lower profits under ME when the marginal social
23

benefit of advertising is less than the marginal social cost, but may make lower or
higher profits when the converse is true. For FPSIM and FPSEQ, the private
broadcaster always makes lower profits under ME.
Therefore, depending on the marginal social benefits and costs of
advertising, the state broadcaster takes actions that are either to the benefit of, or to
the detriment of the private broadcaster (where benefit means profit and detriment
means loss). These results are the two sided market analog of the result in
Sappington and Sidak (2003, pg.184), who find that a “reduced focus on profit can
provide state enterprises with stronger incentives to pursue activities that
disadvantage competitors.”
In the two sided market setting, the state broadcaster tailors the
aggressiveness with which it competes against the private broadcaster depending
on the relationship between the marginal social benefit and the marginal social
cost of advertising. More aggressive competition means more aggressive ad space
reductions by the state broadcaster when the marginal social benefit of advertising
is less than the marginal social cost, while less aggressive competition means
carrying a greater ad space on the state broadcaster when the marginal social
benefit of advertising exceeds the marginal social cost. Subject to net effect of the
two forces i.e. (1) the potential tradeoff between welfare from program content
and welfare from advertising under both simultaneous and sequential settings, and
24

(2) the second mover advantage of the private broadcaster under sequential
settings, the private broadcaster may end up with greater ad space and / or greater
audience share under ME, in which case it could make greater profits under ME,
or vice versa.
Given that the complaints from private broadcasters about making lower
profits because of “anti-competitive” practices of state broadcasters have some
merit in certain situations, the question is whether any regulatory action is called
for to discipline state broadcasters. Under the FTA FTA industry structure, every
situation where private broadcasters make lower profits under ME than under PE
is consistent with a welfare improvement under ME. Therefore, no regulatory
action is called for.
Under FPSIM, lower profits for the private broadcaster is accompanied by
lower social welfare only when the state broadcaster carries 100% ad space i.e.
telecasts only advertisements and no programming! Since this situation rarely
arises in practice, and even if it does arise is easily detectable by the simple
expedient of watching the telecast for any evidence of non-advertising
programming, the solution would be simply to regulate advertising on the state
broadcaster rather than its privatization. Such regulation could potentially increase
social welfare to a level greater than possible under PE, but would not result in a
level of private broadcaster profit greater than possible under PE.
25

Under FPSEQ, when the marginal social benefit of advertising is low
enough compared to the marginal social cost, not only does the private broadcaster
make lower profit under ME, but social welfare is also lower under ME. However,
this situation occurs only when the state broadcaster carries 0% ad space i.e.
telecasts only programming and no advertisements! Since this situation is easily
detectable by the simple expedient of watching the telecast for any evidence of
advertising, the solution would be simply to regulate advertising on the state
broadcaster instead of its privatization. Such regulation could potentially increase
social welfare to a level greater than possible under PE, but would not result in a
level of private broadcaster profit greater than possible under PE.
In a PPSIM setting, Pan (2009) finds that the private broadcaster always
makes lower profits under ME than under PE. Since the state broadcaster cares
only about viewer surplus (subject to a binding break even constraint), it is a much
more aggressive competitor than another private broadcaster, thus lowering the
profits of the private broadcaster under ME. Given the binding break even
constraint, the state broadcaster is unable to tailor the aggressiveness of its
competition depending on the relationship between the profitability and the
nuisance value of advertising.
In the models considered in the present dissertation, depending on the
relationship between the marginal social benefit and the marginal social cost of
26

advertising, the state broadcaster may incur losses while attempting to carry out its
PSB mandate, and usually makes lower profit that it would have if it was a
privately owned profit maximizing entity. In such situations, if the deadweight
loss incurred in raising public funds to provide budgetary support to the state
broadcaster is low enough, a welfare argument can be made for the continued
existence and operation of state broadcasters, even if they are loss making.
Although privatization of the state broadcaster may increase its profits, it may well
lead to a reduction in social welfare, or deliver a welfare increase that is smaller
than could be achieved by regulation of ad space on the state broadcaster. We now
turn our attention to the analysis of agricultural cooperatives.
Agricultural Cooperatives
Agenda
In India, agricultural cooperatives may have monopoly status in their area
of operation (typically a number of villages). The Registrar of Cooperatives may
forbid the establishment of another cooperative in the same area if it is
“detrimental to the cooperative movement” (for example by interfering with the
operation of existing cooperatives). In practice, the exercise of this power of the
Registrar of Cooperatives has resulted in the overwhelming majority of
agricultural co-ops in India having local area monopoly status.
27

One purpose of according local monopoly status to agricultural
cooperatives was to rectify the market power imbalance caused by a large number
of small farmers facing a few large traders. While monopolies in general may
lower social welfare by restricting output, this may not true of cooperative
monopolies on account of legal requirements for co-ops to practice open
membership and / or to process or market all produce supplied to it by members.
Since this would remove one avenue for members to exploit the market power of a
monopoly, the existence of cooperatives would have to be justified by other
considerations such as scale economies, access to government subsidies, or the
ability to exploit market power while making decisions on product quality, product
range, or product grading / certification (if the output quality is unobservable).
Given the numerous other avenues for monopoly co-ops to exploit their
market power, simply removing the ability of a monopoly co-op to practice output
restriction would be insufficient to preserve efficiency. In light of the findings of
the literature on the quality distorting effects of monopolies both when quality is
observable and when quality is unobservable, it is necessary to investigate the
welfare implications of permitting the unregulated operation of monopoly
cooperatives in quality differentiated industries. Chapter IV of this dissertation
focuses on the choices of monopoly co-ops regarding the provision of product
28

quality in a setting where quality is unobservable and grading capacity is
insufficient to grade all available output.
Regarding the issue of market structure and quality choice when quality is
observable, the quality distortions under monopoly have been well known since
Spence (1975) for the case where the monopolist chooses a single level of quality,
and Mussa & Rosen (1978) for the case where the monopolist chooses a range of
qualities. When the quality and quantity setting monopolist chooses a range of
qualities, it is known that quality is under-supplied when there is full market
coverage, but is optimally supplied when there is partial market coverage – in the
latter case, the monopolist under-supplies quantity (Berges-Sennou, Bontems and
Requillart, 2003).
With regard to market structure and quality choice when the quality of
individual units of output is unobservable prior to purchase, Hollander, Monier-
Dilhan and Raynal
11
(1999) consider an oligopoly situation where producers
choose both the average quality of output (which is observed by the consumer),
and whether or not to grade their output, i.e. reveal the quality of individual units
of output. The cost of grading is a fixed cost, and there is sufficient grading
capacity to grade the entire output. The amount of output produced by each firm is
given exogenously – however, the firm can choose the proportion of output that is

11
Referred to as HMR hereafter.
29

of high quality (and hence can choose the average quality of output). Thus, HMR
impose a technological restriction on the firm's quantity quality tradeoff. This
restriction captures the fact that it is not possible to have perfect control over
quality and quantity in an agricultural setting.
Surprisingly, they find that an increase in the cost of grading has an
ambiguous effect on the average quality of output that undergoes grading. If the
cost of grading increases, the amount of ungraded output increases, while the
amount of graded output decreases. One would expect that only producers of
output that is on average of higher quality would have the incentive to go for
costly grading, since they could get higher prices for their high quality output from
quality conscious consumers. So, the average quality of output sent for grading
should increase.
However, since the total quantity of output is given exogenously, market
clearing requires some consumers to switch away from graded output to ungraded
output. If the number of switchers from high quality output to ungraded output
exceeds the number of switchers from low quality to ungraded output, then the
average quality of output sent for grading will fall, since the most lucrative market
for graded output (consumers who buy high quality output) has shrunk. So, an
increase in grading cost can actually lower the average quality of output sent for
grading. Further, they find that the average industry quality increases with the cost
30

of grading if the increased cost of grading reduces the quality of output that
undergoes grading.
Marette and Crespi
12
(2003) consider a model where there are two types of
producers, one producing high quality output and the other producing low quality
output, and demonstrate welfare improvements under stable producer cartels. In
their model, quality is unobservable to the consumer but known to the producer.
The producer can signal his quality with perfect accuracy by engaging in costly
third party certification (the analog of the grading decision in HMR). Producers
may collude in order to share certification costs and also reduce competition by
lowering output. There is sufficient certification capacity to certify the entire
industry output if so desired. Under certain conditions, MC find that a stable cartel
that provides information about product quality through certification may improve
overall social welfare even though competition is reduced.
The present dissertation focuses on market structure and quality choice by
contrasting the polar cases of perfect competition and monopoly co-ops. Producers
are homogeneous, and output can be of two qualities – low and high. We treat
quantity of output as exogenously given, say as a result of decisions on the
quantum of land set aside for a particular crop that were made in the past. We
allow the producer to choose the distribution of the fixed amount of output across

12
Referred to as MC hereafter.
31

low and high quality. This is equivalent to choosing the average quality of output,
and can be done for example by varying cultivation practices post the land
allocation decision. Thus, we impose a technological restriction on the farmer's
tradeoff between quality and quantity that is in the spirit of HMR, with the
difference that we consider the homogeneous producer case.
The more usual treatment of quality, such as that in MC, does not impose
any technological restriction on the tradeoff between quantity and quality. The
HMR treatment is more representative of an agricultural context where it is
difficult to perfectly control the quantity and quality of individual units of output,
while the MC treatment is more representative of a manufacturing context, where
it is possible to achieve perfect control the quantity quality of individual units of
output.
Contrary to HMR and MC however, we consider the situation where the
there are exogenous capacity constraints in grading. This could arise in situations
where grading infrastructure is either too expensive for widespread adoption, or is
newly introduced and therefore not in widespread use. Capacity constraints in
grading are commonplace across most agricultural markets in most developing
countries. In the presence of capacity constraints, as long as grading is profitable
for farmers, all output up to the grading capacity constraint is graded, with the rest
of the output sold as ungraded. Therefore, information is revealed about the
32

quality of individual units of output only for a fraction of total output, with no
information on quality being revealed for the rest of the output.
The fraction of output about which no information on quality is revealed is
treated as being of known average quality. Although average quality is a choice
variable of the producer, the consumer observes this choice prior to making his
consumption decision. This would be the case if the cultivation practices (but not
output) of the farmer could be costlessly and accurately certified by some agency,
or if the consumer could infer information about average quality from the quantity
of output in each quality grade, or from the prices of output in each quality grade.
A real world example of capacity constraints in grading arises in the quality
determination process in the Indian milk industry, which originally consisted of
use of eyeball and tongue, but graduated to the use of a lactometer
13
. The
following anecdote from the milk sector illustrates the linkage between market
structure, grading capacity and quality. Prior to the advent of the AMUL
14

movement in India, milk was sold by private milkmen, who were notorious for the
low quality of their milk. As the joke went, in an effort to adulterate the milk,
milkmen did not add water to milk, but rather added milk to water! This state of
affairs existed because there was no reliable method for the quality of individual

13
A device that measures the specific gravity of milk, and hence the quantum of adulteration with water.

14
The AMUL movement was begun by Dr. Varghese Kurien in Anand, Gujarat, and its principles were
spread across the entire country as part of Operation Flood.
33

units of milk to be made known prior to purchase, and hence no incentive for
either milk farmers or milkmen to provide high quality milk. Consumers got low
quality milk and milk farmers got low prices for their milk from private milkmen.
With the advent of the AMUL movement, payment to milk farmers began
to be made based on the specific gravity of milk, as measured by the lactometer.
With the increase in grading capacity that the slow diffusion of the lactometer
represented, milk farmers finally had the incentive to supply higher quality milk to
the dairy and hence to the consumer. The use of lactometers to adjudge the quality
of milk became an industry standard due to the pioneering work of the AMUL
movement leaders.
Prior to 1992, dairies (such as those of AMUL) had local area monopoly
status in their area of operation. During the registration process, dairies had to
declare an area known as a “milk shed”, and were given sole rights to procure milk
from their milk sheds. However, the government removed restrictions on new milk
processing capacity in 1992, and allowed competition between dairies even in the
same milk shed. Thus, at one stroke, the nature of the dairy industry changed from
being a monopoly, to being a competitive industry.
Once the capacity constraint in grading had been alleviated to the point
where the quality of each individual unit of milk could be determined with near
certainty through the use of the lactometer, there could be no welfare grounds for
34

the continued existence of monopoly dairies
15
. Although at the time the
government did not adduce welfare reasons for stripping dairies of local monopoly
status, its action could very well have led to welfare gains. What is interesting to
investigate is if the existence of monopoly co-op dairies during the period where
milk grading infrastructure was constrained because the lack of widespread use of
the lactometer was detrimental to welfare, or if as in the case of MC's welfare
enhancing cartels, there was some way that these monopoly dairies could have
actually increased welfare. This is the starting point of analysis for the present
dissertation.
Like all other agricultural cooperatives, the cooperative dairies which
dominated the Indian dairy scene prior to deregulation were subject to legal
restrictions on their ability to control the quantum of output of their members. The
first cooperative principle of the International Cooperative Alliance is “voluntary
and open membership”. This principle states that “Co-operatives are voluntary
organizations, open to all persons able to use their services and willing to accept
the responsibilities of membership, without gender, social, racial, political or
religious discrimination” (International Cooperative Alliance, 2009). This
principle was adopted in Indian Cooperative Law, through legal provisions to the
effect that any primary producer in the geographical jurisdiction of the co-op

15
As would be argued by the usual finding in the literature on market structure and quality choice, that
monopolies distort quality when quality is observable
35

would be eligible to become a member of the co-op. This removed much of the
freedom of the cooperative to indulge in quantity distortions by restricting
participation in the co-op. As a result, the co-op had to take the quantity of output
supplied by members as given. This exogeneity of output quantity is in the spirit
of HMR. MC however allow the producer cartel to use two routes to distort
quantity: (1) exclude producers from membership of the cartel, and (2) impose
quantity limits on producers who join the cartel.
In our setting, the co-op can influence the quality choice of their members
by instituting an appropriate internal quality payment schedule. By varying the
slope of the price quality schedule, the co-op can induce a given level of quality
from members. We do not however model the internal price quality schedule of
the co-op, but assume instead that the co-op determines the profit maximizing
quality level, and induces that from members through the use of an appropriate
price quality schedule.
Preview of Main Findings
We find that for given capacity constraint in the grading technology, if the
cost of quality is high enough, or if the quality gap between low and high quality
output is low enough, the co-op member farmer may provide a greater amount of
high quality output at a lower price than the equivalent perfectly competitive
farmer. Similarly, for a given cost of quality and a given quality gap between low
36

and high quality output, if the grading capacity is low enough, the co-op member
farmer may provide a greater amount of high quality output at a lower price than
the equivalent perfectly competitive farmer. In both cases, welfare under the
monopoly co-op exceeds that under perfect competition.
This finding is in the spirit of MC, who find welfare gains from stable
producer cartels under certain conditions. MC’s results require that there be no
binding constraint on grading capacity which would prevent producers from
realizing their grading intentions. In contrast, we impose binding exogenous
constraints on grading capacity, and find welfare gains under monopoly only if the
grading capacity is low enough!
The intuition behind our result is as follows. When there is no grading
capacity constraint i.e. quality of each unit of output can be determined with
certainty, the monopoly market structure provides sub-optimally low levels of
average quality, while the competitive market structure provides optimal levels of
average quality. As the grading capacity reduces, farmers in both market structures
lose their incentive to make costly effort to produce high quality output, and hence
reduce the average quality of their output.
However, the quality choice of the co-op member farmer is less sensitive to
the reduction in grading capacity than that of the competitive farmer. This is
because the co-op member farmer, acting through the co-op, has two tools at his
37

disposal to handle the weakening incentives to provide quality – (1) he can reduce
average quality, and / or (2) he can distort prices. As a result, his reliance on the
first tool i.e. average quality reduction is less than that of the competitive farmer.
Therefore, for a sufficiently low grading capacity, the quality choice of the co-op
member farmer will be higher than that of the competitive farmer.
Recall that when there is a grading capacity constraint, farmers in both
market structures provide sub-optimally low levels of average quality. Since
average quality is the only choice variable, welfare will be higher under the market
structure where the extent of quality under-provision is less. Since the extent of
quality under-provision is lower under monopoly when grading capacity is low
enough, welfare would be higher under monopoly than under perfect competition!
This provides a nice welfare justification for the government to protect the
monopoly status of dairies when grading capacity is low i.e. no widespread use of
lactometers, but strip them of their monopoly status after the use of the lactometer
became the industry standard!
In summary, Chapter IV of the dissertation contributes to the literature on
market structure and quality choice by considering the welfare implications of the
equilibrium choice of quality by a monopoly cooperative that produces a quality
differentiated good of unobservable quality in the presence of grading capacity
38

constraints. Chapter V of the dissertation concludes the discussion of the
economics of non-profit institutions in the television industry and agriculture.
39

Chapter II
The FTA FTA Industry Structure
In this chapter, we consider program content and advertising levels in a
mixed economy television broadcast industry under the FTA FTA industry
structure. The FTA FTA industry structure is modeled as one where both
broadcasters use the FTA broadcast technology. With regard to competition
between advertisers, we consider a homogenous product market structure. We
treat the choice of program content and advertising intensity as endogenous, and
contrast the equilibrium in a mixed economy with that in a purely private economy
industry structure.
We analyze competition between broadcasters under two settings: (1) a
simultaneous setting, and (2) a sequential setting. The simultaneous setting
involves both broadcasters entering the industry at the same time, while the
sequential setting involves one broadcaster entering the industry before the other.
In the sequential setting under ME, we focus attention on the case where the
broadcaster who enters the industry first is the state broadcaster. This is because
the observed pattern of entry in most world television markets is that the earlier
entrant is usually a state broadcaster, while the latter entrant is usually a private
broadcaster. The chapter is organized as follows. The next section analyzes the
40

simultaneous setting. The subsequent section analyzes the sequential setting, and
the final section summarizes and concludes the discussion.
The Simultaneous Setting
Consider first the simultaneous setting where both broadcasters enter the
industry simultaneously. We adapt the models of Gabszewicz et. al. (2004) and
Nasser et. al. (2007), where the three participants in the industry i.e. viewers,
advertisers, and broadcasters, together play the following multi-stage game. In
stage 1, both broadcasters simultaneously choose program type. In stage 2, both
broadcasters simultaneously choose the level of advertising. In stage 3, advertisers
choose which broadcasters to advertise with, and consumers make their program
viewing choices and good purchase decisions.
Consider first the decision problem of viewers. There is a mass N of
viewers. Viewers are distributed uniformly along the Hotelling interval, with taste
for programs ( ) 0,1 U β ∼ . There are two program types e.g. news and music,
indexed by 1 and 2. Producing program 1 is equivalent to the broadcaster choosing
a location of 0 on the Hotelling interval. Producing program 2 is equivalent to the
broadcaster choosing a location of 1 on the Hotelling interval.
Let the broadcasters be called PUBTV and PRITV. Program choice of
PRITV is denoted by distance
1
d from the end of the Hotelling interval indexed by
0, and program choice of PUBTV by distance of
2
d from the end of the Hotelling
41

interval indexed by 1. Thus, if both PRITV and PUBTV produced program 1, then
( ) ( )
1 2
, 0,1 d d = . Similarly, if both PRITV and PUBTV produced program 2, then
( ) ( )
1 2
, 1,0 d d = . These are the cases where program content is duplicated by
broadcasters. If PRITV produced program 1 and PUBTV produced program 2, or
vice versa, then ( ) ( )
1 2
, 0,0 d d = or ( ) ( )
1 2
, 1,1 d d = respectively. These are the cases
where program content is differentiated by broadcasters.
With quadratic transportation costs, a viewer located at position β on the
Hotelling interval is at a distance β from program 1, and hence incurs a disutility
2
β if she watches program 1. Similarly, she is at a distance of ( ) 1 β − from
program 2, and hence incurs a disutility of ( )
2
1 β − if she watches program 2.
Each viewer gets a pure benefit v from watching a program of length T.
Without loss of generality, we set 1 T = , which is equivalent to having an hour of
programming. If program i is broadcast under FTA broadcast technology, then it
will contain advertising. Let a fraction
i
a of program i consist of advertisements,
with the remaining fraction ( ) 1
i
a − consisting of actual programming, say news or
music. We call fraction
i
a the ad space on program i.
The viewer’s distaste for advertisements is measured by the advertising
nuisance parameter δ . If the viewed program has ad space
i
a , then the viewer gets
a disutility
i
a δ from watching that program. We restrict the ad nuisance parameter
42

to values greater than 1. This restriction ensures that the viewer benefit of
watching a program of length 75 minutes that is interrupted by 15 minutes of
advertising is less than the viewer benefit of watching an advertising free program
of length 60 minutes.
If the viewer is located at position β on the Hotelling interval, her indirect
utility of watching program 1 is then
2
1 1
IU v a δ β = − − , while that of watching
program 2 is
2
2 2
IU v a δ β = − − . If the viewer does not watch TV, then her
indirect utility is
0
0 IU = . We set v sufficiently large so that both
1
IU and
2
IU are
always positive. Thus, every viewer watches TV, and there is full market
coverage. Every viewer can watch only one program, and thus chooses the
program that gives her higher indirect utility.
Suppose program content is differentiated i.e. both programs are produced.
Then, there is a consumer located at position
1
b on the Hotelling interval who is
indifferent between the two programs. All consumers to the left of
1
b prefer to
watch program 1, while all consumers to the right of
1
b prefer to watch program 2.
The audience shares of programs 1 and 2 are displayed below.

( )
1
2 1 2 1
2 1
1
1
b a a
b b
δ = + −  
 
= −

43

Total viewership of each program is then
1
Nb for program 1 and
2
Nb for
program 2. Suppose on the other hand that program content is duplicated i.e. both
broadcasters produce the same program. Then, the audience share of program 1 is
given below.

1 2
1
2 1 1 2
1 2
1 if
if
0 if
a a
b a a
a a
< 

= =


>


As before, the audience share of program 2 is
2 1
1 b b = − , and the total viewership
of each program is
1
Nb for program 1 and
2
Nb for program 2.
A large number of advertisers (also called producers) costlessly
manufacture identical goods. Each advertiser makes a profit of α per viewer when
the viewer (also called consumer) buys the product. The advertiser cannot sell to a
consumer unless the consumer sees the advertisement for the good. If a program
carries the advertisement for a particular producer’s good, then every viewer of
that program sees the advertisement and becomes a consumer of that producer’s
good. Producers can advertise on neither, either or both programs.
The broadcaster of program i sells a fraction
i
a of program time to
advertisers at a unit price
i
r . An advertisement on program i reaches
i
Nb viewers,
each of whom gives the producer a profit equal to α . Since producers are price
takers on the advertising market, profit maximization by producers implies that
44

i i
Nb r α = . As Gabszewicz et. al. (2004) point out, this condition implies that the
equilibrium price per viewer i.e. /
i i
r Nb is the same on both channels, and equals
the producer profit per viewer i.e. α .
Under FTA broadcast technology, the only source of revenue for
broadcasters is ad revenue
i i
a r . Using the fact that producers are price takers in the
advertising market, we can express revenue for the broadcaster of program i as
i i
Nb aα . Given an exogenously fixed cost of broadcasting K, the profit of the
broadcaster of program i is then
i i i
Nb a K π α = − .
The objective of the private broadcaster is to maximize profits. The
objective of the state broadcaster is to maximize welfare. Welfare in this economy
is the sum of broadcaster profits
( )
bro
π , advertiser profits
( )
ad
π , and consumer
surplus (CS). We treat the state broadcaster as a welfare maximizer since we feel
that the diverse nature of PSB objectives can best be captured by a welfare
function. Although it may seem strange to include the profit of the private
broadcaster in the state broadcaster’s objective function
1
, PSB objectives such as
“offering a wide range of quality in all program genres” may not be achieved if the
actions of a state broadcaster that is heedless of the effects of its actions on private
broadcaster profits results in the private broadcaster being driven out of the market
altogether.

1
Recall that social welfare includes broadcaster profits.
45

In light of the anecdotal evidence that Norris et. al. (2003) provide of some
state broadcasters engaging in audience share maximization type activities, it is
possible to conceive of a state broadcaster objective function that is a weighted
combination of profit and social welfare
2
. Given that the profit of the state
broadcaster is already included in welfare, this would amount to reweighting the
components of social welfare in favor of state broadcaster profit. However, there is
little evidence to suggest that this kind of audience share maximization activity by
state broadcasters is widespread enough to justify increasing the weight of own
profits in the state broadcaster’s objective function. The components of social
welfare are displayed below.

2
1
2 2
1 1
2
2
1
0
1,2
2
i
bro
i i
i
ad
i i i i
i i
b
i i
i
i
a r K
N b a a r
CS N v b a d
π
π α
δ β β
=
= =
=
=
= −
= −
 
= − −
 
 
∑
∑ ∑
∑ ∑
∫

Following Peitz and Valletti (2008), we can rearrange terms to express
social welfare as 2
co ad
W Nv W W K = + + − , where
co
W i.e. the welfare from
program content and
ad
W i.e. the welfare from advertising are displayed below.

2
This approach would be in the spirit of that adopted in the mixed economy literature, where the objective
of the state owned enterprise is usually a weighted combination of revenue and profit, in order to take
account of bureaucratic budget maximization activities.
46

( )
2
2
1
0
2
1
i
b
co
i
ad
i i
i
W N d
W N b a
β β
α δ
=
=
= −
= −
∑
∫
∑

The welfare from program content consists of the viewer transport costs
incurred from viewing a program whose content is different from that of the
viewer’s bliss point. As is usual in a Hotelling type model, transport costs are
lower under content differentiation than under content duplication, regardless of
the audience shares of broadcasters. Thus, welfare from program content is higher
under content differentiation than under content duplication. The limiting case
under content differentiation is when one broadcaster captures the entire audience
share. In this case, viewer transport costs are equal under content differentiation
and content duplication.
Welfare from advertising consists of the gains from trade between
consumers (viewers) and producers (advertisers), less the disutility to viewers
from watching advertisements. Clearly, advertising increases social welfare only if
the gains from trade exceed the disutility to viewers. Note that the gains from trade
accrue completely to the producer. Thus, welfare per ad is positive only if the
producer’s incremental profit from placing an advertisement exceeds the
incremental disutility to viewers from seeing the advertisement. Accordingly,
47

welfare from advertising is highest if all producers advertise when α δ > , and if
no producer advertises when α δ < .
We look for the Sub-game Perfect Nash Equilibria (SPNEs) under ME and
PE industry structures
3
. The stage 1 choices of PRITV are from the set
{ } 1,2
PRI
A = , and those of PUBTV from the set { } 1,2
PUB
A = , where 1 stands for
broadcast program 1, and 2 stands for broadcast program 2.
All terms in the analysis that follows are expressed per-person. Since we set
v high enough to ensure full coverage of the market, the equilibria will be the same
whether they are computed using per-person payoffs or gross payoffs. The
advantage in using per-person payoffs is that we have one less parameter to deal
with, namely N.
The only SPNEs that can occur under both ME and PE broadcasting market
structures are those where both broadcasters operate with content differentiation.
The private broadcaster prefers content differentiation because content duplication
would lead to Bertrand type competition in terms of advertising levels, and thus
zero advertising revenue. The state broadcaster prefers content differentiation
because content duplication would lead to a reduction in welfare from advertising
due to lost gains from trade between advertisers (producers) and viewers
(consumers) on account of zero ad space, as well a reduction in welfare from

3
All calculations for the equilibria are available upon request from the author.
48

program content due to increased viewer transport costs. As a result, both
broadcasters prefer content differentiation to content duplication in equilibrium,
regardless of broadcasting market structure (PE or ME). We now characterize and
compare the SPNEs under PE and ME.
The PE Case
In the PE case, both broadcasters are privately owned profit maximizers.
By the symmetry of the locations of the two programs on the unit Hotelling
interval, the sub-game equilibria yield the same profit levels for PRITV and
PUBTV as well as the same level of social welfare regardless of the broadcasters’
specific program choices. Recall that the SPNE involves content differentiation.
Without loss of generality, we characterize the sub-game equilibrium where
PRITV produces program 1 and PUBTV produces program 2.
Since program content is differentiated, we have
1 2
0 d d = = , and market
share is displayed below.

( )
1
2 1 2 1
2 1
1
1
b a a
b b
δ = + −  
 
= −

Under the PE broadcasting market structure, both broadcasters are privately
owned, and thus seek to maximize profits. The optimization problem confronting
49

PRITV (denoted by subscript 1) and PUBTV (denoted by subscript 2) is displayed
below.

[ ] [ ]
[ ] [ ]
1 1
2 2
1 1 1
0,1 0,1
2 2 2
0,1 0,1
max max
max max
a a
a a
b a k
b a k
π α
π α
∈ ∈
∈ ∈
= −
= −

From the Kuhn-Tucker formulation, the equilibrium choices of ad space
depend on the relationship between the advertiser profit per advertisement α and
the viewer nuisance per advertisement δ as displayed in Table 2.1. Superscript P
indicates PE equilibrium level. The socially optimal level of ad space is the same
on both broadcasters and equals 0 when α δ < , and equals 1 when α δ > .
Table 2.1: The FFSIM PE Equilibrium Actions
α δ
1
P
a
2
P
a
( ) 0,∞ ( ) 1,∞
1/δ 1/δ

If advertiser profit per advertisement is less than the viewer nuisance per
advertisement i.e. α δ < , welfare maximization requires that no producer be
allowed to advertise on either channel. However, in equilibrium, both channels
carry too much (positive levels of) advertising. If advertiser profit per
advertisement is greater than the viewer nuisance per advertisement i.e. α δ > ,
welfare maximization requires that all producers be allowed to advertise on both
50

channels. However, in equilibrium, both channels carry too little advertising
(recall that we impose the restriction 1 δ > ).
Note that there are two sources of market imperfections in the FTA FTA
industry structure: (1) viewers who receive benefits from watching programs
cannot be charged for watching programs due to the public good nature of the
broadcast under FTA broadcast technology, and (2) advertisements imposed a
negative externality on viewers, who experienced disutility at the rate of δ per
advertisement viewed. These imperfections are caused by the FTA nature of the
broadcast technology used by both broadcasters, as well as consumers’ attitude
towards advertising.
As a result of these market imperfections, we observe the Anderson and
Coate (2005) type pattern of sub-optimal provision of ad space by both
broadcasters, with social over provision of ad space when α δ < and social under
provision of ad space when α δ > . This is due to the fact that neither broadcaster
the incentive to equate the marginal social benefit of advertisements i.e. the
incremental gains from trade between advertisers and viewers from one more
advertisement, with the marginal social cost of advertising, i.e. the incremental
disutility to viewers from watching one more advertisement. The tradeoff relevant
to the profit maximizing broadcasters is the revenue gains from carrying more
51

advertisements versus the loss of market share due to carrying more
advertisements.
Consumer surplus, social welfare, and the payoffs of broadcasters are
displayed in Table 2.2. Recall that advertisers make zero profits since they are
price takers in the advertising market. We now analyze the ME case.
Table 2.2: Payoffs under FFSIM PE
1
P
π
2
P
π
P
CS
P
w
2
k
α
δ
−
2
k
α
δ
−
13
12
v−
13
12
2 v k
α
δ
− − +

The ME Case
Consider now the equilibrium under the ME market structure. Here, PRITV
is a profit maximizing privately owned broadcaster, while PUBTV is a social
welfare maximizing state owned broadcaster. By the symmetry of the locations of
the two programs on the unit Hotelling interval, the sub-game equilibria yield the
same profit levels for PRITV and PUBTV as well as the same level of social
welfare regardless of the broadcasters’ specific program choices. Recall that the
SPNE involves content differentiation. Without loss of generality, we characterize
the sub-game equilibrium where PRITV produces program 1 and PUBTV
produces program 2.
52

Since program content is differentiated, we have
1 2
0 d d = = , and market
share is displayed below.

( )
1
2 1 1 2
2 1
1
1
b a a
b b
δ = − −  
 
= −

The optimization problem confronting PRITV (denoted by subscript 1) and
PUBTV (denoted by subscript 2) is displayed below.

[ ] [ ]
[ ] [ ]
1 1
2 2
1 1 1
0,1 0,1
0,1 0,1
max max
max max 2
a a
ad co
a a
b a k
w v w w k
π α
∈ ∈
∈ ∈
= −
= + + −

From the Kuhn-Tucker formulation, the equilibrium choices of ad space
depend on the relationship between the advertiser profit per advertisement α and
the viewer nuisance per advertisement δ as displayed in Table 2.3 and Figure 2.1.
Superscript M indicates ME equilibrium level. The socially optimal level of ad
space is the same on both broadcasters and equals 0 when α δ < , and equals 1
when α δ > . The second order conditions for social welfare maximization require
that we have 2 δ α < , so we restrict attention to the parameter space
( ) ( ) ( )
1
2
, , 1,2 α δ α ∈ ∞ × . Broadcaster profits are displayed in Table 2.4. Consumer
surplus and social welfare are displayed in Table 2.5. Recall that advertisers make
zero profits since they are price takers in the advertising market.
53

Table 2.3: The FFSIM ME Equilibrium Actions
Zone α δ
1
M
a
2
M
a
1
( )
3
4
,∞ ( )
4
3
1,
α

( )
3 2
2
α δ
δ α δ
−
−

( )
4 3
2
α δ
δ α δ
−
−

2
( )
1
2
,∞
{ } ( )
4
3
max 1, ,2
α
α
1
2δ
0

Figure 2.1: The FFSIM ME Equilibrium

54

Table 2.4: Broadcaster Profits under FFSIM ME
Zone
1
M
π
2
M
π
1 ( )
( )
2
2
3 2
2 2
k
α α δ
δ α δ
−
−
−
( )
( )
2
2
4 3
2 2
k
α α δ
δ α δ
−
−
−
2
8
k
α
δ
−
0

Table 2.5: CS and Welfare under FFSIM ME
Zone
M
CS
M
w
1
( )
( )
2
9 4
7
3
4 2
v
α α δ
α δ
−
−
− +
( )
( )
13 4
7
3 4 2
2 v k
α α δ
δ α δ
−
−
− − +
2
13
48
v−
13
48 8
2 v k
α
δ
− − +

Comparing the PE and ME Equilibria
Just as in the PE case, in the ME case too one or both broadcasters over-
provides ad space when profit per ad exceeds viewer nuisance per ad, but under-
provides ad space when the viewer nuisance per ad exceeds the profit per ad.
However, the following proposition holds.
Proposition 2.1 In the simultaneous setting, both broadcasters carry more
optimal ad space under ME than under PE.
55

Proof. It follows from the expressions for equilibrium ad space that

1 2 1 2
1 2 1 2
when >
when <
P P M M
W
P P M M
W
a a a a a
a a a a a
α δ
α δ
= < < <
= > > ≥

▲
When α δ > , both broadcasters under-provide ad space under PE and ME.
When α δ < , both broadcasters over-provide ad space under PE and ME.
However, the extent of under and over provision of ad space by both broadcasters
is lower under ME. Since PUBTV acts as a welfare maximizer under ME, it is not
surprising that its choice of advertising space is more optimal under ME than
under PE. What is interesting to note is that PUBTV forces PRITV also to make
the more optimal ad space decisions under ME!
Recall that when viewer nuisance per ad is less than profit per ad, it is
optimal to have all producers advertising. Under PE, neither broadcaster cares
about the welfare consequences of their actions, hence ad space is under provided.
Under ME however, PUBTV cares about the welfare consequences of its actions,
and hence carries greater ad space than it would have under PE. This permits
PRITV also to carry greater ad space under ME than it would have under PE, since
there is strategic complementarity in choice of ad space. The welfare concerns of
PUBTV thus permit both broadcasters to carry more ad space under ME, and
hence reduce the extent of under provision of ad space.
56

Similarly, when viewer nuisance per ad is greater than profit per ad, it is
optimal to have no producers advertising. Under PE, neither broadcaster cares
about the welfare consequences of their actions, hence ad space is over provided.
Under ME however, PUBTV cares about the welfare consequences of its actions,
and hence carries lower ad space than it would have under PE. This forces PRITV
also to carry lower ad space under ME than it would have under PE. The welfare
concerns of PUBTV thus force both broadcasters to carry less ad space under ME,
and hence reduce the extent of over provision of ad space.
In a sense, the state broadcaster tailors the level of aggressiveness with
which it competes with the private broadcaster depending on the relationship
between the viewer nuisance from ads and the profitability of advertising. When
α δ > , PUBTV is a less aggressive competitor under ME, allowing PRITV to
carry more ads, and thus ameliorating the under provision of ad space. When
α δ < , PUBTV is a more aggressive competitor under ME, disciplining
4
PRITV
by forcing it to carry fewer ads, and thus ameliorating the over provision of ad
space. We now turn our attention to social welfare.
Proposition 2.2 In the simultaneous setting, social welfare under ME is
always greater than social welfare under PE
Proof. See Appendix▲

4
Shaw (1999, pg.46) reports that the disciplining role of a public broadcaster was summed up by Michael
Grade, formerly of the BBC, as follows: “The BBC is there to keep private broadcasters honest.”
57

We have already established that the ad space decisions of both broad
casters are more optimal under ME than under PE. It follows that welfare from
advertising is higher under ME than under PE. However, welfare from program
content is lower under ME than under PE. Symmetric advertising level choices by
identical private broadcasters in a PE cause viewers to choose between programs
solely on the basis of which program’s content is most preferred. However, in a
ME, the advertising space choices of both broadcasters are different. As a result,
some viewers under ME watch the program whose content is less preferred on
account of its carrying lower advertising levels than the other program. This leads
to larger viewer transport costs and hence lower welfare from program content
under ME.
Nevertheless, it can be shown that the welfare gains from more optimal ad
space decisions under ME outweigh the welfare losses from program content
regardless of the level of viewer nuisance from ads or the profitability of
advertising. Since all other components of social welfare
5
are the same under ME
and PE, it follows that social welfare under ME is always greater than that under
PE. We now turn our attention to broadcaster profits under ME.
Proposition 2.3 In the simultaneous setting, PRITV’s profit is higher (lower)
under ME than under PE when ( ) δ α < > .

5
The other components of social welfare are the pure viewer benefits of watching TV and the costs of
broadcasting.
58

Proof. See Appendix▲
When δ α < , we have
1 2 1 2
P P M M
W
a a a a a = < < < . Notice that PRITV (the
producer of program 1) carries higher ad space under ME than it does under PE.
Moreover, both PRITV and PUBTV have the same ad space under PE, but
PUBTV carries greater ad space than PRITV under ME. Thus, PRITV’s market
share is greater under ME than under PE. Greater ad space and greater market
share implies greater revenues and hence greater profits for PRITV under ME.
Thus, when the viewer nuisance per ad is less than the profit per ad, the state
broadcaster’s actions are to the benefit of the private broadcaster, in the sense of
allowing the private broadcaster a greater level of profit under ME than under PE.
The non-profit focus of the state broadcaster renders it a less aggressive
6

competitor than another private broadcaster.
When δ α > , we have
1 2 1 2
P P M M
W
a a a a a = > > ≥ . Notice that PRITV (the
producer of program 1) carries lower ad space under ME than it does under PE.
Moreover, both PRITV and PUBTV have the same ad space under PE, but PRITV
carries greater ad space than PUBTV under ME. Thus, PRITV’s market share is
lower under ME than under PE. Lower ad space and lower market share implies
lower revenues and hence lower profits for PRITV under ME. Thus, when the

6
By less aggressive, we mean that the state broadcaster carries a larger ad space than a private broadcaster
would under the same conditions.

59

viewer nuisance per ad is greater than the profit per ad, the state broadcaster’s
actions are to the detriment of the private broadcaster, in the sense of allowing the
private broadcaster a lower level of profit under ME than under PE. The non-profit
focus of the state broadcaster renders it a more aggressive competitor than another
private broadcaster.
This is the two-sided market analog of the result in Sappington and Sidak
(2003, pg.184), who find that “… a reduced focus on profit can provide state
enterprises with stronger incentives (than profit maximizing firms) to pursue
activities that disadvantage competitors.” Note that in the two-sided market
setting, the state broadcaster is not always a more aggressive competitor than a
private broadcaster – when δ α < , the state broadcaster’s actions are to the benefit
of the private broadcaster in the sense of allowing it a higher profit under ME than
it could have got under PE.
In the situation where δ α > , the private broadcaster ends up making lower
profits under ME. If the real world situation corresponds to the one where δ α > ,
then there is some justice to the complaints by private broadcasters of an uneven
playing field and overly aggressive competition from state broadcasters. How then
are we to know whether δ α > , or whether δ α < ?
The model finds that ad space is over provided under ME when δ α > . The
common perception is also that ad space is over provided; this is the reason why
60

regulations have been enacted the world over to curb over provision of ad space. If
the common perception is correct, we are in a situation where δ α > . In this case,
while there may be some justice to the complaints of private broadcasters, no
regulatory action is called for since welfare is invariably enhanced by the presence
and operation of state broadcasters. Of course, even if the common perception is
wrong, and we are in a situation where δ α < , no regulatory action is called for
since the complaints by private broadcasters could then be dismissed as false. We
now consider profits of the state broadcaster.
Proposition 2.4 In the simultaneous setting, PUBTV’s profit is higher (lower)
under ME than under PE when ( ) δ α < > .
Proof. See Appendix▲
When δ α < , PUBTV carries greater ad space and has lower market share
under ME. However, the positive revenue effects of greater ad space outweigh the
negative revenue effects of lower market share. For a given cost of broadcasting,
this translates into higher PUBTV profits under ME. Thus, it is possible for both
broadcasters to make higher profits under ME if the advertising nuisance level is
low enough. This is a “win-win-win” situation since both broadcasters make
higher profits, and society as a whole is better off.
When δ α > , PUBTV carries lower ad space and has higher market share
under ME. However, the negative revenue effects of lower ad space outweigh the
61

positive revenue effects of higher market share. For a given cost of broadcasting,
this translates into lower PUBTV profits under ME. In this situation, the welfare
gains under ME do not come for free. When the viewer nuisance per ad exceeds
the profit per ad, the increased aggression of the state broadcaster in terms of ad
space reductions results in welfare improving ad space choices by both state and
private broadcasters, but at the cost of the state broadcaster making lower profits
under ME than under PE.
When
( )
5 4
4 3
,
α α
δ ∈ , the state broadcaster makes lower profits than the
private broadcaster under ME. This would tend to support the contention that the
real world corresponds to the case where δ α > , since it is not uncommon to find
that state broadcasters are less profitable than their private sector counterparts.
When
4
3
α
δ > , the state broadcaster does not carry any ads, and therefore makes
losses from operating. Clearly, it would require some form of budgetary support in
order to perform its PSB mandate. As long as the deadweight losses involved in
raising funds to support the state broadcaster do not negate the welfare gains from
having an operational state broadcaster, a welfare argument could be made to
support the continuing operations of loss making state broadcasters. We now turn
our attention to the sequential setting.
62

The Sequential Setting
In the sequential setting under ME, we focus attention on the case where
the broadcaster who enters the industry first is the state broadcaster. This is
because the observed pattern of entry in most world television markets is that the
earlier entrant is usually a state broadcaster, while the latter entrant is usually a
private broadcaster.
We adapt the models of Gabszewicz et. al. (2004) and Nasser et. al. (2007),
where the three participants in the industry i.e. viewers, advertisers and
broadcasters, together play the following multi-stage game. In stage 1, the first
broadcaster enters the industry, and chooses its program type and advertising
space. In stage 2, the second broadcaster enters the industry, observes the program
type and advertising space choices of the first broadcaster, and then chooses its
own program type and advertising space. In stage 3, advertisers choose which
broadcasters to advertise with, and consumers make their program viewing
choices and goods purchase decisions. Aside from the change in the sequence of
actions, all other aspects of players’ preferences and actions remain the same as in
the simultaneous setting.
The only SPNEs that can occur under both ME and PE broadcasting market
structures are those where both broadcasters operate with content differentiation.
The private broadcaster prefers content differentiation because content duplication
63

would lead to Bertrand type competition in terms of advertising levels, and thus
zero advertising revenue. The state broadcaster prefers content differentiation
because content duplication would lead to a reduction in welfare from advertising
due to lost gains from trade between advertisers (producers) and viewers
(consumers) on account of zero ad space, as well a reduction in welfare from
program content due to increased viewer transport costs. As a result, both
broadcasters prefer content differentiation to content duplication in equilibrium,
regardless of broadcasting market structure (PE or ME). We now characterize and
compare the SPNEs under PE and ME.
The PE Case
In the PE market structure, both broadcasters are privately owned profit
maximizers. By the symmetry of the locations of the two programs on the unit
Hotelling interval, the sub-game equilibria yield the same profit levels for PRITV
and PUBTV as well as the same level of social welfare regardless of the
broadcasters’ specific program choices. Recall that the SPNE involves content
differentiation. Without loss of generality, we characterize the sub-game
equilibrium where PRITV produces program 1 and PUBTV produces program 2.
Since program content is differentiated, we have
1 2
0 d d = = , and market
share is displayed below.
64

( )
1
2 1 1 2
2 1
1
1
b a a
b b
δ = − −  
 
= −

Since PUBTV moves first, it must calculate the response of PRITV to its
choice of program type and advertising space. Thus, PUBTV solves the
optimization problem of PRITV for a given choice of program type and
advertising space. The optimization problem of PRITV is displayed below.

[ ]
( )
[ ]
( )
1 1
1 2 1 2 1
0,1 0,1
max max
a a
a b a a k π α
∈ ∈
= −
The use of the parenthesis term indicates the fact that the optimization
problem is solved for a particular choice of program type (set equal to program 2
for ease of exposition) and advertising space by PUBTV. Solving PRITV’s
optimization problem yields PRITV’s best response as ( )
2
1
2 1 2
a P
s a
δ +
= . Plugging
this back into the profit function of PUBTV gives the following optimization
problem for PUBTV.

[ ]
( ) ( )
[ ]
( ) ( )
2 2
2 1 2 2 1 2 2
0,1 0,1
max max
P P
a a
a a b a a a k π α
∈ ∈
= −
From the Kuhn-Tucker formulation, the equilibrium choice of ad space by
each broadcaster depends on the relationship between the advertiser profit per
advertisement α and the viewer nuisance per advertisement δ as displayed in
Table 2.6 and Figure 2.2. Superscript P indicates PE equilibrium level. The
socially optimal level of ad space is the same on both broadcasters and equals 0
65

when α δ < , and equals 1 when α δ > . As in the case of the simultaneous setting,
we restrict the viewer nuisance per advertisement to assume values greater than 1.
We restrict the advertiser profit per advertisement to lie above
2
δ
, since this is
necessary to satisfy the second order condition for welfare maximizing choice of
advertising space by PUBTV under ME market structure. Broadcaster profits are
displayed in Table 2.7. Consumer surplus and social welfare are displayed in
Table 2.8. Recall that advertisers make zero profits since they are price takers in
the advertising market.
Table 2.6: The FFSEQ PE Equilibrium Actions
Zone α δ
1
P
a
2
P
a
1
( )
2
,
δ
∞ ( )
3
2
1,
1
2
δ +
1
2
( )
2
,
δ
∞ ( )
3
2
,∞
5
4δ

3
2δ

66

Figure 2.2: The FFSEQ PE Equilibrium

Table 2.7: Broadcaster Profits under FFSEQ PE
Zone
1
P
π
2
P
π
1
( )
2
1
8
k
δ
δ
+
−
( ) 3
4
k
α δ −
−
2
25
32
k −
9
16
k
α
δ
−

67

Table 2.8: CS and Welfare under FFSEQ PE
Zone
P
CS
P
w
1
( ) 13 14
48
v
δ δ + −
−
( ) ( )
2
8 1
14
13
48 8 16
2 v k
α δ δ
δ δ
δ
+ −
−
− − + −
2
277
192
v−
277 43
192 32
2 v k
α
δ
− − +

If advertiser profit per advertisement is less than the viewer nuisance per
advertisement i.e. α δ < , welfare maximization requires that no producer be
allowed to advertise on either broadcaster. However, in equilibrium, both
broadcasters carry too much advertising. If advertiser profit per advertisement is
greater than the viewer nuisance per advertisement i.e. α δ > , welfare
maximization requires that all producers be allowed to advertise on both
broadcasters. However, in equilibrium, both broadcasters carry too little
advertising.
Note that there are two sources of market imperfections in the FTA FTA
industry structure: (1) viewers who receive benefits from watching programs
cannot be charged for watching programs due to the public good nature of the
broadcast under FTA broadcast technology, and (2) advertisements imposed a
negative externality on viewers, who experienced disutility at the rate of δ per
advertisement viewed. These imperfections are caused by the FTA nature of the
68

broadcast technology used by both broadcasters, as well as consumers’ attitude
towards advertising.
As a result of these market imperfections, we observe the Anderson and
Coate (2005) type pattern of sub-optimal provision of ad space by both
broadcasters, with social over provision of ad space when α δ < and social under
provision of ad space when α δ > . This is due to the fact that neither broadcaster
the incentive to equate the marginal social benefit of advertisements i.e. the gains
from trade between advertisers and viewers, with the marginal social cost of
advertising, i.e. the disutility to viewers from watching advertisements. The
tradeoff relevant to the profit maximizing broadcasters is the revenue gains from
carrying more advertisements versus the loss of market share due to carrying more
advertisements.
It is interesting to compare the PE equilibrium in the sequential setting with
that in the simultaneous setting. Consider the payoffs of PRITV, who enters the
industry after PUBTV. It can be shown that PRITV makes greater profits under PE
in the sequential setting than in the simultaneous setting regardless of the values of
the advertiser profit per ad or the viewer nuisance per ad. Thus, the well known
result of Gal-Or (1985) regarding second mover advantage when there is strategic
complementarity extends to this particular two sided markets setting.
69

The ME Case
In the ME market structure, PRITV is a privately owned profit maximizer,
while PUBTV is a state owned welfare maximizer. By the symmetry of the
locations of the two programs on the unit Hotelling interval, the sub-game
equilibria yield the same profit levels for PRITV and PUBTV as well as the same
level of social welfare regardless of the broadcasters’ specific program choices.
Recall that the SPNE involves content differentiation. Without loss of generality,
we characterize the sub-game equilibrium where PRITV produces program 1 and
PUBTV produces program 2.
Since program content is differentiated, we have
1 2
0 d d = = , and market
share is displayed below.

( )
1
2 1 2 1
2 1
1
1
b a a
b b
δ = + −  
 
= −

Since PUBTV (the state broadcaster) moves first, it must calculate the
response of PRITV (the private broadcaster) to its choice of program type,
advertising space, and advertising price. Thus, PUBTV solves the optimization
problem of PRITV for a given choice of program type, advertising space, and
advertising price. The optimization problem of PRITV is displayed below.

[ ]
( )
[ ]
( )
1 1
1 2 1 2 1
0,1 0,1
max max
a a
a b a a k π α
∈ ∈
= −
70

The use of the parenthesis term indicates the fact that the optimization
problem is solved for a particular choice program type (set equal to program 2 for
ease of exposition), advertising space, and advertising price by PUBTV. Solving
PRITV’s optimization problem yields PRITV’s best response as ( )
2
1
2 1 2
a P
a a
δ
δ
+
= .
Plugging this back into the profit function of PUBTV gives the following
optimization problem for PUBTV.

[ ]
( ) ( )
[ ]
( ) ( ) ( ) ( )
2 2
1 2 1 2 1 2
0,1 0,1
max max 2
P co P ad P
a a
w a a v w a a w a a k
∈ ∈
= + + −
From the Kuhn-Tucker formulation, the equilibrium choices of ad space
depend on the relationship between the advertiser profit per advertisement α and
the viewer nuisance per advertisement δ as displayed in Table 2.9 and associated
Figure 2.3. Superscript M indicates ME equilibrium level. The socially optimal
level of ad space is the same on both broadcasters and equals 0 when α δ < , and
equals 1 when α δ > . As in the case of the simultaneous setting, we restrict the
viewer nuisance per advertisement to assume values greater than 1. We restrict the
advertiser profit per advertisement to lie above
2
δ
, since this is necessary to satisfy
the second order condition for welfare maximizing choice of advertising space on
PUBTV under ME market structure. Broadcaster profits are displayed in Table
2.10. Consumer surplus and social welfare are displayed in Table 2.11. Recall that
advertisers make zero profits since they are price takers in the advertising market.
71

Table 2.9: The FFSEQ ME Equilibrium Actions
Zone α δ
1
M
a
2
M
a
1
( ) 2 , δ ∞ ( ) 1,3
1
2
δ
δ
+
1
2
( )
4
5
,2
δ
δ ( ) 1,∞
( )
5 4
2
α δ
δ α δ
−
−

( )
8 7
2
α δ
δ α δ
−
−

3
( )
4
2 5
,
δ δ
( ) 1,∞
1
2δ
0

Figure 2.3: The FFSEQ ME Equilibrium

72

Table 2.10: Broadcaster Profits under FFSEQ ME
Zone
1
M
π
2
M
π
1
( )
2
1
8
k
α δ +
−
( ) 3
4
k
α δ −
−
2
( )
( )
2
2
5 4
2 2
k
α α δ
δ α δ
−
−
−
( )( )
( )
2
8 7 2
2 2
k
α α δ δ α
δ α δ
− −
−
−
3
8
k
α
δ
− k −

Table 2.11: CS and Welfare under FFSEQ ME
Zone
M
CS
M
w
1
( ) 13 3 14
48
v
δ δ + −  
 
−
( ) ( )
2
8 1
14
13
48 8 16
2 v k
α δ δ
δ δ
δ
+ −
−
− − + −
2
( )
2
2
10 9
3
4 2
v
α
α δ −
− +
( )
( )
17 4
10
3 4 2
2 v k
α α δ
δ α δ
−
−
− − +
3
13
48
v−
13
48 8
2 v k
α
δ
− − +

Just as in the case of the sequential PE equilibrium, in the sequential ME
equilibrium also we have both broadcasters setting sub-optimal levels of
advertising space. The pattern of sub-optimality is the same, with both
broadcasters over-providing advertising space when α δ < , and under-providing
advertising space when α δ > . As usual, the reason for the sub-optimality is that
PRITV does not have the incentive to equate the marginal social benefit of
73

advertising with its marginal social cost. Although PUBTV has the incentive to do
this, the strategic complementarity between actions of PRITV and PUBTV means
that over-provision of advertising space by PRITV leads to over-provision of
advertising space by PUBTV. We now compare broadcasters’ choices of viewer
subscription fee and advertising space under the PE and ME sequential settings.
Comparing the PE and ME Equilibria
Proposition 2.5 Broadcasters’ choices of advertising space in the sequential
setting may be less optimal under ME when α δ > .
Proof. Follows from the expressions for viewer subscription fee and
advertising space ▲
Contrary to the simultaneous setting, it is not always the case that both
broadcasters make more optimal ad space decisions under ME in the sequential
setting. In fact, the ad space decisions of both broadcasters are more optimal under
ME only when the advertiser profit per ad is less than the viewer nuisance per ad
i.e. when α δ < . When the advertiser profit per ad is more than the viewer
nuisance per ad i.e. when α δ > , broadcasters’ choices of ad space may or may
not be more optimal under ME depending on the exact values of the advertiser
profit per ad and the viewer nuisance per ad.
The second mover advantage of PRITV on account of the strategic
complementarity in the choice of advertising space may counteract the more
74

optimal choice of ad space by PUBTV when α δ > . When α δ > , there is under-
provision of ad space by both broadcasters in equilibrium. Although PUBTV
would like to increase its ad space in order to mitigate the under provision of ad
space, its ability to do so is curtailed by the second mover advantage of PRITV,
which has incentive to undercut PUBTV and thus gain market share. Given that
this would result in a loss of welfare from program content on account of
increased viewer transport costs, PUBTV is restrained in its ability to guide the
industry towards a mutual ad space increase. For the right parameter values, in
order to stave off welfare losses from viewer transport costs, it is possible that
PUBTV would set even lower ad space than a private broadcaster would similar
conditions, thus worsening the problem of under-provision of ad space. This is the
reason why it is possible, but not necessary, to have less optimal ad space choices
under ME when α δ > . We shall see consequently that such regions may also be
associated with lower welfare under ME than under PE.
Proposition 2.6 Both broadcasters make lower profits under ME when α δ < .
Proof. See Appendix ▲
When α δ < , there is social over-provision of ad space on both
broadcasters in equilibrium. Given its welfare concerns, PUBTV curtails its ad
space more than it would have under PE, resulting in PRITV curtailing its own ad
space in response, in order to avoid losing market share. Lower ad space on both
75

broadcasters results in lower profits for both broadcasters under ME. No clear
pattern regarding broadcasters’ profits under PE and ME is apparent when α δ > ,
with both broadcasters sometimes making greater profit under PE, and sometimes
making greater profit under ME. Subject to net effect of the two forces i.e. (1) the
potential tradeoff between welfare from program content and welfare from
advertising under both simultaneous and sequential settings, and (2) the second
mover advantage of the private broadcaster under sequential settings, the private
broadcaster may end up with greater ad space and / or greater audience share
under ME, in which case it could make greater profits under ME, or vice versa.
From the point of view of the private broadcaster therefore, under certain
parameter values, there is justice to their complaints of earning lower profits under
ME on account of the “anti-competitive” practices of the state broadcaster. Thus,
the results of Sappington and Sidak (2003) hold for certain parameter values in
this two sided market setting as well. What needs to be determined is whether
these “anti-competitive” practices should be tolerated in view of their potential for
welfare improvements.
Proposition 2.7 Social welfare in the sequential setting is equal or higher
under ME than under PE except when ( ) ( ) ( )
3
2
, 2 , ,3 α δ δ ∈ ∞ × .
Proof. See Appendix ▲
76

Contrary to the case in the simultaneous setting, it is not always the case
that the presence of a state broadcaster leads to a welfare improvement. The region
in which social welfare is lower under ME than under PE i.e. where
( ) ( ) ( )
3
2
, 2 , ,3 α δ δ ∈ ∞ × , is easily identified as the only one in which the state
broadcaster carries 100% advertising space. Only in this region would there be a
welfare decrease on account of the operation of a state broadcaster. In all other
parameter regions, social welfare is enhanced by the operation of a state
broadcaster.
Note that in this region, we have α δ > , so there is under-provision of ad
space by both broadcasters in both PE and ME equilibria, although the extent of
under-provision of ad space is lower in the ME equilibrium. Consequently, the
welfare from advertising is higher under ME in this parameter region. However,
welfare from program content is lower under ME in this parameter region, on
account of excessively skewed market share in favor of the private broadcaster.
The market share of PRITV under PE is
5
8
, while that under ME ranges from
5
8
to
1 as δ increases from
3
2
to 1. Recall that welfare from program content is
maximum when the audience is shared equally between channels, and decreases as
the market share of any one broadcaster increases beyond
1
2
. As δ increases,
PRITV’s market share becomes so large under ME that the welfare losses from
program content under ME begin to outweigh the welfare gains from more optimal
77

ad space choices under ME, with the result that total welfare under ME becomes
smaller than total welfare under PE.
The appropriate regulatory response in this region is to regulate ad space on
the state broadcaster, not to privatize the state broadcaster. This is because
regulation of ad space can yield welfare improvements that are greater than those
that could be realized by privatization of the state broadcaster. Ironically, this
region is characterized by a reduced profit for the state broadcaster, but an
increased profit for the private broadcaster vis-à-vis the PE setting. Therefore, in
the FFSEQ setting in general, private broadcasters’ complaints about overly
aggressive competition from the state broadcaster can safely be ignored, since the
only situations in which such complaints of reduced profits are valid are also those
situations where social welfare is improved.
Summary of Results
We now summarize the main findings in this chapter, dealing with the
simultaneous setting first and the sequential setting next. In the simultaneous
setting, with regard to ad space decisions, we found that competition between the
state and the private broadcaster causes the private broadcaster to choose levels of
ad space that are more optimal than when the private broadcaster competes with
another private broadcaster. Since the state broadcaster also chooses more optimal
78

levels of ad space than it would if it were privately owned, welfare from
advertising is higher under ME than under PE.
Recall that when viewer nuisance per ad is less than profit per ad, it is
optimal to have all producers advertising. Under PE, neither broadcaster cares
about the welfare consequences of their actions, hence ad space is under provided.
Under ME however, PUBTV cares about the welfare consequences of its actions,
and hence carries greater ad space than it would have under PE. This permits
PRITV also to carry greater ad space under ME than it would have under PE, since
there is strategic complementarity in choice of ad space. The welfare concerns of
PUBTV thus permit both broadcasters to carry more ad space under ME, and
hence reduce the extent of under provision of ad space.
Similarly, when viewer nuisance per ad is greater than profit per ad, it is
optimal to have no producers advertising. Under PE, neither broadcaster cares
about the welfare consequences of their actions, hence ad space is over provided.
Under ME however, PUBTV cares about the welfare consequences of its actions,
and hence carries lower ad space than it would have under PE. This forces PRITV
also to carry lower ad space under ME than it would have under PE. The welfare
concerns of PUBTV thus force both broadcasters to carry less ad space under ME,
and hence reduce the extent of over provision of ad space.
79

With regard to program content, we find that broadcasters prefer content
differentiation to content duplication in equilibrium, regardless of broadcasting
market structure (ME or PE). This is because content duplication would lead to a
race to the bottom in terms of advertising levels and thus zero advertising revenue
for both broadcasters, as well as a reduction in social welfare on account of (1) lost
opportunities for gains from trade between advertisers and viewers, and (2)
increased viewer transport costs.
Welfare from program content is lower under ME than under PE.
Symmetric advertising level choices by identical private broadcasters in a PE
cause viewers to choose between programs solely on the basis of which program’s
content is most preferred. However, in a ME, the advertising level choices of both
broadcasters are different. As a result, some viewers under ME watch the program
whose content is less preferred on account of its carrying lower advertising levels
than the other program. This leads to lower welfare from program content under
ME. However, we show that the welfare gains from more optimal ad space
decisions more than outweigh the welfare losses from program content, regardless
of the profitability of advertising or the nuisance value of advertising. Therefore,
total welfare under ME is always higher than total welfare under PE in the
simultaneous setting.
80

Private broadcaster profit is lower under ME when the nuisance value of
advertising exceeds the profitability of advertising. However, when the
profitability of advertising exceeds the nuisance value of advertising, private
broadcaster profit is actually higher under ME! Thus, it is possible to have a “win-
win-win” situation under ME, with both broadcasters as well as society as a whole
being better off.
When the profitability of advertising exceeds the nuisance value of
advertising, competing with a state broadcaster allows the private broadcaster to
carry a larger ad space as well as have a greater market share than when
competing with another private broadcaster. For a given cost of broadcasting, this
gives the private broadcaster greater revenues and thus greater profits under ME.
However, if the nuisance value of advertising exceeds the profitability of
advertising, competing with a state broadcaster forces the private broadcaster to
curtail its ad space to a greater extent as well as have a lower market share than
when competing with another private broadcaster. For a given cost of
broadcasting, this gives the private broadcaster lower revenues and thus lower
profits under ME.
Therefore, depending on the level of nuisance from ads and the profitability
of advertising, the state broadcaster takes actions that are either to the benefit of,
or to the detriment of the private broadcaster (where benefit means profit and
81

detriment means loss). This result is the two sided market analog of the result in
Sappington and Sidak (2003, pg.184), who find that a “reduced focus on profit can
provide state enterprises with stronger incentives to pursue activities that
disadvantage competitors.” When the nuisance value of advertising exceeds it
profitability, there may be some justice to the complaints of private broadcasters
of an uneven playing field and overly aggressive competition from state
broadcasters, but no regulatory action is called for since welfare in the
simultaneous setting under ME is invariably enhanced by the presence and
operation of state broadcasters.
Under certain conditions on the nuisance level and profitability of
advertising, the state broadcaster may incur losses while attempting to carry out its
PSB mandate. In such situations, if the deadweight losses incurred in raising
public funds to provide budgetary support to the state broadcaster is low enough, a
welfare argument can be made for the continued existence and operation of state
broadcasters, even if they are loss making. Privatizing the state broadcaster,
although it may result in increased profits for the state broadcaster, would
inevitably reduce social welfare on account of less optimal ad space decisions by
the privatized entity.
We now turn our attention to the sequential setting. With regard to the
choice of ad space, we note that both broadcasters carry more optimal ad space
82

under ME when the advertiser profit per ad is less than the viewer nuisance per ad,
but not always when the advertiser profit per ad exceeds the viewer nuisance per
ad. The second mover advantage of PRITV on account of the strategic
complementarity in the choice of advertising space may counteract the more
optimal choice of ad space by PUBTV when advertiser profit per ad exceeds the
viewer nuisance per ad.
Recall that when advertiser profit per ad exceeds the viewer nuisance per
ad, there is under-provision of ad space by both broadcasters in equilibrium.
Although PUBTV would like to increase its ad space in order to mitigate the under
provision of ad space, its ability to do so is curtailed by the second mover
advantage of PRITV, which has incentive to undercut PUBTV and thus gain
market share. Given that this would result in a loss of welfare from program
content on account of increased viewer transport costs, PUBTV is restrained in its
ability to guide the industry towards a mutual ad space increase. For the right
parameter values, in order to stave off welfare losses from viewer transport costs,
it is possible that PUBTV would set even lower ad space than a private
broadcaster would similar conditions, thus worsening the problem of under-
provision of ad space. This is the reason why it is possible, but not necessary, to
have less optimal ad space choices under ME when advertiser profit per ad
exceeds the viewer nuisance per ad.
83

With regard to social welfare, contrary to the case in the simultaneous
setting, it is not always the case that the presence of a state broadcaster leads to a
welfare improvement. The parameter region in which social welfare is lower under
ME than under PE i.e. when ( ) ( ) ( )
3
2
, 2 , ,3 α δ δ ∈ ∞ × is easily identified as the only
one in which the state broadcaster carries 100% advertising space. Only in such a
parameter region would there be a welfare decrease on account of the operation of
a state broadcaster.
Note that in this region, we have α δ > , so there is under-provision of ad
space by both broadcasters in both PE and ME equilibria, although the extent of
under-provision of ad space is lower in the ME equilibrium. Consequently, the
welfare from advertising is higher under ME in this parameter region. However,
welfare from program content is lower under ME in this parameter region, on
account of excessively skewed market share in favor of the private broadcaster.
The market share of PRITV under PE is
5
8
, while that under ME ranges from
5
8
to
1 as δ increases from
3
2
to 3. Recall that welfare from program content is
maximum when the audience is shared equally between channels, and decreases as
the market share of any one broadcaster increases beyond
1
2
. As δ increases,
PRITV’s market share becomes so large under ME that the welfare losses from
program content under ME begin to outweigh the welfare gains from more optimal
84

ad space choices under ME, with the result that total welfare under ME becomes
smaller than total welfare under PE.
The appropriate response however is not to privatize the state broadcaster,
but merely to regulate the quantum of advertising it is allowed to carry – a
judicious regulation of advertising space can then deliver a welfare improvement
over the outcome that would prevail if privatization of the state broadcaster was
instead adopted. Since this parameter region is associated with increased profit for
the private broadcaster vis-à-vis the PE setting, it is clear that a private broadcaster
would not feel obliged to complain against the state broadcaster to competition
authorities. Therefore, private broadcasters’ complaints about overly aggressive
competition from the state broadcaster can safely be ignored, since the only
situations in which such complaints of reduced profits are valid are also those
situations where social welfare is improved.
Interestingly, when we compare the sequential PE setting to the
simultaneous PE setting, we find that the private broadcaster always makes greater
profits in the sequential setting. Recall that the private broadcaster moves second
in the sequential setting. Thus, the well known result of Gal-Or (1985) regarding
second mover advantage in the presence of strategic complementarity extends to
this particular two sided markets setting.

85
Chapter III
The FTA PTV Industry Structure
In this chapter, we consider program content and advertising levels in a
mixed economy television broadcast industry under the FTA PTV industry
structure. The FTA PTV industry structure is modeled as one where one
broadcaster uses the FTA broadcast technology, while the other broadcaster uses
the PTV broadcast technology. With regard to competition between advertisers,
we consider a homogenous product market structure. We treat the choice of
program content, advertising intensity, and viewer subscription fee as endogenous,
and contrast the equilibrium in a mixed economy with that in a purely private
economy industry structure.
We analyze competition between broadcasters under two settings: (1) a
simultaneous setting, and (2) a sequential setting. The simultaneous setting
involves both broadcasters entering the industry at the same time, while the
sequential setting involves one broadcaster entering the industry before the other.
In the sequential setting, we focus attention on the case where the broadcaster who
enters the industry first uses the FTA broadcast technology, while the broadcaster
who enters the industry subsequently uses the PTV broadcast technology. We
focus on this case because the PTV broadcast technology is newer than the FTA
broadcast technology.

86
After the introduction of the PTV broadcast technology, the observed
pattern in most world television markets is that the earlier entrant, usually a state
broadcaster, used the FTA broadcast technology, while the latter entrant, usually a
private broadcaster, used the PTV broadcast technology. The chapter is organized
as follows. The next section analyzes the simultaneous setting. The subsequent
section analyzes the sequential setting, and the final section summarizes and
concludes the discussion.
The Simultaneous Setting
Consider first the simultaneous setting where both broadcasters enter the
industry simultaneously. We adapt the models of Gabszewicz et. al. (2004) and
Nasser et. al. (2007), where the three participants in the industry i.e. viewers,
advertisers and broadcasters, together play the following multi-stage game. In
stage 1, both broadcasters simultaneously choose program type. In stage 2, both
broadcasters simultaneously choose the level of advertising if they use the FTA
broadcast technology, and the viewer subscription fee if they use the PTV
broadcast technology. In stage 3, advertisers choose which broadcasters to
advertise with, and consumers make their program viewing choices and goods
purchase decisions.
Consider first the decision problem of viewers. There is a mass N of
viewers. Viewers are distributed uniformly along the Hotelling interval, with taste

87
for programs ( ) 0,1 U β ∼ . There are two program types e.g. news and music,
indexed by 1 and 2. Producing program 1 is equivalent to the broadcaster choosing
a location of 0 on the Hotelling interval. Producing program 2 is equivalent to the
broadcaster choosing a location of 1 on the Hotelling interval.
Program choice of PRITV is denoted by distance
1
d from the end of the
Hotelling interval indexed by 0, and program choice of PUBTV by distance of
2
d
from the end of the Hotelling interval indexed by 1. Thus, if both PRITV and
PUBTV produced program 1, then ( ) ( )
1 2
, 0,1 d d = . Similarly, if both PRITV and
PUBTV produced program 2, then ( ) ( )
1 2
, 1,0 d d = . These are the cases where
program content is duplicated by broadcasters. If PRITV produced program 1 and
PUBTV produced program 2, or vice versa, then ( ) ( )
1 2
, 0,0 d d = or ( ) ( )
1 2
, 1,1 d d =
respectively. These are the cases where program content is differentiated by
broadcasters.
With quadratic transportation costs, a viewer located at position β on the
Hotelling interval is at a distance β from program 1, and hence incurs a disutility
2
β if she watches program 1. Similarly, she is at a distance of ( ) 1 β − from
program 2, and hence incurs a disutility of ( )
2
1 β − if she watches program 2.
Each viewer gets a pure benefit v from watching a program of length T.
Without loss of generality, we set 1 T = , which is equivalent to having an hour of

88
programming. If program i is broadcast under FTA broadcast technology, then it
will contain advertising. Let a fraction
i
a of program i consist of advertisements,
with the remaining fraction ( ) 1
i
a − consisting of actual programming, say news or
music. We call fraction
i
a the ad space on program i.
The viewer’s distaste for advertisements is measured by the advertising
nuisance parameter δ . If the viewed program has ad space
i
a , then the viewer gets
a disutility
i
a δ from watching that program. We restrict the ad nuisance parameter
to values greater than 1. This restriction ensures that the viewer benefit of
watching a program of length 75 minutes that is interrupted by 15 minutes of
advertising is less than the viewer benefit of watching an advertising free program
of length 60 minutes.
If program j is broadcast under PTV broadcast technology, then it will not
contain advertisements, i.e. 0
j
a = . Viewers will however have to pay a
subscription fee
j
s in order to view the program. Without loss of generality, we let
program 1 be broadcast under PTV broadcast technology, while program 2 is
broadcast under FTA broadcast technology. If the viewer is located at position β
on the Hotelling interval, her indirect utility of watching program 1 is then
2
1 1
IU v s β = − − , while that of watching program 2 is
2
2 2
IU v a δ β = − − . If the
viewer does not watch TV, then her indirect utility is
0
0 IU = . We set v

89
sufficiently large so that both
1
IU and
2
IU are always positive. Thus, every
viewer watches TV, and there is full market coverage. Every viewer can watch
only one program, and thus chooses the program that gives her higher indirect
utility.
Suppose program content is differentiated i.e. both programs are produced.
Then, there is a consumer located at position
1
b on the Hotelling interval who is
indifferent between the two programs. All consumers to the left of
1
b prefer to
watch program 1, while all consumers to the right of
1
b prefer to watch program 2.
The audience shares of programs 1 and 2 are displayed below.

[ ]
1
2 1 2 1
2 1
1
1
b a s
b b
δ = + −
= −

Total viewership of each program is then
1
Nb for program 1 and
2
Nb for
program 2. Suppose on the other hand that program content is duplicated i.e. both
broadcasters produce the same program. Then, the audience share of program 1 is
given below.

1 2
1
2 1 1 2
1 2
1 if
if
0 if
s a
b s a
s a
δ
δ
δ
< 

= =


>


90
As before, the audience share of program 2 is
2 1
1 b b = − , and the total viewership
of each program is
1
Nb for program 1 and
2
Nb for program 2.
A large number of advertisers (also called producers) costlessly
manufacture goods. Each advertiser makes a profit of α per viewer when the
viewer (also called consumer) buys the product. The advertiser cannot sell to a
consumer unless the consumer sees the advertisement for the good. If a program
carries the advertisement for a particular producer’s good, then every viewer of
that program sees the advertisement and becomes a consumer of that producer’s
good. Producers can advertise on neither, either or both programs.
The broadcaster of program 2 sells a fraction
2
a of program time to
advertisers at a unit price
2
r . An advertisement on program 2 reaches
2
Nb
viewers, each of whom gives the producer a profit equal to α . Since producers are
price takers on the advertising market, profit maximization by producers implies
that
2 2
Nb r α = . As Gabszewicz et. al. (2004) point out, this condition implies that
the equilibrium price per viewer i.e.
2 2
/ r Nb equals the producer profit per viewer
i.e. α .
Under FTA broadcast technology, the only source of revenue for the
broadcaster of program 2 is ad revenue
2 2
a r . Using the fact that producers are
price takers in the advertising market, we can express revenue for the broadcaster

91
of program 2 as
2 2
Nb aα . Given an exogenously fixed cost of broadcasting K, the
profit of the broadcaster of program 2 is then
2 2 2
Nb a K π α = − .
Under the PTV broadcast technology, the only source of revenue for the
broadcaster of program 1 is subscription fees from viewers. Revenue from
subscription fees from viewers is
1 1
Nb s . Given an exogenously fixed cost of
broadcasting K, the profit of the broadcaster of program 1 is then
1 1 1
Nb s K π = − .
The objective of the private broadcaster is to maximize profits. The
objective of the state broadcaster is to maximize welfare. Welfare in this economy
is the sum of broadcaster profits
( )
bro
π , advertiser profits
( )
ad
π , and consumer
surplus (CS). We treat the state broadcaster as a welfare maximizer since we feel
that the diverse nature of PSB objectives can best be captured by a welfare
function. Although it may seem strange to include the profit of the private
broadcaster in the state broadcaster’s objective function
1
, PSB objectives such as
“offering a wide range of quality in all program genres” may not be achieved if the
actions of a state broadcaster that is heedless of the effects of its actions on private
broadcaster profits results in the private broadcaster being driven out of the market
altogether.
In light of the anecdotal evidence that Norris et. al. (2003) provide of some
state broadcasters engaging in audience share maximization type activities, it is

1
Recall that social welfare includes broadcaster profits.

92
possible to conceive of a state broadcaster objective function that is a weighted
combination of profit and social welfare
2
. Given that the profit of the state
broadcaster is already included in welfare, this would amount to reweighting the
components of social welfare in favor of state broadcaster profit. However, there is
little evidence to suggest that this kind of audience share maximization activity by
state broadcasters is widespread enough to justify increasing the weight of own
profits in the state broadcaster’s objective function. The components of social
welfare are displayed below.

1 1 2 2
2 2 2 2
2
1 1 2 2
0
1,2
2
i
bro
ad
b
i
Nb s a r K
Nb a a r
CS N v b s b a d
π
π α
δ β β
=
= + −
= −
 
= − − −
 
 
∑
∫

Following Peitz and Valletti (2008), we can rearrange terms to express
social welfare as 2
co ad
W Nv W W K = + + − , where
co
W i.e. the welfare from
program content and
ad
W i.e. the welfare from advertising are displayed below.

( )
2
0
1,2
2 2
i
b
co
i
ad
W N d
W Nb a
β β
α δ
=
= −
= −
∑
∫

2
This approach would be in the spirit of that adopted in the mixed economy literature, where the objective
of the state owned enterprise is usually a weighted combination of revenue and profit, in order to take
account of bureaucratic budget maximization activities.

93
The welfare from program content consists of the viewer transport costs
incurred from viewing a program whose content is different from that of the
viewer’s bliss point. As is usual in a Hotelling type model, transport costs are
lower under content differentiation than under content duplication, regardless of
the audience shares of broadcasters. Thus, welfare from program content is higher
under content differentiation than under content duplication. The limiting case
under content differentiation is when one broadcaster captures the entire audience
share. In this case, viewer transport costs are equal under content differentiation
and content duplication.
Welfare from advertising consists of the gains from trade between
consumers (viewers) and producers (advertisers), less the disutility to viewers
from watching advertisements. Clearly, advertising increases social welfare only if
the gains from trade exceed the disutility to viewers. Note that the gains from trade
accrue completely to the producer. Thus, welfare per ad is positive only if the
producer’s incremental profit from placing an advertisement exceeds the
incremental disutility to viewers from seeing the advertisement. Accordingly,
welfare from advertising is highest if all producers advertise when α δ > , and if
no producer advertises when α δ < .

94
We look for the Sub-game Perfect Nash Equilibria (SPNEs) under ME and
PE industry structures
3
. Let the two broadcasters be called PRITV and PUBTV.
The stage 1 choices of PRITV are from the set { } 1,2
PRI
A = , and those of PUBTV
from the set { } 1,2
PUB
A = , where 1 stands for broadcast program 1, and 2 stands
for broadcast program 2.
All terms in the analysis that follows are expressed per-person. Since we set
v high enough to ensure full coverage of the market, the equilibria will be the same
whether they are computed using per-person payoffs or gross payoffs. The
advantage in using per-person payoffs is that we have one less parameter to deal
with, namely N.
The only SPNEs that can occur under both ME and PE broadcasting market
structures are those where both broadcasters operate with content differentiation.
The private broadcaster prefers content differentiation because content duplication
would lead to Bertrand type competition in terms of advertising levels and
subscription fees, and thus zero advertising and subscription revenue. The state
broadcaster prefers content differentiation because content duplication would lead
to a reduction in welfare from advertising due to lost gains from trade between
advertisers (producers) and viewers (consumers) on account of zero ad space, as
well a reduction in welfare from program content due to increased viewer

3
All calculations for the equilibria are available upon request from the author.

95
transport costs. As a result, both broadcasters prefer content differentiation to
content duplication in equilibrium, regardless of broadcasting market structure (PE
or ME). We now characterize and compare the SPNEs under PE and ME.
The PE Case
In the PE case, both broadcasters are privately owned profit maximizers.
By the symmetry of the locations of the two programs on the unit Hotelling
interval, the sub-game equilibria yield the same profit levels for PRITV and
PUBTV as well as the same level of social welfare regardless of the broadcasters’
specific program choices. Recall that the SPNE involves content differentiation.
Without loss of generality, we characterize the sub-game equilibrium where
PRITV produces program 1 and PUBTV produces program 2.
Since program content is differentiated, we have
1 2
0 d d = = , and market
share is displayed below.

[ ]
1
2 1 2 1
2 1
1
1
b a s
b b
δ = + −
= −

Under the PE broadcasting market structure, both broadcasters are privately
owned, and thus seek to maximize profits. The optimization problem confronting
PRITV (denoted by subscript 1) and PUBTV (denoted by subscript 2) is displayed
below.

96

[ ] [ ]
1 1
2 2
1 1 1
2 2 2
0,1 0,1
max max
max max
s s
a a
b s k
b a k
π
π α
∈ ∈
= −
= −

From the Kuhn-Tucker formulation, the equilibrium choices of viewer
subscription fee and ad space depend on the relationship between the advertiser
profit per advertisement α and the viewer nuisance per advertisement δ as
displayed in Table 3.1. Superscript P indicates PE equilibrium level. The socially
optimal levels of viewer subscription fee and ad space are respectively
( ) ( )
1 2
, 0,0
W W
s a = when α δ < , and
( ) ( )
1 2
, ,1
W W
s a α = when α δ > .
Table 3.1: The FPSIM PE Equilibrium Actions
α δ
1
P
s
2
P
a
( ) 0,∞ ( ) 1,∞
1 1/δ

If advertiser profit per advertisement is less than the viewer nuisance per
advertisement i.e. α δ < , welfare maximization requires that no producer be
allowed to advertise on PUBTV (the broadcaster of program 2), while the
optimum level of viewer subscription fee for PRITV (the broadcaster of program
1) is zero. However, in equilibrium, PRITV charges a higher than optimal viewer
subscription fee, and PUBTV carries too much advertising.

97
If advertiser profit per advertisement is greater than the viewer nuisance per
advertisement i.e. α δ > , welfare maximization requires that all producers be
allowed to advertise on PUBTV (the broadcaster of program 2), while the
optimum level of viewer subscription fee for PRITV (the broadcaster of program
1) is set equal to the advertiser profit per ad. However, in equilibrium, PRITV
charges a lower than optimal viewer subscription fee, and PUBTV carries too little
advertising.
Recall that there were two sources of market imperfections in the FTA FTA
industry structure: (1) viewers who receive benefits from watching programs
cannot be charged for watching programs due to the public good nature of the
broadcast under FTA broadcast technology, and (2) advertisements imposed a
negative externality on viewers, who experienced disutility at the rate of δ per
advertisement viewed. These imperfections were caused by the FTA nature of the
broadcast technology used by both broadcasters.
In the FTA PTV industry structure, both market imperfections are partially
alleviated by the fact that one broadcaster uses the PTV broadcast technology
under which (1) consumers can be charged for the privilege of watching programs,
and (2) programs are advertisement free, thereby mitigating both market
imperfections. Of course, both market imperfections cannot be eliminated entirely
due to the operation of the broadcaster that uses the FTA broadcast technology.

98
In spite of the fact that the FTA PTV industry structure partially alleviates
both market imperfections, we still observe the Anderson and Coate (2005) type
pattern of sub-optimal provision of ad space by PUBTV, with social over
provision of ad space when α δ < and social under provision of ad space when
α δ > . Once again, this is due to the fact that PUBTV does not have the incentive
to equate the marginal social benefit of advertisements i.e. the gains from trade
between advertisers and viewers, with the marginal social costs of advertising, i.e.
the disutility to viewers from watching advertisements. The tradeoff relevant to the
profit maximizing PUBTV is the revenue gains from carrying more
advertisements versus the loss of market share due to carrying more
advertisements.
Not only does PUBTV carry sub-optimal levels of advertising, but PRITV
also charges sub-optimal levels of viewer subscription fees. Just as in the case of
PUBTV, this sub-optimality is due to the fact that PRITV does not have the
incentive to equate the marginal social benefit of advertisements with the marginal
social costs of advertising. The particular pattern of sub-optimality in viewer
subscription fees observed on the part of PRITV is over-charging when α δ < and
under charging when α δ > . This pattern of sub-optimality is similar to that
observed in the case of PUBTV, and is a consequence of the fact that there is
strategic complementarity in the choices of viewer subscription fee and advertising

99
space, as can be readily observed from the reaction functions of each broadcaster.
Therefore, when α δ > , PUBTV carries too little advertising space, and PRITV
under-charges viewer subscription fees, and vice versa. Given that the partial
mitigation of market imperfections cannot eliminate sub-optimal choices by both
broadcasters in equilibrium, there may still be scope for a welfare improvement
under ME. Consumer surplus and the payoffs of broadcasters are displayed in
Table 3.2. Recall that advertisers make zero profits since they are price takers in
the advertising market.
Table 3.2: Payoffs under FPSIM PE
1
P
π
2
P
π
P
CS
P
w
1
2
k −
2
k
α
δ
−
13
12
v−
7
12 2
2 v k
α
δ
− − +

Before analyzing the equilibrium under ME, it is interesting to contrast
social welfare in the FTA PTV industry structure under PE with that in the FTA
FTA industry structure under PE. It can be shown that even though both market
imperfections are alleviated to a degree under FTA PTV industry structure as
compared to no alleviation of market imperfections under FTA FTA industry
structure, social welfare is lower under FTA PTV industry structure when α δ > ,
and vice versa. This is an interesting application of the general theory of second
best. As Lipsey and Lancaster (1957, pg.12) observe, “… it is not true that a

100
situation in which more, but not all, of the optimum conditions are fulfilled (e.g.
FTA PTV) is necessarily, or is even likely to be, superior to a situation in which
fewer are fulfilled (e.g. FTA FTA).”
In the present context, welfare is lower under FTA PTV when α δ >
because this is the situation where optimality demands that every producer be
allowed to advertise – since programming is advertisement free under PTV
broadcast technology and the channel using the FTA broadcast technology carries
the same amount of advertising under both FTA PTV and FTA FTA, it is no
wonder that welfare is lower under FTA PTV. This is a perverse case where
partially alleviating both market imperfections through the use of PTV broadcast
technology actually lowered welfare!
Similarly, welfare is higher under FTA PTV when α δ < because this is the
situation where optimality demands that no producer be allowed to advertise –
since programming is advertisement free under PTV broadcast technology and the
channel using the FTA broadcast technology carries the same amount of
advertising under both FTA PTV and FTA FTA, it is no wonder that welfare is
higher under FTA PTV.
The ME Case
Consider now the equilibrium under the ME market structure. Here, PRITV
is a profit maximizing privately owned broadcaster, while PUBTV is a social

101
welfare maximizing state owned broadcaster. By the symmetry of the locations of
the two programs on the unit Hotelling interval, the sub-game equilibria yield the
same profit levels for PRITV and PUBTV as well as the same level of social
welfare regardless of the broadcasters’ specific program choices. Recall that the
SPNE involves content differentiation. Without loss of generality, we characterize
the sub-game equilibrium where PRITV produces program 1 and PUBTV
produces program 2.
Since program content is differentiated, we have
1 2
0 d d = = , and market
share is displayed below.

[ ]
1
2 1 2 1
2 1
1
1
b a s
b b
δ = + −
= −

The optimization problem confronting PRITV (denoted by subscript 1) and
PUBTV (denoted by subscript 2) is displayed below.

[ ] [ ]
1 1
2 2
1 1 1
0,1 0,1
max max
max max 2
s s
ad co
a a
b s k
w v w w k
π
∈ ∈
= −
= + + −

From the Kuhn-Tucker formulation, the equilibrium choices of viewer
subscription fee and ad space depend on the relationship between the advertiser
profit per advertisement α and the viewer nuisance per advertisement δ as
displayed in Table 3.3 and Figure 3.1. Superscript M indicates ME equilibrium

102
level. The socially optimal levels of viewer subscription fee and ad space are
respectively
( ) ( )
1 2
, 0,0
W W
s a = when α δ < , and
( ) ( )
1 2
, ,1
W W
s a α = when α δ > . The
second order conditions for social welfare maximization require that
2
δ
α > , so we
restrict attention to the parameter space ( ) ( ) ( )
2
, , 1,
δ
α δ ∈ ∞ × ∞ . Broadcaster profits
are displayed in Table 3.4. Consumer surplus and social welfare are displayed in
Table 3.5. Recall that advertisers make zero profits since they are price takers in
the advertising market.
Table 3.3: The FPSIM ME Equilibrium Actions
Zone α δ
1
M
s
2
M
a
1
( )
2
3
,
δ
∞ ( ) 1,∞ 1
1
δ

2
( )
2
2 3
,
δ δ
( ) 1,3
1
2
1
2
1
δ +






1
1
0
δ






3
( )
2
2 3
,
δ δ
( ) 3,∞
1
2
1




1
0
δ




103
Figure 3.1: The FPSIM ME equilibrium

Table 3.4: Broadcaster Profits under FPSIM ME
Zone
1
M
π
2
M
π
1
1
2
k −
2
k
α
δ
−
2 ( )
2
1
8
1
2
1
8
k
k
k
δ +

−


−


−



( ) 3
4
2
k
k
k
α δ
α
δ
−

−

−


−


3
1
2
1
8
k
k
− 

−


2
k
k
α
δ
−


−


104
Table 3.5: CS and Welfare under FPSIM ME
Zone
M
CS
M
w
1
13
12
v−
7
12 2
2 v k
α
δ
− − +
2
( ) 13 3 14
48
13
12
13
48
v
v
v
δ δ + −

−

−


−


( ) ( ) 3 10 3
7
48 4 16
7
12 2
7
48
2
2
2
v k
v k
v k
α δ δ δ
α
δ
− −

− − + −

− − +


− −


3
13
12
13
48
v
v
− 

−


7
12 2
7
48
2
2
v k
v k
α
δ
− − + 

− −


Comparing PE and ME Equilibria
Notice that in the ME case, there are multiple equilibria when ( )
2
2 3
,
δ δ
α∈ .
Specifically, there are three equilibria when ( ) ( ) ( )
2
2 3
, , 1,3
δ δ
α δ ∈ × . It can be
shown that the ranking by each broadcaster of the equilibria in terms of their
payoffs are as follows: PRITV (the producer of program 1) ranks the equilibria as
( ) ( ) ( )
1 1 1
2 2 1 1 1
,1 1, ,0
δ
δ
π π π
+
> > , while PUBTV (the producer of program 2) ranks
the equilibria as ( ) ( ) ( )
1 1 1
2 2
,1 1, ,0 w w w
δ
δ
+
< < . Since neither broadcaster agrees on
the ranking of the equilibria, no equilibrium dominates the others in terms of
payoffs of both broadcasters. Consequently, it is not possible to determine ex ante
which equilibrium will manifest itself when ( ) ( ) ( )
2
2 3
, , 1,3
δ δ
α δ ∈ × . Note that
equilibrium 2 is the same as the PE equilibrium.

105
Consider the choices of viewer subscription fee by PRITV and advertising
space by PUBTV. In this parameter space, we have α δ < . Just as in the case of
PE, both broadcasters choose sub-optimally high levels of viewer subscription fee
and advertising space in all three equilibria. However, both broadcasters choose
the same or lower levels of viewer subscription fee and advertising space in
equilibria 2 and 3 than they do under PE. Thus, in equilibria 2 and 3 under ME, the
extent of over-charging of viewer subscription fee and over-provision of
advertising space is the same or lower than it would have been under PE. Since
PUBTV acts as a welfare maximizer under ME, it is not surprising that its choice
of advertising space is the same or more optimal under equilibria 2 and 3 of ME
than under PE. What is interesting to note is that PUBTV forces PRITV also to
make the same or more optimal choice of viewer subscription fee.
Recall that when advertiser profit per ad is less than viewer nuisance per ad,
it is optimal to have no producers advertising with the FTA broadcaster, as well as
a zero viewer subscription fee on the PTV broadcaster. Under PE, neither
broadcaster cares about the welfare consequences of their actions, hence PRITV
over-charges viewer subscription fees and PUBTV over-supplies advertising
space. Under ME however, PUBTV cares about the welfare consequences of its
actions, and hence carries the same or lower advertising space in equilibria 2 and
3. This “aggressive” action by PUBTV forces PRITV to charge the same or lower

106
viewer subscription fee in equilibria 2 and 3 than it would have under PE, since
there is strategic complementarity in the choices of advertising space and viewer
subscription fee. The welfare concerns of PUBTV therefore force both
broadcasters to make decisions that are at least optimal under ME as they would
be under PE in equilibria 2 and 3.
In equilibrium 1 however, both PRITV and PUBTV choose higher levels of
viewer subscription fee and advertising space than they would have under PE, thus
worsening the problem of over-charging of viewer subscription fee and over-
supply of advertising space. This is surprising since we have the perverse outcome
of the welfare maximizing PUBTV choosing a less optimal level of advertising
space under ME than it would have if it was a privately owned entity operating
under PE. This may occur because of (1) the strategic complementarity between
the decisions of viewer subscription fee and advertising space, and (2) the low
level of viewer disutility from advertising.
At low levels of viewer disutility from advertising, a choice of higher
viewer subscription fee by PRITV will force PUBTV to carry a larger advertising
space. This is because PUBTV has to trade off the welfare from program content
(viewer transport costs of watching non preferred programs) with the welfare from
advertising (gains from trade from advertising). Choosing a low level of
advertising space will minimize the loss of welfare from advertising when α δ < ,

107
but will result in large welfare loss from program content since PUBTV will end
up with a large market share, which will result in large viewer transport costs. At
low levels of viewer disutility from advertising, the tradeoff may be in favor of
PUBTV choosing a higher advertising space since welfare losses from advertising
are likely to be lower than welfare losses from program content. This is how it is
possible to have a perverse outcome where PUBTV chooses higher advertising
space under ME than it would under PE. Note that at higher levels of viewer
disutility from advertisements, PUBTV would always resolve its tradeoff in favor
of minimizing welfare loss from advertising at the cost of increased welfare loss
from program content, so it would not be possible to have this kind of perverse
outcome.
There are two equilibria when ( ) ( ) ( )
2
2 3
, , 3,
δ δ
α δ ∈ × ∞ . It can be shown that
the ranking by each broadcaster of the equilibria in terms of their payoffs are as
follows: PRITV (the producer of program 1) ranks the equilibria as
( ) ( )
1 1
2 1 1
1, ,0
δ
π π > , while PUBTV (the producer of program 2) ranks the equilibria
as ( ) ( )
1 1
2
1, ,0 w w
δ
< . Since neither broadcaster agrees on the ranking of the
equilibria, no equilibrium dominates the other in terms of payoffs of both
broadcasters. Consequently, it is not possible to determine ex ante which
equilibrium will manifest itself when ( ) ( ) ( )
2
2 3
, , 3,
δ δ
α δ ∈ × ∞ .

108
In this parameter space, we have α δ < . Even though there are multiple
equilibria, the relatively higher values of viewer disutility from advertising mean
that the perverse outcome where PUBTV chooses a higher level of advertising
space than it would under PE does not arise. In both equilibria, we have both
broadcasters choosing the same or more optimal levels of viewer subscription fee
and advertising space under ME than they would have under PE.
Consider now the parameter space ( ) ( ) ( )
2
3
, , 1,
δ
α δ δ ∈ × ∞ . Here too we
have α δ < . There is a unique equilibrium, which is the same as the PE
equilibrium, in which both broadcasters over-charge viewer subscription fee and
over-supply advertising space. In the parameter space ( ) ( ) ( ) , , 1, α δ δ ∈ ∞ × ∞ , we
have α δ > . There is a unique equilibrium, which is the same as the PE
equilibrium, in which both broadcasters under-charge viewer subscription fee and
under-supply advertising space. Thus, at relatively higher levels of advertiser
profit per advertisement, regardless of the level of viewer nuisance per
advertisement, we have a unique equilibrium that is the same as the PE
equilibrium. In these parameter spaces, both broadcasters do not make either more
or less optimal decisions regarding viewer subscription fee and advertising space
than they would have under PE, even though one of the broadcasters i.e. PUBTV
is a state owned welfare maximizer! Summarizing the discussion on broadcasters’

109
choices of viewer subscription fee and advertising space, we have the following
proposition.
Proposition 3.1 Broadcasters’ choices of viewer subscription fee and
advertising space are the same or more optimal under ME than under PE when
( ) ( ) ( )
2
, , 3,
δ
α δ ∈ ∞ × ∞ .
Proof. Follows from the expressions for viewer subscription fee and
advertising space ▲
We now turn our attention to social welfare.
Proposition 3.2 Social welfare is the same or higher under ME than under PE
when ( ) ( ) ( )
2
, , 3,
δ
α δ ∈ ∞ × ∞ .
Proof. See Appendix ▲
We have already established that the viewer subscription fee and
advertising space decisions of both broadcasters are the same or more optimal
under ME than under PE for all possible equilibria in this parameter region. It
follows that the welfare from advertising is the same or higher under ME than
under PE. Note that in equilibrium under PE, we have
1 2
P P
s a δ = . We call this
choice of viewer subscription fee and advertising space “symmetric choice”.
Under symmetric choice, the viewer chooses between programs solely on the basis
of which program’s content is most preferred. However, under ME, we may or
may not have symmetric choice depending on the parameter space. In regions

110
where we do not have symmetric choice, some viewers under ME may watch the
program whose content is less preferred since the utility loss from paying the
viewer subscription fee is lower than the utility loss from watching
advertisements. In such regions, we have larger viewer transport costs, and hence
lower welfare from program content under ME.
Nevertheless, it can be shown that the welfare gains from the same or more
optimal decisions of viewer subscription fee and advertising space under ME
outweigh the welfare losses from program content whenever we have
( ) ( ) ( )
2
3
, , 3,
δ
α δ ∈ ∞ × ∞ . Since all other components of social welfare are the same
under ME and PE, it follows that welfare under ME is the same or greater than
welfare under PE whenever ( ) ( ) ( )
2
3
, , 3,
δ
α δ ∈ ∞ × ∞ . Therefore, we can realize a
welfare improvement under ME, albeit in a restricted parameter space.
Only when ( ) ( ) ( )
2
2 3
, , 1,3
δ δ
α δ ∈ × do we have the possibility of a perverse
equilibrium where both broadcasters choose levels of ad space and viewer
subscription fee that are more sub-optimal under ME, leading to lower welfare
under ME. This equilibrium is easily recognizable by the fact that it is the only
equilibrium in which PUBTV carries 100% ad space. The correct regulatory
response in this situation is not to privatize PUBTV, but merely to regulate
(curtail) the ad space on PUBTV. While privatization of PUBTV might increase
social welfare, it would deliver a welfare improvement smaller than could be

111
achieved by the optimal regulation of ad space on PUBTV. We now turn our
attention to broadcaster profits under ME.
Proposition 3.3 PRITV profit is the same or lower under ME than under PE
when ( ) ( ) ( )
2
, , 3,
δ
α δ ∈ ∞ × ∞ .
Proof. See Appendix ▲
In this parameter space, we have PRITV charging the same or lower viewer
subscription fee under ME than it does under PE. Moreover, PRITV has the same
or lower audience share under ME than it does under PE
4
. Since viewer
subscription fees are the only source of revenue of PRITV, the same or lower
viewer subscription fees in addition to the same or lower market share translates
into the same or lower profits for PRITV under ME.
Thus, regardless of whether the advertiser profit per ad is greater or less
than the viewer nuisance per ad in this parameter region, we have the actions of
the state broadcaster being to the detriment of the private broadcaster in the sense
of allowing the private broadcaster only the same or lower profits under ME than
under PE. The non profit focus of the state broadcaster renders it a more
aggressive
5
competitor than another private broadcaster.

4
PRITV market share under ME is either ½ or ¼ depending on the parameter values, while its market share
under PE is always ½

5
By more aggressive, we mean that the state broadcaster carries an advertising space that is less than or
equal to the advertising space that would be carried by a private broadcaster under the same conditions.

112
Thus, the two sided market analog of the result in Sappington and Sidak
(2003) holds for the case of FTA PTV industry structure as well. Recall that the
Sappington and Sidak (2003, pg.184) result was “… a reduced focus on profit can
provide state enterprises with stronger incentives (than profit maximizing firms) to
pursue activities that disadvantage competitors.” There is therefore some justice to
the complaints of private broadcasters of an uneven playing field and overly
aggressive competition from state broadcasters. However, no regulatory action is
called for since welfare is potentially improved by the presence and operation of
state broadcasters.
The only parameter region where welfare is lower under ME is in the case
of the perverse equilibrium (equilibrium 1) when ( ) ( ) ( )
2
2 3
, , 1,3
δ δ
α δ ∈ × , i.e. where
the state broadcaster’s advertising space is 100% of programming time. However,
since this situation rarely arises in practice, and is easily detected even if it does
arise (since advertising space is observable), appropriate regulatory action can
then be undertaken in the form of curtailing advertising space on the state
broadcaster.
Proposition 3.4 PUBTV profit is the same or lower under ME than under PE
when ( ) ( ) ( )
2
, , 3,
δ
α δ ∈ ∞ × ∞ .
Proof. See Appendix ▲

113
In this parameter space, we have PUBTV carrying the same or lower
advertising space under ME than it does under PE. Moreover, PUBTV has the
same or lower audience share under ME than it does under PE
6
. Since advertising
revenue is the only source of revenue of PUBTV, the same or lower advertising
space in addition to the same or lower market share translates into the same or
lower profits for PUBTV under ME.
Thus, the potential welfare gains under ME for the FTA PTV industry
structure too do not come for free. The increased aggression of the state
broadcaster in terms of advertising space reductions results in welfare improving
choices of viewer subscription fee and advertising space by both broadcasters.
However, this comes at the cost of the state broadcaster making lower profits
under ME than under PE. In the case where the state broadcaster does not carry
any advertisements at all (e.g. BBC), it would clearly require some sort of
budgetary support in order to perform its public service mandate. As long as the
deadweight losses involved in raising funds to support the state broadcaster do not
negate the welfare gains from having an operational state broadcaster, a welfare
argument can be made to support the continuing operations of loss making state
broadcasters.

6
PRITV market share under ME is either ½ or 0 depending on the parameter values, while its market share
under PE is always ½

114
As regards the consumer surplus of viewers, it is clear that it is the same or
higher under ME than it would be under PE (except in the case of the perverse
equilibrium). This is because the profits of both broadcasters are the same or lower
under ME than under PE, advertisers make zero profits under both ME and PE,
and consumer surplus is the only other component of welfare, which is the same or
higher under ME than under PE. Contrary to the case of the FTA FTA industry
structure however, it is not possible to have a “win-win-win” situation, since the
cost of the improvement in social welfare and consumer surplus is the lower
profits of both broadcasters. We now turn our attention to the sequential setting.
The Sequential Setting
In the sequential setting, we focus attention on the case where the
broadcaster who enters the industry first uses the FTA broadcast technology, while
the broadcaster who enters the industry subsequently uses the PTV broadcast
technology. We focus on this case because the PTV broadcast technology is newer
than the FTA broadcast technology. After the introduction of the PTV broadcast
technology, the observed pattern in most world television markets was that the
earlier entrant, usually a state broadcaster, used the FTA broadcast technology,
while the latter entrant, usually a private broadcaster, used the PTV broadcast
technology.

115
We adapt the models of Gabszewicz et. al. (2004) and Nasser et. al. (2007),
where the three participants in the industry i.e. viewers, advertisers and
broadcasters, together play the following multi-stage game. In stage 1, the FTA
broadcaster enters the industry, and chooses its program type and advertising
space. In stage 2, the PTV broadcaster enters the industry, observes the program
type and advertising space choices of the FTA broadcaster, and then chooses its
own program type and viewer subscription fee. In stage 3, advertisers choose
which broadcasters to advertise with, and consumers make their program viewing
choices and goods purchase decisions. Aside from the change in the sequence of
actions, all other aspects of players’ preferences and actions remain the same as in
the simultaneous setting.
The only SPNEs that can occur under both ME and PE broadcasting market
structures are those where both broadcasters operate with content differentiation.
The private broadcaster prefers content differentiation because content duplication
would lead to Bertrand type competition in terms of advertising levels and
subscription fees, and thus zero advertising and subscription revenue. The state
broadcaster prefers content differentiation because content duplication would lead
to a reduction in welfare from advertising due to lost gains from trade between
advertisers (producers) and viewers (consumers) on account of zero ad space, as
well a reduction in welfare from program content due to increased viewer

116
transport costs. As a result, both broadcasters prefer content differentiation to
content duplication in equilibrium, regardless of broadcasting market structure (PE
or ME). We now characterize and compare the SPNEs under PE and ME.
The PE Case
In the PE market structure, both broadcasters are privately owned profit
maximizers. By the symmetry of the locations of the two programs on the unit
Hotelling interval, the sub-game equilibria yield the same profit levels for PRITV
and PUBTV as well as the same level of social welfare regardless of the
broadcasters’ specific program choices. Recall that the SPNE involves content
differentiation. Without loss of generality, we characterize the sub-game
equilibrium where PRITV produces program 1 and PUBTV produces program 2.
Since program content is differentiated, we have
1 2
0 d d = = , and market
share is displayed below.

[ ]
1
2 1 2 1
2 1
1
1
b a s
b b
δ = + −
= −

Since PUBTV (the FTA broadcaster) moves first, it must calculate the
response of PRITV (the PTV broadcaster) to its choice of program type and
advertising space. Thus, PUBTV solves the optimization problem of PRITV for a
given choice of program type and advertising space. The optimization problem of
PRITV is displayed below.

117
( ) ( )
1 1
1 2 1 2 1
max max
s s
a b a s k π = −
The use of the parenthesis term indicates the fact that the optimization
problem is solved for a particular choice of advertising space and program type
(set equal to program 2 for ease of exposition) by PUBTV. Solving PRITV’s
optimization problem yields PRITV’s best response as ( )
2
1
2 1 2
a P
s a
δ +
= . Plugging
this back into the profit function of PUBTV gives the following optimization
problem for PUBTV.

[ ]
( ) ( )
[ ]
( ) ( )
2 2
2 1 2 2 1 2 2
0,1 0,1
max max
P P
a a
s a b s a a k π α
∈ ∈
= −
From the Kuhn-Tucker formulation, the equilibrium choices of viewer
subscription fee and ad space depend on the relationship between the advertiser
profit per advertisement α and the viewer nuisance per advertisement δ as
displayed in Table 3.6 and Figure 3.2. Superscript P indicates PE equilibrium
level. The socially optimal levels of viewer subscription fee and ad space are
respectively
( ) ( )
1 2
, 0,0
W W
s a = when α δ < , and
( ) ( )
1 2
, ,1
W W
s a α = when α δ > . As
in the case of the simultaneous setting, we restrict the viewer nuisance per
advertisement to assume values greater than 1. We restrict the advertiser profit per
advertisement to lie above
3
4
δ
, since this is necessary to satisfy the second order
condition for welfare maximizing choice of advertising space on PUBTV under
ME market structure. Thus, the words “No Eqm” in Figure 3.2 refer to the absence

118
of equilibrium under ME, but not under PE. Broadcaster profits are displayed in
Table 3.7. Consumer surplus and social welfare are displayed in Table 3.8. Recall
that advertisers make zero profits since they are price takers in the advertising
market.
Table 3.6: The FPSEQ PE Equilibrium Actions
Zone α δ
1
P
s
2
P
a
1
( )
3
4
,
δ
∞ ( )
3
2
1,
1
2
δ +
1
2
( )
3
4
,
δ
∞ ( )
3
2
,∞
5
4

3
2δ

Table 3.7: Broadcaster Profits under FPSEQ PE
Zone
1
P
π
2
P
π
1
( )
2
1
8
k
δ +
−
( ) 3
4
k
α δ −
−
2
25
32
k −
9
16
k
α
δ
−

Table 3.8: CS and Welfare under FPSEQ PE
Zone
P
CS
P
w
1
( )
1
48
13 14 v δ δ − + −  
 

( ) ( ) 3 10 3
7
48 4 16
2 v k
α δ δ δ − −
− − + −
2
277
192
v−
127 9
192 16
2 v k
α
δ
− − +

119
Figure 3.2: The FPSEQ PE equilibrium

If advertiser profit per advertisement is less than the viewer nuisance per
advertisement i.e. α δ < , welfare maximization requires that no producer be
allowed to advertise on PUBTV (the broadcaster of program 2), while the
optimum level of viewer subscription fee for PRITV (the broadcaster of program
1) is zero. However, in equilibrium, regardless of the value of viewer nuisance per
advertisement, PRITV charges a higher than optimal viewer subscription fee, and
PUBTV carries too much advertising.
If advertiser profit per advertisement is greater than the viewer nuisance per
advertisement i.e. α δ > , welfare maximization requires that all producers be
allowed to advertise on PUBTV (the broadcaster of program 2), while the

120
optimum level of viewer subscription fee for PRITV (the broadcaster of program
1) is set equal to the advertiser profit per ad. However, in equilibrium, regardless
of the value of viewer nuisance per advertisement, PRITV charges a lower than
optimal viewer subscription fee, and PUBTV carries too little advertising.
Recall that there were two sources of market imperfections in the FTA FTA
industry structure: (1) viewers who receive benefits from watching programs
cannot be charged for watching programs due to the public good nature of the
broadcast under FTA broadcast technology, and (2) advertisements imposed a
negative externality on viewers, who experienced disutility at the rate of δ per
advertisement viewed. These imperfections were caused by the FTA nature of the
broadcast technology used by both broadcasters.
In the FTA PTV industry structure, both market imperfections are partially
alleviated by the fact that one broadcaster uses the PTV broadcast technology
under which (1) consumers can be charged for the privilege of watching programs,
and (2) programs are advertisement free, thereby mitigating both market
imperfections. Of course, both market imperfections cannot be eliminated entirely
due to the operation of the broadcaster that uses the FTA broadcast technology.
In spite of the fact that the FTA PTV industry structure partially alleviates
both market imperfections, we still observe the Anderson and Coate (2005) type
pattern of sub-optimal provision of ad space by PUBTV, with social over

121
provision of ad space when α δ < and social under provision of ad space when
α δ > . Once again, this is due to the fact that PUBTV does not have the incentive
to equate the marginal social benefit of advertisements i.e. the gains from trade
between advertisers and viewers, with the marginal social costs of advertising, i.e.
the disutility to viewers from watching advertisements. The tradeoff relevant to the
profit maximizing PUBTV is the revenue gains from carrying more
advertisements versus the loss of market share due to carrying more
advertisements.
Not only does PUBTV carry sub-optimal levels of advertising, but PRITV
also charges sub-optimal levels of viewer subscription fees. Just as in the case of
PUBTV, this sub-optimality is due to the fact that PRITV does not have the
incentive to equate the marginal social benefit of advertisements with the marginal
social costs of advertising. The particular pattern of sub-optimality in viewer
subscription fees observed on the part of PRITV is over-charging when α δ < and
under charging when α δ > . This pattern of sub-optimality is similar to that
observed in the case of PUBTV, and is a consequence of the fact that there is
strategic complementarity in the choices of viewer subscription fee and advertising
space, as can be readily observed from the reaction functions of each broadcaster.
Therefore, when α δ > , PUBTV carries too little advertising space, and PRITV
under-charges viewer subscription fees, and vice versa. Given that the partial

122
mitigation of market imperfections cannot eliminate sub-optimal choices by both
broadcasters in equilibrium, there may still be scope for a welfare improvement
under ME.
It is interesting to compare the PE equilibrium in the sequential setting with
that in the simultaneous setting. Consider first the actions and payoffs of PRITV,
in the parameter space ( ) ( ) ( )
3 3
4 2
, , 1,
δ
α δ ∈ ∞ × . In this space, PRITV charges a
viewer subscription fee of
1
2
δ +
, has a market share of between
1
2
and
5
8
, and makes
profits of
( )
2
1
8
k
δ +
− under the sequential setting, whereas it charges a viewer
subscription fee of 1, has a market share of
1
2
, and makes profits of
1
2
k − under
the simultaneous setting. It is clear that PRITV does better in the sequential setting
than in the simultaneous setting under PE for the parameter space under
consideration.
Consider now the actions and payoffs of PRITV, in the parameter space
( ) ( ) ( )
3 3
4 2
, , ,
δ
α δ ∈ ∞ × ∞ . In this space, PRITV charges a viewer subscription fee of
5
4
, has a market share of
5
8
, and makes profits of
25
32
k − under the sequential
setting, whereas it charges a viewer subscription fee of 1, has a market share of
1
2
,
and makes profits of
1
2
k − under the simultaneous setting. It is clear that PRITV
does better in the sequential setting than in the simultaneous setting under PE for
the parameter space under consideration. Thus, the well known result of Gal-Or

123
(1985) regarding second mover advantage in the presence of strategic
complementarity extends to this particular two sided markets setting. We now turn
our attention to the equilibrium in the ME case.
The ME Case
In the ME market structure, PRITV is a privately owned profit maximizer,
while PUBTV is a state owned welfare maximizer. By the symmetry of the
locations of the two programs on the unit Hotelling interval, the sub-game
equilibria yield the same profit levels for PRITV and PUBTV as well as the same
level of social welfare regardless of the broadcasters’ specific program choices.
Recall that the SPNE involves content differentiation. Without loss of generality,
we characterize the sub-game equilibrium where PRITV produces program 1 and
PUBTV produces program 2.
Since program content is differentiated, we have
1 2
0 d d = = , and market
share is displayed below.

[ ]
1
2 1 2 1
2 1
1
1
b a s
b b
δ = + −
= −

Since PUBTV (the FTA broadcaster) moves first, it must calculate the
response of PRITV (the PTV broadcaster) to its choice of program type and
advertising space. Thus, PUBTV solves the optimization problem of PRITV for a

124
given choice of program type and advertising space. The optimization problem of
PRITV is displayed below.
( ) ( )
1 1
1 2 1 2 1
max max
s s
a b a s k π = −
The use of the parenthesis term indicates the fact that the optimization
problem is solved for a particular choice of advertising space and program type
(set equal to program 2 for ease of exposition) by PUBTV. Solving PRITV’s
optimization problem yields PRITV’s best response as ( )
2
1
2 1 2
a P
s a
δ +
= . Plugging
this back into the profit function of PUBTV gives the following optimization
problem for PUBTV.

[ ]
( ) ( )
[ ]
( ) ( ) ( ) ( )
2 2
1 2 1 2 1 2
0,1 0,1
max max 2
P co P ad P
a a
w s a v w s a w s a k
∈ ∈
= + + −
From the Kuhn-Tucker formulation, the equilibrium choices of viewer
subscription fee and ad space depend on the relationship between the advertiser
profit per advertisement α and the viewer nuisance per advertisement δ as
displayed in Table 3.9 and associated Figure 3.3. The term
0
α in the Figure 3.3 is
equal to
( ) 5 3
6 4
δ δ
δ
−
−
. Superscript M indicates ME equilibrium level. The socially
optimal levels of viewer subscription fee and ad space are respectively
( ) ( )
1 2
, 0,0
W W
s a = when α δ < , and
( ) ( )
1 2
, ,1
W W
s a α = when α δ > . As in the case of
the simultaneous setting, we restrict the viewer nuisance per advertisement to
assume values greater than 1. We restrict the advertiser profit per advertisement to

125
lie above
3
4
δ
, since this is necessary to satisfy the second order condition for
welfare maximizing choice of advertising space on PUBTV under ME market
structure. Broadcaster profits are displayed in Table 3.10. Consumer surplus and
social welfare are displayed in Table 3.11. Recall that advertisers make zero
profits since they are price takers in the advertising market.
Just as in the case of the sequential PE equilibrium, in the sequential ME
equilibrium also we have sub-optimal levels of viewer subscription fee and
advertising space. The pattern of sub-optimality is the same, with both
broadcasters over-charging (over-providing) viewer subscription fee (advertising
space) when α δ < , and under-charging (under-providing) viewer subscription fee
(advertising space) when α δ > . As usual, the reason for the sub-optimality is that
PRITV does not have the incentive to equate the marginal social benefit of
advertising with its marginal social cost. Although PUBTV has the incentive to do
this, the strategic complementarity between actions of PRITV and PUBTV means
that over / under charging of viewer subscription fee by PRITV leads to over /
under provision of ad space by PUBTV. We now compare broadcasters’ choices
of viewer subscription fee and advertising space under PE and ME.

126
Table 3.9: The FPSEQ ME Equilibrium Actions
Zone α δ
1
M
s
2
M
a
1
( )
( )
5 3
6 4
,
δ δ
δ
−
−
∞
( )
3
2
1,
1
2
δ +
1
2
( )
( )
5 3
5
6 6 4
,
δ δ
δ
δ
−
−

( )
3
2
1,
5 4
4 3
α δ
α δ
−
−

( )
6 5
4 3
α δ
δ α δ
−
−

2
( )
5
6
,
δ
∞ ( )
3
2
,∞
5 4
4 3
α δ
α δ
−
−

( )
6 5
4 3
α δ
δ α δ
−
−

3
( )
3 5
4 6
,
δ δ
( ) 1,∞
1
2
0

Table 3.10: Broadcaster Profits under FPSEQ ME
Zone
1
M
π
2
M
π
1
( )
2
1
8
k
δ +
−
( ) 3
4
k
α δ −
−
2
( )
( )
2
2
5 4
2 4 3
k
α δ
α δ
−
−
−
( )( )
( )
2
6 5 3 2
2 4 3
k
α α δ α δ
δ α δ
− −
−
−
3
1
8
k − k −

127
Figure 3.3: The FPSEQ ME equilibrium

Table 3.11: CS and Welfare under FPSEQ ME
Zone
M
CS
M
w
1
( )
1
48
13 3 14 v δ − + −  
 

( ) ( ) 3 10 3
7
48 4 16
2 v k
α δ δπ δ − −
− − + −
2
( )
2 2
2
277 432 168
12 4 3
v
α αδ δ
α δ
− +
−
− [ ]
27 2 1
3 48 4 3
2 v k
α
α δ δ −
− − + +
3
13
48
v−
7
48
2 v k − −

128
Comparing the PE and ME Equilibria
Proposition 3.5 Broadcasters’ choices of viewer subscription fee and
advertising space in the sequential setting are the same or more optimal under ME
than under PE.
Proof. Follows from the expressions for viewer subscription fee and
advertising space ▲
Contrary to the simultaneous setting, in the sequential setting we have more
optimal choices by both broadcasters for the entire admissible parameter space.
The second mover advantage of PRITV was not sufficient to push PUBTV’s
resolution of the tradeoff between welfare from advertising and welfare from
program content in the direction of sacrificing welfare from advertising (by
choosing less optimal ad space) in order to preserve welfare from program
content. Therefore, both broadcasters made more optimal choices under ME. We
now turn our attention to social welfare.
Proposition 3.6 Social welfare in the sequential setting is the same or higher
under ME than under PE if
5
6
δ
α > , else it is lower.
Proof. See Appendix ▲
We have already established that the viewer subscription fee and
advertising space decisions of both broadcasters are the same or more optimal
under ME than under PE. It follows that the welfare from advertising is the same

129
or higher under ME than under PE. Note that under full market coverage, the level
of viewer subscription fee does not impact welfare since it is merely a transfer
from the viewers to a broadcaster. Like in the case of the simultaneous setting, in
the sequential setting too we can realize a welfare improvement under ME only
under a restricted region of the parameter space.
For low levels of advertiser profit per advertisement i.e. when
5
6
δ
α < ,
regardless of the levels of viewer nuisance per advertisement, we have lower
social welfare under ME than under PE. In this situation, the overly aggressive
competition by the public broadcaster results in a complete shut down of
advertising space, forcing the private broadcaster to drastically lower its viewer
subscription fee. The result is a market share under ME that it skewed in favor of
the state broadcaster, thus increasing welfare losses from viewer transport costs to
a level greater than would be the case under PE, as well as causing the state
broadcaster to make losses. Even though welfare from more optimal advertising
choices is greater under ME, it is insufficient to outweigh the greater loss of
welfare from program content as well as the lower profits of both broadcasters.
The result is a reduction in welfare under ME.
In this situation, the complaints of private broadcasters of overly aggressive
competition from public broadcasters do merit some regulatory action, since not
only are both broadcasters worse off under ME, but social welfare is also reduced!

130
The optimal regulatory response however is not to privatize PUBTV but to
regulate (increase) ad space on PUBTV. While privatization may increase
PUBTV’s profits as well as increase social welfare, a greater increase in social
welfare may be had by the optimal regulation of ad space. We now turn our
attention to broadcaster profits under ME.
Proposition 3.7 PRITV’s profit in the sequential setting is the same or lower
under ME than under PE.
Proof. See Appendix ▲
Just as in the simultaneous setting, in the sequential setting also we have
PRITV making lower profits under ME than under PE. Thus, the two sided market
analog of the result of Sappington and Sidak (2003) continues to hold in the
sequential case as well, with aggressive competition by the state broadcaster
resulting in the same or lower profits for the private broadcaster under ME than
under PE. However, except in the region where
5
6
δ
α < , no regulatory action is
called for since lower profits for the private broadcaster is consistent with a
welfare improvement. In the region where
5
6
δ
α < , not only does the private
broadcaster make a lower profit under ME, but social welfare is also lower under
ME. The optimal regulatory response however is not to privatize PUBTV, but to
regulate (increase) ad space on PUBTV. While privatization may increase

131
PUBTV’s profits as well as increase social welfare, a greater increase in social
welfare may be had by the optimal regulation of ad space.
Proposition 3.8 PUBTV’s profit in the sequential setting is the same or lower
under ME than under PE.
Proof. See Appendix ▲
Just as in the simultaneous setting, in the sequential setting also we have
PUBTV making lower profits under ME than under PE. Thus, the welfare gains
under ME do not come for free in the sense that both broadcasters earn lower
profits under ME. In Zone 3 i.e. where
5
6
δ
α < , the state broadcaster does not carry
any advertisements, and consequently makes losses while carrying out its PSB
mandate. This is also the zone where welfare is lower under ME on account of
overly aggressive ad space reductions by PUBTV. Even though it would be
tempting to privatize PUBTV and therefore simultaneously solve the twin
problems of losses and lower social welfare, a greater welfare improvement can be
realized by regulating (increasing) ad space on PRITV, thus moving it from losses
to profits while still increasing welfare.
Summary of Results
We now summarize the main findings in this chapter, dealing with the
simultaneous setting first and the sequential setting next. In the simultaneous
setting, with regard to ad space decisions, we find that the partial mitigation of the

132
two market failures associated with the FTA FTA industry structure on account of
the use of the PTV broadcast technology by one of the broadcasters, was not
sufficient to restore optimality in broadcasters’ choices of viewer subscription fee
and advertising space in the PE market structure. We still observe the Anderson
and Coate (2005) type pattern of sub-optimal provision of ad space and sub-
optimal charging of viewer subscription fee by both broadcasters. The pattern of
sub-optimality observed is social over-provision (over-charging) of ad space
(viewer subscription fee) when α δ < , and vice versa. This is due to the fact that
neither broadcaster has the incentive to equate the marginal social benefit of
advertisements i.e. the gains from trade between advertisers and viewers, with the
marginal social costs of advertising, i.e. the disutility to viewers from watching
advertisements.
With regard to social welfare under PE, despite the fact that both market
imperfections are alleviated to a degree under FTA PTV industry structure as
compared to no alleviation of market imperfections under FTA FTA industry
structure, social welfare is lower under FTA PTV industry structure when α δ > ,
and vice versa. This is an interesting application of the general theory of second
best. As Lipsey and Lancaster (1957, pg.12) observe, “… it is not true that a
situation in which more, but not all, of the optimum conditions are fulfilled (e.g.

133
FTA PTV) is necessarily, or is even likely to be, superior to a situation in which
fewer are fulfilled (e.g. FTA FTA).”
Welfare is lower under FTA PTV than under FTA FTA for the PE
equilibrium in the simultaneous setting when α δ > because this is the situation
where optimality demands that every producer be allowed to advertise – since
programming is advertisement free under PTV broadcast technology, but both
channels carry advertisements under FTA FTA industry structure, it is no wonder
that welfare is lower under FTA PTV. This is a perverse case where partially
alleviating both market imperfections through the use of PTV broadcast
technology actually lowered welfare!
In the simultaneous setting of the FTA PTV industry structure under ME,
we find that there are multiple equilibria when ( )
2
2 3
,
δ δ
α∈ . Since neither
broadcaster agrees on the ranking of any of the equilibria, no equilibrium
dominates the others in terms of payoffs of both broadcasters. Consequently, it is
not possible to determine ex ante which equilibrium will manifest itself.
In all but one of the equilibria (equilibrium 1), we find that PUBTV’s
choice of advertising space is the same or more optimal under ME than under PE.
This is not surprising given that PUBTV is a social welfare maximizer, and thus
attempts to mitigate the social over-provision of ad space that obtains in these
equilibria by curtailing its own ad space. This forces PRITV to lower the viewer

134
subscription fee it charges, thus forcing it also to make the same or more optimal
choices of viewer subscription fee under ME than under PE.
In equilibrium 1 however, even though we have the problems of social
over-charging of viewer subscription fee and social over-provision of ad space,
both PRITV and PUBTV choose higher levels of viewer subscription fee and
advertising space than they would have under PE, thus worsening the sub-
optimality problem. This is surprising since we have the perverse outcome of the
welfare maximizing PUBTV choosing a less optimal level of advertising space
under ME than it would have if it was a privately owned entity operating under
PE. This may occur because of (1) the strategic complementarity between the
decisions of viewer subscription fee and advertising space, (2) the low level of
viewer disutility from advertising, and (3) the tradeoff between welfare from
advertising and welfare from program content.
At low levels of viewer disutility from advertising i.e. when ( ) 1,3 δ ∈ , a
choice of higher viewer subscription fee by PRITV will force PUBTV to carry a
larger advertising space. This is because PUBTV has to trade off the welfare from
program content (viewer transport costs of watching non preferred programs) with
the welfare from advertising (gains from trade from advertising). Choosing a low
level of advertising space will minimize the loss of welfare from advertising when
α δ < , but will result in large welfare loss from program content since PUBTV

135
will end up with a large market share, which will result in large viewer transport
costs. At low levels of viewer disutility from advertising, the tradeoff may be in
favor of PUBTV choosing a higher advertising space since welfare losses from
advertising are likely to be lower than welfare losses from program content. This
is how it is possible to have a perverse outcome where PUBTV chooses higher
advertising space under ME than it would under PE.
Note that at higher levels of viewer disutility from advertisements
( ) 3, δ ∈ ∞ , PUBTV would always resolve its tradeoff in favor of minimizing
welfare loss from advertising at the cost of increased welfare loss from program
content, so it would not be possible to have this kind of perverse outcome. In
summary, broadcasters’ choices are the same or more optimal under ME than
under PE, in the restricted parameter space ( ) ( ) ( )
2
, , 3,
δ
α δ ∈ ∞ × ∞ , while they may
or may not be more optimal in the parameter space ( ) ( ) ( )
2
2 3
, , 1,3
δ δ
α δ ∈ × .
We have already established that the viewer subscription fee and
advertising space decisions of both broadcasters are the same or more optimal
under ME than under PE in the simultaneous setting when ( ) ( ) ( )
2
, , 3,
δ
α δ ∈ ∞ × ∞ .
It follows that the welfare from advertising is the same or higher under ME than
under PE. Note that welfare from program content is the same or lower under ME
than it is under PE. In equilibrium under PE, we have
1 2
P P
s a δ = . We call this

136
choice of viewer subscription fee and advertising space “symmetric choice”.
Under symmetric choice, the viewer chooses between programs solely on the basis
of which program’s content is most preferred.
However, under ME, we may or may not have symmetric choice depending
on the parameter space. In regions where we do not have symmetric choice, some
viewers under ME may watch the program whose content is less preferred since
the utility loss from paying the viewer subscription fee is lower than the utility loss
from watching advertisements. In such regions, we have larger viewer transport
costs, and hence lower welfare from program content under ME.
Nevertheless, it can be shown that the welfare gains from the same or more
optimal decisions of viewer subscription fee and advertising space under ME
outweigh the welfare losses from program content whenever we have
( ) ( ) ( )
2
3
, , 3,
δ
α δ ∈ ∞ × ∞ . Since all other components of social welfare are the same
under ME and PE, it follows that welfare under ME is the same or greater than
welfare under PE whenever ( ) ( ) ( )
2
3
, , 3,
δ
α δ ∈ ∞ × ∞ .
In the case of the perverse equilibrium where ad space and viewer
subscription fee choices of both broadcasters are less optimal under ME, we have
lower welfare under ME than under PE. However, this equilibrium is easily
recognized by the fact that it is the only one where there is a corner solution of
100% advertising space on the state broadcaster. Although privatization of the

137
state broadcaster could improve welfare, an even greater level of welfare could be
realized by the optimal regulation of ad space on the state broadcaster.
With regard to broadcaster profit in the simultaneous setting, we find that in
the parameter space ( ) ( ) ( )
2
3
, , 3,
δ
α δ ∈ ∞ × ∞ , PRITV charges the same or lower
viewer subscription fee under ME than it does under PE. Moreover, PRITV has
the same or lower audience share under ME than it does under PE
7
. Since viewer
subscription fees are the only source of revenue of PRITV, the same or lower
viewer subscription fees in addition to the same or lower market share translates
into the same or lower profits for PRITV under ME. Thus, in this parameter
region, we have the actions of the state broadcaster being to the detriment of the
private broadcaster in the sense of allowing the private broadcaster only the same
or lower profits under ME than under PE. The non profit focus of the state
broadcaster renders it a more aggressive
8
competitor than another private
broadcaster.
Thus, the two sided market analog of the result in Sappington and Sidak
(2003) holds for the case of FTA PTV industry structure as well for the FTA FTA
industry structure in the simultaneous setting. Recall that the Sappington and
Sidak (2003, pg.184) result was “… a reduced focus on profit can provide state

7
PRITV market share under ME is either ½ or ¼ depending on the parameter values, while its market share
under PE is always ½

8
By more aggressive, we mean that the state broadcaster carries an advertising space that is less than or
equal to the advertising space that would be carried by a private broadcaster under the same conditions.

138
enterprises with stronger incentives (than profit maximizing firms) to pursue
activities that disadvantage competitors.” There is therefore some justice to the
complaints of private broadcasters of an uneven playing field and overly
aggressive competition from state broadcasters. However, no regulatory action is
called for since welfare is improved by the presence and operation of state
broadcasters in this parameter region i.e. ( ) ( ) ( )
2
3
, , 3,
δ
α δ ∈ ∞ × ∞ .
With regard to the profits of the state broadcaster, we find that in the
parameter region ( ) ( ) ( )
2
3
, , 3,
δ
α δ ∈ ∞ × ∞ , we have PUBTV carrying the same or
lower advertising space under ME than it does under PE. Moreover, PUBTV has
the same or lower audience share under ME than it does under PE
9
. Since
advertising revenue is the only source of revenue of PUBTV, the same or lower
advertising space in addition to the same or lower market share translates into the
same or lower profits for PUBTV under ME.
Just as in the case of the simultaneous setting in the FTA FTA industry
structure, the potential welfare gains under ME in the simultaneous setting of the
FTA PTV industry structure do not come for free. The increased aggression of the
state broadcaster in terms of advertising space reductions results in welfare
improving choices of viewer subscription fee and advertising space by both

9
PRITV market share under ME is either ½ or 0 depending on the parameter values, while its market share
under PE is always ½

139
broadcasters. However, this comes at the cost of the state broadcaster making
lower profits under ME than under PE. In the case where the state broadcaster
does not carry any advertisements at all (e.g. BBC), it would clearly require some
sort of budgetary support in order to perform its public service mandate. As long
as the deadweight losses of raising funds to support the state broadcaster are not
outweighed by the welfare gains from having a state broadcaster, a welfare
argument could be made in order to support the operations of loss making state
broadcasters.
As regards the consumer surplus of viewers in the simultaneous setting of
the FTA PTV industry structure, it is clear that it is the same or higher under ME
than it would be under PE (except in the case of the perverse equilibrium). This is
because the profits of both broadcasters are the same or lower under ME than
under PE; advertisers make zero profits under both ME and PE; and consumer
surplus, being the only other component of welfare, is the same or higher under
ME than under PE. Contrary to the case of the FTA FTA industry structure
however, it is not possible to have a “win-win-win” situation, since the cost of the
improvement in social welfare and consumer surplus is the lower profits of both
broadcasters.
We now turn our attention to the sequential setting of the FTA PTV
industry structure. Just as in the simultaneous setting, in the sequential setting too

140
we observe that both broadcasters make sub-optimal choices under PE, with over-
charging of viewer subscription fee and over-provision of ad space when α δ < ,
and vice versa. Here too, the partial mitigation of both market failures associated
with an FTA FTA industry structure is not sufficient to restore optimality of the
equilibrium.
When comparing the PE equilibrium in the sequential setting with that of
the simultaneous setting, we find that PRITV does better in terms of greater profits
in the sequential setting than in the simultaneous setting under PE when
( ) ( ) ( )
3
4
, , 1,
δ
α δ ∈ ∞ × ∞ . Thus, the well known result of Gal-Or (1985) regarding
second mover advantage in the presence of strategic complementarity extends to
this particular two sided markets setting.
When considering the broadcaster choices under the ME equilibrium of the
sequential setting, we find sub-optimal choices of viewer subscription fee and
advertising space by both broadcasters. However, the choices of both broadcasters
are the same or more optimal under ME than under PE. The welfare concerns of
the state broadcaster result in both broadcasters making the same or more optimal
viewer subscription fee and advertising space decisions under ME than under PE.
With regard to social welfare under the sequential setting, we find that it is
the same or higher under ME than under PE if
5
6
δ
α > , else it is lower. For the low
levels of advertiser profit per advertisement in Zone 3, i.e. for
5
6
δ
α < , regardless

141
of the levels of viewer nuisance per advertisement, we have lower social welfare
under ME than under PE. In this situation, the overly aggressive competition by
the public broadcaster results in a complete shut down of advertising space,
forcing the private broadcaster to drastically lower its viewer subscription fee. The
result is a market share under ME that it skewed in favor of the state broadcaster,
thus increasing welfare losses from viewer transport costs to a level greater than
would be the case under PE, as well as causing the state broadcaster to make
losses. Even though welfare from more optimal advertising choices is greater
under ME, it is insufficient to outweigh the greater loss of welfare from program
content as well as the lower profits of both broadcasters. The result is a reduction
in welfare under ME.
With regard to broadcaster profits, just as in the simultaneous setting, in the
sequential setting also we have both broadcasters making lower profits under ME
than under PE. Thus, the two sided market analog of the result of Sappington and
Sidak (2003) continues to hold in the sequential case as well, with aggressive
competition by the state broadcaster resulting in the same or lower profits for the
private broadcaster under ME than under PE. No regulatory action is called for
except in Zone 3, since lower profits for the private broadcaster is consistent with
a welfare improvement.

142
In Zone 3 of the ME equilibrium in the sequential setting i.e. when
5
6
δ
α < ,
the state broadcaster does not carry any advertisements, and consequently requires
some form of budgetary support in order to carry out its PSB mandate. In this
situation, the complaints of private broadcasters of overly aggressive competition
from public broadcasters do merit some regulatory action, since not only are both
broadcasters worse off under ME, but social welfare is also reduced! The optimal
regulatory response would be to regulate (increase) ad space on the state
broadcaster in order to simultaneously solve the twin problems of losses and lower
social welfare. Privatization in this parameter region would also solve the twin
problems, but would deliver a welfare improvement smaller than could be realized
by optimal regulation of ad space on the state broadcaster.
143

Chapter IV
Quality Choice by Agricultural Cooperatives
Introduction
In India, agricultural cooperatives may have monopoly status in their area
of operation (typically a number of villages). The Registrar of Cooperatives may
forbid the establishment of another cooperative in the same area if it is detrimental
to the cooperative movement (for example by interfering with the operation of
existing cooperatives). In practice, the exercise of this power of the Registrar of
Cooperatives has resulted in the overwhelming majority of agricultural co-ops in
India having local area monopoly status.
One purpose of according local monopoly status to agricultural
cooperatives was to rectify the market power imbalance caused by a large number
of small farmers facing a few large traders. While monopolies in general may
lower social welfare by restricting output, this may not true of cooperative
monopolies on account of legal requirements for co-ops to practice open
membership and / or to process or market all produce supplied to it by members.
Since this would remove one avenue for members of the cooperative to exploit the
market power of a monopoly, the existence of cooperatives would have to be
justified by other considerations such as scale economies, access to government
144

subsidies, or the ability to exploit market power while making decisions on
product quality, product range, or product grading / certification (if the output
quality is unobservable).
Given the numerous other avenues for monopoly co-ops to exploit their
market power, simply removing the ability of a monopoly co-op to practice output
restriction would be insufficient to preserve efficiency. In light of the findings of
the literature on the quality distorting effects of monopolies both when quality is
observable and when quality is unobservable, it is necessary to investigate the
welfare implications of permitting the unregulated operation of monopoly
cooperatives in quality differentiated industries. This chapter focuses on the
choices of monopoly agricultural co-ops regarding the provision of product quality
in a setting where quality is unobservable and grading capacity is insufficient to
grade all available output.
Regarding the issue of market structure and quality choice when quality is
observable, the quality distortions under monopoly have been well known since
Spence (1975) for the case where the monopolist chooses a single level of quality,
and Mussa & Rosen (1978) for the case where the monopolist chooses a range of
qualities. When the quality and quantity setting monopolist chooses a range of
qualities, it is known that quality is under-supplied when there is full market
coverage, but is optimally supplied when there is partial market coverage – in the
145

latter case, the monopolist under-supplies quantity (Berges-Sennou, Bontems and
Requillart, 2003).
With regard to market structure and quality choice when the quality of
individual units of output is unobservable prior to purchase, Hollander, Monier-
Dilhan and Raynal
1
(1999) consider an oligopoly situation where producers
choose both the average quality of output (which is observed by the consumer),
and whether or not to grade their output, i.e. reveal the quality of individual units
of output. The cost of grading is a fixed cost, and there is sufficient grading
capacity to grade the entire output. The amount of output produced by each firm is
given exogenously – however, the firm can choose the proportion of output that is
of high quality (and hence can choose the average quality of output). Thus, HMR
impose a technological restriction on the firm's quantity quality tradeoff. This
restriction captures the fact that it is not possible to have perfect control over
quality and quantity in an agricultural setting.
Surprisingly, they find that an increase in the cost of grading has an
ambiguous effect on the average quality of output that undergoes grading. If the
cost of grading increases, the amount of ungraded output increases, while the
amount of graded output decreases. One would expect that only producers of
output that is on average of higher quality would have the incentive to go for

1
Referred to as HMR hereafter.
146

costly grading, since they could get higher prices for their high quality output from
quality conscious consumers. So, the average quality of output sent for grading
should increase.
However, since the total quantity of output is given exogenously, market
clearing requires some consumers to switch away from graded output to ungraded
output. If the number of switchers from high quality output to ungraded output
exceeds the number of switchers from low quality to ungraded output, then the
average quality of output sent for grading will fall, since the most lucrative market
for graded output (consumers who buy high quality output) has shrunk. So, an
increase in grading cost can actually lower the average quality of output sent for
grading. Further, they find that the average industry quality increases with the cost
of grading if the increased cost of grading reduces the quality of output that
undergoes grading.
Marette and Crespi
2
(2003) consider a model where there are two types of
producers, one producing high quality output and the other producing low quality
output, and demonstrate welfare improvements under stable producer cartels. In
their model, quality is unobservable to the consumer but known to the producer.
The producer can signal his quality with perfect accuracy by engaging in costly
third party certification (the analog of the grading decision in HMR). Producers

2
Referred to as MC hereafter.
147

may collude in order to share certification costs and also reduce competition by
lowering output. There is sufficient certification capacity to certify the entire
industry output if so desired. Under certain conditions, MC find that a stable cartel
that provides information about product quality through certification may improve
overall social welfare even though competition is reduced.
The present chapter focuses on market structure and quality choice by
contrasting the polar cases of perfect competition and monopoly co-ops. Producers
are homogeneous, and output can be of two qualities – low and high. We treat
quantity of output as exogenously given, say as a result of decisions on the
quantum of land set aside for a particular crop that were made in the past. We
allow the producer to choose the distribution of the fixed amount of output across
low and high quality. This is equivalent to choosing the average quality of output,
and can be done for example by varying cultivation practices post the land
allocation decision. Thus, we impose a technological restriction on the farmer's
tradeoff between quality and quantity that is in the spirit of HMR, with the
difference that we consider the homogeneous producer case.
The more usual treatment of quality, such as that in MC, does not impose
any technological restriction on the tradeoff between quantity and quality. The
HMR treatment is more representative of an agricultural context where it is
difficult to perfectly control the quantity and quality of individual units of output,
148

while the MC treatment is more representative of a manufacturing context, where
it is possible to achieve perfect control the quantity quality of individual units of
output.
Contrary to HMR and MC however, we consider the situation where the
there are exogenous capacity constraints in grading. This could arise in situations
where grading infrastructure is either too expensive for widespread adoption, or is
newly introduced and therefore not in widespread use. Capacity constraints in
grading are commonplace in most agricultural markets in most developing
countries. In the presence of capacity constraints, as long as grading is profitable
for farmers, all output up to the grading capacity constraint is graded, with the rest
of the output sold as ungraded. Therefore, information is revealed about the
quality of individual units of output only for a fraction of total output, with no
information on quality being revealed for the rest of the output.
The fraction of output about which no information on quality is revealed is
treated as being of known average quality. Although average quality is a choice
variable of the producer, the consumer observes this choice prior to making his
consumption decision. This would be the case if the cultivation practices (but not
output) of the farmer could be costlessly and accurately certified by some agency,
or if the consumer could infer information about average quality from the quantity
of output in each quality grade, or from the prices of output in each quality grade.
149

A real world example of capacity constraints in grading arises in the quality
determination process in the Indian milk industry, which originally consisted of
use of eyeball and tongue (to visually inspect and then taste the milk), but
graduated to the use of a lactometer
3
. The following anecdote from the milk sector
illustrates the linkage between market structure, grading capacity and quality.
Prior to the advent of the AMUL
4
movement in India, milk was sold by private
milkmen, who were notorious for the low quality of their milk. As the joke went,
in an effort to adulterate the milk, milkmen did not add water to milk, but rather
added milk to water! This state of affairs existed because there was no reliable
method for the quality of individual units of milk to be made known prior to
purchase, and hence no incentive for either milk farmers or milkmen to provide
high quality milk. Consumers got low quality milk and milk farmers got low
prices for their milk from private milkmen.
With the advent of the AMUL movement, payment to milk farmers began
to be made based on the specific gravity of milk, as measured by the lactometer.
With the increase in grading capacity that the slow diffusion of the lactometer
represented, milk farmers finally had the incentive to supply higher quality milk to
the dairy and hence to the consumer. The use of lactometers to adjudge the quality

3
A device that measures the specific gravity of milk, and hence the quantum of adulteration with water.

4
The AMUL movement was begun by Dr. Varghese Kurien in Anand, Gujarat, and its principles were
spread across the entire country as part of Operation Flood.
150

of milk became an industry standard due to the pioneering work of the AMUL
movement leaders.
Prior to 1992, dairies (such as those of AMUL) had local area monopoly
status in their area of operation. During the registration process, dairies had to
declare an area known as a “milk shed”, and were given sole rights to procure milk
from their milk sheds. However, the government removed restrictions on new milk
processing capacity in 1992, and allowed competition between dairies even in the
same milk shed. Thus, at one stroke, the nature of the dairy industry changed from
being a monopoly, to being a competitive industry.
Once the capacity constraint in grading had been alleviated to the point
where the quality of each individual unit of milk could be determined with near
certainty through the use of the lactometer, there could be no welfare grounds for
the continued existence of monopoly dairies
5
. Although at the time the
government did not adduce welfare reasons for stripping dairies of local monopoly
status, its action could very well have led to welfare gains. What is interesting to
investigate is if the existence of monopoly co-op dairies during the period where
milk grading infrastructure was constrained because of the lack of widespread use
of the lactometer was detrimental to welfare, or if as in the case of MC's welfare
enhancing cartels, there was some way that these monopoly dairies could have

5
As would be argued by the usual finding in the literature on market structure and quality choice, that
monopolies distort quality when quality is observable
151

actually increased welfare. This is the starting point of analysis for the present
chapter.
Like all other agricultural cooperatives, the cooperative dairies which
dominated the Indian dairy scene prior to deregulation were subject to legal
restrictions on their ability to control the quantum of output of their members. The
first cooperative principle of the International Cooperative Alliance is “voluntary
and open membership”. This principle states that “Co-operatives are voluntary
organizations, open to all persons able to use their services and willing to accept
the responsibilities of membership, without gender, social, racial, political or
religious discrimination” (International Cooperative Alliance, 2009). This
principle was adopted in Indian Cooperative Law, through legal provisions to the
effect that any primary producer in the geographical jurisdiction of the co-op
would be eligible to become a member of the co-op. This removed much of the
freedom of the cooperative to indulge in quantity distortions by restricting
participation in the co-op. As a result, the co-op had to take the quantity of output
supplied by members as given. This exogeneity of output quantity is in the spirit
of HMR. MC however allow the producer cartel to use two routes to distort
quantity: (1) exclude producers from membership of the cartel, and (2) impose
quantity limits on producers who join the cartel.
152

In our setting, the co-op can influence the quality choice of their members
by instituting an appropriate internal price quality schedule. By varying the slope
of the price quality schedule, the co-op can induce a given level of quality from
members. We do not however model the internal price quality schedule of the co-
op, but assume instead that the co-op determines the profit maximizing quality
level, and induces that from members through the use of an appropriate price
quality schedule.
We find that for given capacity constraint in the grading technology, if the
cost of quality is high enough, or if the quality gap between low and high quality
output is low enough, the co-op member farmer may provide a greater amount of
high quality output at a lower price than the equivalent perfectly competitive
farmer. Similarly, for a given cost of quality and a given quality gap between low
and high quality output, if the grading capacity is low enough, the co-op member
farmer may provide a greater amount of high quality output at a lower price than
the equivalent perfectly competitive farmer. In both cases, welfare under the
monopoly co-op exceeds that under perfect competition.
This finding is in the spirit of MC, who find welfare gains from stable
producer cartels under certain conditions. MC’s results require that there be no
binding constraint on grading capacity which would prevent producers from
realizing their grading intentions. In contrast, we impose binding exogenous
153

constraints on grading capacity, and find welfare gains under monopoly only if the
grading capacity is low enough!
The intuition behind our result is as follows. When there is no grading
capacity constraint i.e. quality of each unit of output can be determined with
certainty, the monopoly market structure provides sub-optimally low levels of
average quality, while the competitive market structure provides optimal levels of
average quality. As the grading capacity reduces, farmers in both market structures
lose their incentive to make costly effort to produce high quality output, and hence
reduce the average quality of their output.
However, the quality choice of the co-op member farmer is less sensitive to
the reduction in grading capacity than that of the competitive farmer. This is
because the co-op member farmer, acting through the co-op, has two tools at his
disposal to handle the weakening incentives to provide quality – (1) he can reduce
average quality, and / or (2) he can distort prices. As a result, his reliance on the
first tool i.e. average quality reduction is less than that of the competitive farmer.
Therefore, for a sufficiently low grading capacity, the quality choice of the co-op
member farmer will be higher than that of the competitive farmer.
Recall that when there are grading capacity constraints, farmers in both
market structures provide sub-optimally low levels of average quality. Since
average quality is the only choice variable, welfare will be higher under the market
154

structure where the extent of quality under-provision is less. Since the extent of
quality under-provision is lower under monopoly when grading capacity is low
enough, welfare would be higher under monopoly than under perfect competition!
This provides a nice welfare justification for the government to protect the
monopoly status of dairies when grading capacity is low i.e. no widespread use of
lactometers, but strip them of their monopoly status after the use of the lactometer
became the industry standard!
In summary, this chapter contributes to the literature on market structure
and quality choice by considering the welfare implications of the equilibrium
choice of quality by a monopoly cooperative that produces a quality differentiated
good of unobservable quality in the presence of grading capacity constraints. The
chapter is organized as follows. The next section lays out the model and analyzes
the equilibrium and optimal choices under monopoly and competitive market
structures, and the final section concludes.
The Model and Results
We adapt the setup used in HMR to the case of homogeneous producers.
The agricultural commodity has two quality levels: high and low. The quality of
individual units of output is not observable by the consumer, but the average
quality of the output is observed by consumers. We assume that the number of
producers is exogenously given; and that all producers are identical. We compare
155

the choices made by perfectly competitive farmers and co-op member farmers
regarding the proportion of high quality output, or equivalently, the average
quality of output.
There are n producers who must exert costly effort in order to increase the
proportion of their output that is of high quality. Setting the proportion of high
quality output to λ requires the producer to incur cost cλ . All producers together
produce a unit mass of output.
When considering the quality choice decision, we assume that the grading
technology is fully efficient. Full efficiency means that the quality of graded
output is identified correctly as being of either low or high quality. Grading is
treated as being costless to the farmer, and we verify that it is individually rational
for each farmer to want to have his output graded. However, due to grading
capacity constraints, only a fraction ξ of total output can be graded. The fraction
of output that could not be graded is treated by consumers as being of known
average quality. The consumer can either observe the producers’ choice of
proportion of high quality output, or infer it from the quantity / price of output of
each quality grade.
Consumers have preferences of the Mussa Rosen type, and decide whether
to consume the high quality output at price
H
P , or the ungraded output at price
M
P ,
or nothing at all. Consumers do not consume the low quality output. Low quality
156

output can be sold for use in industrial applications at a constant price 1
L
P = ,
regardless of the quantity of such output. Consumers are indexed by a taste
parameter θ , that is uniformly distributed in [0,1].
Consumers are risk neutral, and are unaware of the quality of individual
units of output in the absence of grading. However, consumers observe / infer the
average quality of ungraded output. The consumer of type θ gets expected utility
( ) , U s s P θ θ = − from consuming a unit of output of quality s at price P. Denote
the quality of the high quality output as
H
s . Since a fraction λ of output is of high
quality, the average quality of the ungraded output is
H
s s λ = . The consumer of
type θ chooses her purchase decision (ungraded output, high output, or nothing)
in order to maximize her surplus, given below.
( ) { } max , ,0
H H M M
S s P s P θ θ θ = − −
As Hollander, Monier-Dilhan and Ossard (1999) observe, in order for both
categories of output (high and ungraded) to exist in equilibrium, it must be the
case that
H M
P P > . The choice of prices determines the allocation of consumers
across quality grades average and high as given below.

[ ]
[ ]
[ ]
0
0 1
1
0, buys nothing
, buys ungraded output
,1 buys high quality output
θ θ
θ θ θ
θ θ
∈
∈
∈

157

The prices which sustain this allocation are given below.

( )
0
1
M
H M H
P s
P P s s
θ
θ
=
= + −

The distance between partition points for uniform distribution of consumer
types is given below.

( )
0
1 0
1
0 1
1
1
L
M
H
Q
Q
Q
θ ξ λ
θ θ ξ
θ ξλ
= − = −
= − = −
= − =

The right hand side of the above equations represents the total number of
units of each grade of output (low, average and high respectively). Rearranging the
equations yields expressions for the partition points as given below.

( )
0
1
1
1
θ ξ λ
θ ξλ
= −
= −

The expressions for the prices of the ungraded and high quality output are
then
6
:

( )
( )
1
1
M H
H H
P s
P s
ξλ λ
λ
= −
= −

6
Recall that the price of the low quality output is set equal to 1.
158

Assuming the existence of n producers, the profit level of each producer is
given below
7
.

[ ]
1
n L L M M H H
Q P Q P Q P c π λ = + + −
We now analyze the farmers’ decisions of proportion of high quality
output.
Perfect competition
The perfectly competitive farmer takes the prices of the various grades of
output as given when making his decision on the proportion of high quality output.
Profit maximization implies equating the marginal revenue with the marginal cost
of a unit increase in the proportion of high quality output. Assuming an interior
solution, the FOC for profit maximizing choice of λ is given below.
( ) 1
H
P nc ξ − =
The equilibrium proportion of high quality output is given below.

*
1
H
nc
s C
ξ
ξ
λ
+
= −
In order to have an interior solution for
*
C
λ , it must be the case that
H
nc
s
ξ
ξ
+
< . In order for the farmer to realize a greater profit by having his output
graded, as opposed to not grading it at all, it must be the case that
H
nc
s
ξ
ξ
+
< . Thus,

7
Note that each producer must incur a cost cλ in order to set the proportion of his high quality output to
λ .
159

as long as the interior condition is satisfied, the farmer prefers to have his output
graded. Comparative statics for
*
C
λ are given below.

*
*
2
C
H
C
H
d
n
dc s
d
nc
d
s
λ
ξ
λ
ξ
ξ
= −
=

Note that the equilibrium proportion of high quality output decreases when
the marginal cost of quality increases. The equilibrium proportion of high quality
output increases when the fraction of output graded increases. The greater the
fraction of output graded, the stronger the incentive to produce high quality output
since more high quality output will be rewarded as high quality output, and not
sold as ungraded output.
Monopoly
The monopoly coop member farmer takes into account the effect of the
choice of proportion of high quality output on the prices of the various grades of
output. The effects of quality choice on the prices of the various grades of output
are given below.

( ) 1 2
M
H
dP
d H
dP
d H
s
s
λ
λ
ξ λ = −
= −

The price of ungraded output increases as the proportion of high quality
increases, as long as
1
2
λ < . We will show that in equilibrium, the co-op member
160

farmer never sets the proportion of high quality output greater than ½. The price of
high quality output decreases as the proportion of high quality output increases.
This is because the ungraded output becomes relatively more attractive to the
consumer as λ increases, since it would contain on average a higher proportion of
high quality output. Thus, the price of the high quality output must be lowered in
order to induce the consumer to purchase high quality output.
Profit maximization implies equating the marginal revenue with the
marginal cost of a unit increase in the proportion of high quality output. Assuming
an interior solution, the FOC for profit maximizing choice of λ is given below.
( ) ( ) 1 1
H M
P P
H
P nc
λ λ
ξ ξλ ξ
∂ ∂
∂ ∂
− + + − =
The equilibrium proportion of high quality output is given below.

( )
*
1
2 2
1
H
nc
M s
ξ
ξ ξ
λ
+
−
 
= −
 

In order to have an interior solution for
*
M
λ , it must be the case that
( ) 2
H
nc
s
ξ
ξ ξ
+
< − . Note that satisfaction of the condition for interior solution for
*
C
λ
implies satisfaction of the condition for interior solution for
*
M
λ . As long we have
an interior solution, the monopolist will never set the proportion of high quality
output greater than ½. Having their output graded is individually rational for
farmers in the sense of delivering a greater profit level over not grading the output
161

as long as
( ) ( )
( )
2
1 2
4 2
0
H
H
nc s
ns
ξ ξ
ξ ξ
  + − −
 
−
> , which is always true. Comparative statics for
*
M
λ are
given below.

( )
( )
( )
*
2 *
2 2
2 2
2 1
2 2
M
H
M
H
d
n
dc s
nc d
d
s
λ
ξ ξ
ξ ξ λ
ξ
ξ ξ
−
− −
−
= −
=

Note that the equilibrium proportion of high quality output decreases when
the marginal cost of quality increases. It can be shown that the equilibrium
proportion of high quality output increases (decreases) when the proportion of
output graded increases if ( ) ( ) 2 nc nc nc ξ < > + − . The term ( ) 2 nc nc nc + −
ranges between 0 and 1 as nc ranges from 0 to ∞ . Note that the monopoly co-op
always has incentive to restrict the quantity of high quality output in order to keep
its price high (call this the output restriction effect). At the same time, the co-op
has incentive to produce more high quality output as the amount of output graded
increases, since more high quality output would be rewarded as high quality
output and not sold as ungraded output (call this the grading effect). The grading
effect dominates as long as the fraction of output graded is not too large, and thus
*
M
λ increases in ξ when ξ is not too large. Once the fraction of output graded
crosses a certain level, the output restriction effect dominates, and thus
*
M
λ
decreases in ξ when ξ is large enough.
162

Industry Structure and Equilibrium Choice of λ
Proposition 4.1 The co-op member farmer sets a higher (lower) proportion of
high quality output than the perfectly competitive farmer i.e. ( )
* *
M C
λ λ > < if
( )
( ) 2
3 2
H
nc
s
ξ ξ ξ
ξ
− +
−
> <
Proof. Follows from the expressions for
*
M
λ and
*
C
λ ▲
The right hand side of the above expression is convex increasing in ξ , and
lies in [0,1]. Proposition 4.1 can be explained by the fact that the co-op farmer
takes into account the effect of the choice of λ on the prices of ungraded output
and high quality output, while the competitive farmer does not. For the co-op
farmer, increasing λ increases the price of ungraded output, while it decreases the
price of the high quality output. Recall that the co-op farmer’s FOC for profit
maximization is as given below.
( ) 1
H M
P P
H H M
P Q Q nc
λ λ
ξ
∂ ∂
∂ ∂
− + + =
This is the same as the FOC of the competitive farmer, with the exception
of the second and third terms in the LHS. For low levels of ξ , there is a large
amount of ungraded output (large
M
Q ) and a small amount of high quality output
(small
H
Q ). For given nc and
H
s , if ξ is low enough, the co-op farmer will
choose a larger level of λ than the competitive farmer since the gains from the
price increase of the ungraded output will more than offset the losses from the
163

price decrease of the high quality output. Not only is there a larger amount of high
quality output under monopoly, but the price of high quality output is also lower.
For high levels of ξ , there is a small amount of ungraded output and a large
amount of high quality output. For given nc and
H
s , if ξ is high enough, the co-op
farmer will choose a smaller level of λ than the competitive farmer since the
gains from the price increase of the ungraded output will not be able to offset the
losses from the price decrease of the high quality output. Thus, there is not only a
lower amount of high quality output under monopoly, but its price is also higher.
Proposition 4.2 The co-op member farmer’s choice of proportion of high
quality output is less sensitive to the fraction of output graded than that of the
competitive farmer.
Proof. Follows from the expressions for
*
M
d
d
λ
ξ
and
*
C
d
d
λ
ξ
▲
The intuition behind the proposition is as follows. When ξ increases, there
is better information about quality, and hence farmers under both market structures
have incentives to incur more (costly) effort and increase the proportion of high
quality output. However, the co-op member farmer has two tools at his disposal to
reap the gains of an increase in grading capacity – (1) he can increase the
proportion of high quality output (but incur additional cost), and (2) he can distort
prices (at no additional cost). As a result, the increase in proportion of high quality
output in response to an increase in grading capacity by the co-op member farmer
164

will be lower than that by the equivalent competitive farmer, since the co-op
farmer can always substitute costly effort with costless price distortion through the
cooperative. This fact is key to the demonstration of welfare gains under
monopoly co-ops in the next sub-section.
Welfare Analysis
Social welfare in this economy consists of the sum of farmer profits and
consumer surplus, as follows.

1
0 1
1
M M L L
W s d s d P Q nc
θ
θ θ
θ θ θ θ λ = + + −
∫ ∫

The socially optimal proportion of high quality output is given below.

( )
( )
2
1
2 2
1
H
H
nc s
W s
ξ
ξ ξ
λ
+ −
−
 
= −
 

In order to have an interior solution for
W
λ it must be the case that
( )
1
2
1 2
H
nc
s
ξ
ξ ξ
+
<  + − 
 
. Note that satisfaction of the interior condition for the
perfectly competitive case implies satisfaction of the interior condition for the
socially optimal choice of proportion of high quality output.
Note that
*
M W
λ λ < i.e. the monopoly co-op farmer provides a sub-optimally
low proportion of high quality output whenever we have 1
H
nc
s
ξ +
< . This condition is
satisfied whenever the interior condition for the choice of
*
M
λ is satisfied. Further,
*
C W
λ λ < i.e. the competitive farmer provides a sub-optimally low proportion of
165

high quality output whenever we have ( ) ( ) ( ) 1 2 1 0
H
nc s ξ ξ ξ −  + + −  >
 
, which is
true whenever 1 ξ < . Thus, neither the perfectly competitive farmer nor the co-op
member farmer chooses the first best proportion of high quality output in the
presence of grading constraints i.e. both farmers provide a sub-optimally low
proportion of high quality output. Of course, when 1 ξ = , the perfectly competitive
farmer provides the optimal proportion of high quality output. This is because both
market imperfections (insufficient grading capacity and imperfect competition) are
absent when 1 ξ = . The relative levels of under-provision of high quality output
under each market structure are given below.

( )
( )
2 * *
3 2
2 * *
3 2
when 0
when 0
H
H
nc
s C M W
nc
s M C W
ξ ξ ξ
ξ
ξ ξ ξ
ξ
λ λ λ ξ
λ λ λ ξ
− +
−
− +
−
< < < < <
< < < < <

Whenever
* *
M C
λ λ > , the extent of under-provision of high quality output is
greater under the perfect competition market structure, and vice versa. We now
turn our attention to social welfare under perfect competition and monopoly.
Proposition 4.3 Social welfare under monopoly exceeds social welfare under
perfect competition whenever
* *
M C
λ λ >
Proof. See the Appendix ▲
Given that farmers in both market structures under-provide quality in the
presence of grading constraints, welfare is higher under the market structure in
166

which the extent of quality under-provision is less. The extent of under-provision
of high quality output is lower under monopoly whenever nc is high enough and /
or ξ is low enough. Figure 4.1 displays the parameter region in which social
welfare is higher under the monopoly market structure.
Figure 4.1: Welfare Comparison

When 1 ξ = i.e. when there is no grading capacity constraint, we have
optimal provision of high quality output under perfect competition but under-
provision of high quality output under monopoly. As ξ decreases, farmers in both
market structures respond by decreasing the proportion of high quality output.
However, the choice of quality by the competitive farmer is more sensitive to ξ
167

than that by the co-op farmer. Hence, the competitive farmer reduces quality by a
greater amount in response to a given decrease in grading capacity. Therefore,
beyond a point, further decreases in ξ will lead to the extent of quality under-
provision becoming greater under perfect competition than under monopoly. This
will occur when ξ decreases sufficiently that we have
( ) 2
3 2
H
nc
s
ξ ξ ξ
ξ
− +
−
> . When this
occurs, welfare under monopoly will exceed that under perfect competition.
Our result on welfare gains under imperfect competition is similar in spirit
to that of MC, who find that the rent gains from cartelization permit the cartel to
engage in costly quality certification. Under certain conditions, the increased
information on quality can be welfare enhancing even though the cartel can
collude to reduce competition. The MC result however depends on the availability
of sufficient quality certification infrastructure to satisfy the cartel’s needs. In
contrast, we demonstrate welfare gains under imperfect competition (monopoly)
specifically in those conditions in which there is a large shortfall in grading
infrastructure.
Our result provides yet another interesting example of the general theory of
second best. In the presence of one market imperfection i.e. insufficiency of
grading capacity (the situation when 1 ξ < ), introducing yet another market
imperfection (by granting the co-op local area monopoly status) can actually
deliver a welfare increase when the grading capacity is low enough!
168

Conclusion
This chapter focuses on the choices of monopoly co-ops regarding the
provision of product quality in a setting where quality is unobservable and grading
capacity is insufficient to grade all available output. We find that for given
capacity constraint in the grading technology, if the cost of quality is high enough,
or if the quality gap between low and high quality output is low enough, the co-op
member farmer may provide a greater amount of high quality output at a lower
price than the equivalent perfectly competitive farmer. Similarly, for a given cost
of quality and a given quality gap between low and high quality output, if the
grading capacity is low enough, the co-op member farmer may provide a greater
amount of high quality output at a lower price than the equivalent perfectly
competitive farmer. In both cases, welfare under the monopoly co-op exceeds that
under perfect competition.
The intuition behind our result is as follows. When there is no grading
capacity constraint i.e. quality of each unit of output can be determined with
certainty, the monopoly market structure provides sub-optimally low levels of
average quality, while the competitive market structure provides optimal levels of
average quality. As the grading capacity reduces, farmers in both market structures
lose their incentive to make costly effort to produce high quality output, and hence
reduce the average quality of their output.
169

However, the quality choice of the co-op member farmer is less sensitive to
the reduction in grading capacity than that of the competitive farmer. This is
because the co-op member farmer, acting through the co-op, has two tools at his
disposal to handle the weakening incentives to provide quality – (1) he can reduce
average quality, and / or (2) he can distort prices. As a result, his reliance on the
first tool i.e. average quality reduction is less than that of the competitive farmer.
Therefore, for a sufficiently low grading capacity, the quality choice of the co-op
member farmer will be higher than that of the competitive farmer.
Recall that when there is a grading capacity constraint, farmers in both
market structures provide sub-optimally low levels of average quality. Since
average quality is the only choice variable, welfare will be higher under the market
structure where the extent of average quality under-provision is less. Since the
extent of average quality under-provision is lower under monopoly when grading
capacity is low enough, welfare would be higher under monopoly than under
perfect competition! This provides a nice welfare justification for the government
to protect the monopoly status of dairies when grading capacity is low i.e. no
widespread use of lactometers, but strip them of their monopoly status after the
use of the lactometer became the industry standard! In fact, our result provides yet
another interesting example of the general theory of second best. In the presence
of one market imperfection i.e. insufficiency of grading capacity (the situation
170

when 1 ξ < ), introducing yet another market imperfection (by granting the co-op
local area monopoly status) can actually deliver a welfare increase when the
grading capacity is low enough!
In summary, this chapter contributes to the literature on market structure
and quality choice by considering the welfare implications of the equilibrium
choices of average quality of output by a monopoly cooperative that produces a
quality differentiated good of unobservable quality in the presence of grading
capacity constraints.
171

Chapter V
Conclusion
This dissertation examines the economics of non-profit institutions in the
television industry and agriculture. We analyze two kinds of non-profit
institutions: (1) state owned television broadcasters, and (2) agricultural
cooperatives. For state owned television broadcasters, we focus on competition
with privately owned television broadcasters over advertising levels, viewer
subscription fees and program content. For agricultural cooperatives, we focus on
the quality provision decision when the cooperative has local area monopoly status
in a situation where quality of individual units of output is unobservable and there
are capacity constraints on output grading infrastructure. We first conclude the
discussion of the television industry, and then the discussion of agricultural
cooperatives.
The Television Industry
Recent papers on program content and advertising intensity in the television
broadcast industry, with the notable exceptions of Kind et. al. (2007) and Pan
(2009) have modeled competition between television broadcasters in the context
of a purely private economy industry structure, where each broadcaster’s sole
objective is to maximize profits. Data on the world TV industry however
172

demonstrates that the industry is better modeled as a mixed economy industry
structure, where state owned broadcasters with possibly non-profit objectives
compete with profit maximizing privately owned broadcasters.
Heap (2005) finds that public service broadcasting objectives are similar
the world over, and do not emphasize profit maximization as an objective,
although Norris et. al. (2003) do find instances where dedicated public service
broadcasters seem to be more interested in increasing audience shares than in
fulfilling their PSB mandates. PSB mandates alone however do not provide and
economic rationale for having a state broadcaster, since they seem to be more
concerned with issues of culture, national identity, citizenship etc.
The economic rationale for having a state broadcaster i.e. public provision
of television broadcast services can only come from an attempt to rectify some or
the other market imperfections in the television broadcast industry, which cannot
be corrected more economically through alternate means. As it happens, one of the
two main broadcasting technologies in use i.e. the FTA broadcasting technology
does create two market imperfections, and therefore offers the possibility of
welfare enhancing public provision of television broadcast services. The two
market imperfections are (1) the public good nature of FTA broadcasts, and (2) the
fact that advertising, while necessary to realize gains from trade between
173

advertisers and viewers, imposes a negative externality in the form of a utility
penalty on viewers, who would rather not watch advertisements at all.
The PTV broadcast technology partially mitigates one of these
imperfections since it provides broadcasters with the ability to charge viewers for
the telecast, and thus solves the public good problem of FTA broadcasts. Since not
all broadcasters use the PTV broadcast technology, it is still possible to find
markets where all broadcasters use the FTA broadcast technology, as well as
markets where some broadcasters use the FTA broadcast technology while some
broadcasters use the PTV broadcast technology.
It is far from clear that the partial mitigation of the twin market
imperfections of the FTA broadcast technology, through the use of the PTV
broadcast technology by some broadcasters in conjunction with the use of the FTA
broadcast technology by other broadcasters, will deliver welfare improvements. In
fact, the general theory of second best, as laid out in Lipsey and Lancaster (1957)
is quite clear on the fact that there is no a priori reason to expect a welfare
improvement from partial mitigation of market imperfections. Thus, there may
still be an economic rationale for the existence and operation of state broadcasters.
The mixed economy literature in one sided market
1
settings has long
recognized the fact that when state owned enterprises with objectives that differ

1
The one sided market setting here refers to markets where buyers and sellers interact directly i.e. without
intermediaries who may charge for the privilege of bringing buyers and sellers together.
174

from profit maximization compete with profit maximizing private enterprises, they
may have incentives to undertake “anti-competitive” actions that hurt their private
enterprise counterparts. Further, even if the objective of the state owned enterprise
is welfare maximization, its actions may reduce welfare in equilibrium
(Sappington and Sidak, 2003). It is important therefore to see whether such
concerns hold good in two sided market settings such as the television broadcast
industry.
Complaints against “anti-competitive” practices of state broadcasters have
been leveled not only by private broadcasters, but also by other state broadcasters
operating in the same market! Recent regulatory actions in France have resulted in
a ban on advertising in the state broadcaster, while loss making state broadcasters
the world over are facing pressure to improve their “bottom lines” or be privatized.
Privatization pressures may also be driven not just by profit and loss
considerations, but also by a belief in the efficiency of markets. There is therefore
no practical consensus on the need for state broadcasters, and if they are needed,
whether they are to be left alone or regulated in their actions.
Since it is theoretically unclear whether the presence of state broadcasters
can be justified in terms of delivering welfare improvements by internalizing the
twin market imperfections, and whether the cost of their doing so would be losses
for them and lower profits for their private sector counterparts, there is a need for a
175

comprehensive analysis of the issue in a variety of settings that obtain in the real
world. This is exactly what this dissertation attempts to do.
The main questions this dissertation asks are:
1. Do broadcasters make more optimal ad space and viewer subscription
fee decisions under ME i.e. do prices do a better job of internalizing the
externalities inherent in the FTA FTA and FTA PTV market structures
when one broadcaster is state owned, and has welfare maximization as
its objective?
2. Is social welfare enhanced by the presence and operation of state
broadcasters?
3. Do private broadcasters make lower profits under ME due to the “anti-
competitive” actions of state broadcasters? If so, what regulatory action
is called for?
We answer these questions under various settings of industry structure
(FTA FTA and FTA PTV), market structure (PE and ME), and entry timing
(simultaneous and sequential). This dissertation therefore extends the literature on
program content and advertising intensity in the television broadcast industry by
analyzing the mostly ignored mixed economy setting, where a welfare maximizing
state broadcaster competes against a profit maximizing private broadcaster. We
provide theoretical results identifying conditions under which the public provision
176

of TV broadcasting services results in a welfare improvement, identify the effects
of such provision on competing privately owned broadcasters, and thereby inform
the debate in regulatory circles regarding the appropriate disciplines on state
broadcasters as well as the appropriate treatment of loss making state broadcasters.
The first question this dissertation asks is whether prices do a better job of
internalizing the twin market imperfections of the FTA broadcast technology when
one broadcaster is state owned, and has welfare maximization as its objective. We
find that the answer is “yes, but not always”. Under FFSIM, it is always true that
both broadcasters make more optimal ad space choices under ME, while it is true
only under certain conditions on the marginal social benefits and costs of
advertising under FFSEQ. Under FPSIM, both broadcasters make more optimal ad
space and viewer subscription fee choices under ME only under certain conditions
on the marginal social benefits and costs of advertising, while it is always the case
that broadcasters’ decisions are more optimal under ME for FPSEQ.
In the models under consideration in this dissertation, we permit only two
content choices: maximal content differentiation, and perfect content duplication.
By an analogous argument to that in Peitz and Valletti (2008), perfect content
duplication would lead to a “race to the bottom” in terms of viewer subscription
fees and advertising levels, and thus zero viewer subscription and advertising
revenue under all industry structures (FTA FTA and FTA PTV), all entry
177

situations (simultaneous and sequential), and all market structures (PE and ME).
Further, it would lead to a reduction in social welfare on account of lost
opportunities for gains from trade between advertisers and viewers, as well as
increased viewer transport costs. Therefore, both broadcasters in all situations
prefer content differentiation to content duplication in equilibrium.
The second question this dissertation asks is if social welfare is enhanced
by the operation of state broadcasters. We find that the answer is “yes, but not
always”. In FFSIM, we find that welfare under ME is always higher than total
welfare under PE. For FFSEQ, we find that welfare is higher under ME except in a
small parameter range characterized by high marginal social benefit of advertising
and low marginal social cost of advertising. For FPSIM, we find that in equilibria
where the choice of ad space is the same or more optimal under ME i.e. when the
marginal social cost of advertising is high, we have welfare improvements under
ME, and vice versa. For FPSEQ, we find that welfare is the same or higher under
ME whenever the marginal social benefit of advertising is high enough relative to
the marginal social cost, and vice versa.
Despite the fact that social welfare is not always higher under ME, the
situations where social welfare is lower under ME are easily identified, since these
are the situation where there is a corner solution for the choice of ad space by the
178

state broadcaster i.e. 100% ad space or 0% ad space
2
. The appropriate response is
then the regulation of ad space on the state broadcaster, not the privatization of the
state broadcaster. This is because privatization of the state broadcaster may give a
welfare improvement that is smaller than could be realized by regulation of ad
space on the state broadcaster.
When comparing the FTA FTA and the FTA PTV equilibria for PE in the
simultaneous setting, we find that even though both market imperfections are
alleviated to a degree under FTA PTV, social welfare is lower under FTA PTV
than under FTA FTA when the marginal social benefit of advertising exceeds the
marginal social cost, and vice versa. This is an interesting application of the
general theory of second best. As Lipsey and Lancaster (1957, pg.12) observe, “…
it is not true that a situation in which more, but not all, of the optimum conditions
are fulfilled (e.g. FTA PTV) is necessarily, or is even likely to be, superior to a
situation in which fewer are fulfilled (e.g. FTA FTA).”
The third question this dissertation asks is whether the private broadcaster
makes lower profits under ME. This question is motivated by complaints of
private broadcasters against “anti-competitive” practices of state broadcasters. We
find that the answer is “yes, but not always”. For FFSIM, we find that the private
broadcaster actually makes greater profits under ME when the marginal social

2
There are state broadcasters with 0% ad space e.g. BBC, but no state broadcasters with 100% ad space.

179

benefit of advertising exceeds the marginal social cost! For FFSEQ, the private
broadcaster always makes lower profits under ME when the marginal social
benefit of advertising is less than the marginal social cost, but may make lower or
higher profits when the converse is true. For FPSIM and FPSEQ, the private
broadcaster always makes lower profits under ME.
Therefore, depending on the marginal social benefits and costs of
advertising, the state broadcaster takes actions that are either to the benefit of, or to
the detriment of the private broadcaster (where benefit means profit and detriment
means loss). These results are the two sided market analog of the result in
Sappington and Sidak (2003, pg.184), who find that a “reduced focus on profit can
provide state enterprises with stronger incentives to pursue activities that
disadvantage competitors.”
Given that the complaints from private broadcasters about making lower
profits because of “anti-competitive” practices of state broadcasters have some
merit in certain situations, the question is whether any regulatory action is called
for to discipline state broadcasters. Under the FTA FTA industry structure, every
situation where private broadcasters make lower profits under ME than under PE
is consistent with a welfare improvement under ME. Therefore, no regulatory
action is called for.
180

Under FPSIM, lower profits for the private broadcaster is accompanied by
lower social welfare only when the state broadcaster carries 100% ad space i.e.
telecasts only advertisements and no programming! Since this situation rarely
arises in practice, and even if it does arise is easily detectable by the simple
expedient of watching the telecast for any evidence of non-advertising
programming, the solution would be simply to regulate advertising on the state
broadcaster rather than its privatization. Such regulation could potentially increase
social welfare to a level greater than possible under PE, but would not result in a
level of private broadcaster profit greater than possible under PE.
Under FPSEQ, when the marginal social benefit of advertising is low
enough compared to the marginal social cost, not only does the private broadcaster
make lower profit under ME, but social welfare is also lower under ME. However,
this situation occurs only when the state broadcaster carries 0% ad space i.e.
telecasts only programming and no advertisements! Since this situation is easily
detectable by the simple expedient of watching the telecast for any evidence of
advertising, the solution would be simply to regulate advertising on the state
broadcaster instead of its privatization. Such regulation could potentially increase
social welfare to a level greater than possible under PE, but would not result in a
level of private broadcaster profit greater than possible under PE.
181

In the models considered in the present dissertation, depending on the
relationship between the marginal social benefit and the marginal social cost of
advertising, the state broadcaster may incur losses while attempting to carry out its
PSB mandate, and usually makes lower profit that it would have if it was a
privately owned profit maximizing entity. In such situations, if the deadweight
loss incurred in raising public funds to provide budgetary support to the state
broadcaster is low enough, a welfare argument can be made for the continued
existence and operation of state broadcasters, even if they are loss making.
Although privatization of the state broadcaster may increase its profits, it may well
lead to a reduction in social welfare, or deliver a welfare increase that is smaller
than could be achieved by regulation of ad space on the state broadcaster.
The analysis of the television industry can be extended along several
directions. Firstly, it would be fruitful to repeat the analysis for alternate objective
functions of the state broadcaster. For example, one could envisage a convex
combination of own profits and viewer surplus instead of social welfare. Secondly,
one could allow for alternate product market structures such as quality
differentiated products as opposed to homogenous or horizontally differentiated
products. Thirdly, one could allow for broadcasters using the PTV broadcast
technology to carry advertisements. Interestingly, the Italian government recently
attempted to legislate caps on advertising in PTV broadcasts, an action with
182

uncertain welfare consequences in a mixed economy setting. Fourthly, one could
allow for more content choices as opposed to the polar choices of content
differentiation versus content duplication. Fifthly, one could extend the analysis to
issues of program quality, as opposed to issues of program content.
Agricultural Cooperatives
In India, the overwhelming majority of agricultural co-ops have local area
monopoly status in their area of operation. One purpose of according local
monopoly status to agricultural cooperatives was to rectify the market power
imbalance caused by a large number of small farmers facing a few large traders.
While monopolies in general may lower social welfare by restricting output, this
may not true of cooperative monopolies on account of legal requirements for co-
ops to practice open membership and / or to process or market all produce
supplied to it by members. Since this would remove one avenue for members to
exploit the market power of a monopoly, the existence of cooperatives would have
to be justified by other considerations such as scale economies, access to
government subsidies, or the ability to exploit market power while making
decisions on product quality, product range, or product grading / certification (if
the output quality is unobservable).
Given the numerous other avenues for monopoly co-ops to exploit their
market power, simply removing the ability of a monopoly co-op to practice output
183

restriction would be insufficient to preserve efficiency. In light of the findings of
the literature on the quality distorting effects of monopolies both when quality is
observable and when quality is unobservable, it is necessary to investigate the
welfare implications of permitting the unregulated operation of monopoly
cooperatives in quality differentiated industries. Motivated by an example from
the Indian milk industry, this dissertation focuses on the choices of monopoly co-
ops regarding the provision of product quality in a setting where quality is
unobservable and grading capacity is insufficient to grade all available output.
We find that for given capacity constraint in the grading technology, if the
cost of quality is high enough, or if the quality gap between low and high quality
output is low enough, the co-op member farmer may provide a greater amount of
high quality output at a lower price than the equivalent perfectly competitive
farmer. Similarly, for a given cost of quality and a given quality gap between low
and high quality output, if the grading capacity is low enough, the co-op member
farmer may provide a greater amount of high quality output at a lower price than
the equivalent perfectly competitive farmer. In both cases, welfare under the
monopoly co-op exceeds that under perfect competition.
The intuition behind our result is as follows. When there is no grading
capacity constraint i.e. quality of each unit of output can be determined with
certainty, the monopoly market structure provides sub-optimally low levels of
184

average quality, while the competitive market structure provides optimal levels of
average quality. As the grading capacity reduces, farmers in both market structures
lose their incentive to make costly effort to produce high quality output, and hence
reduce the average quality of their output.
However, the quality choice of the co-op member farmer is less sensitive to
the reduction in grading capacity than that of the competitive farmer. This is
because the co-op member farmer, acting through the co-op, has two tools at his
disposal to handle the weakening incentives to provide quality – (1) he can reduce
average quality, and / or (2) he can distort prices. As a result, his reliance on the
first tool i.e. average quality reduction is less than that of the competitive farmer.
Therefore, for a sufficiently low grading capacity, the quality choice of the co-op
member farmer will be higher than that of the competitive farmer.
Recall that when there is a grading capacity constraint, farmers in both
market structures provide sub-optimally low levels of average quality. Since
average quality is the only choice variable, welfare will be higher under the market
structure where the extent of average quality under-provision is less. Since the
extent of average quality under-provision is lower under monopoly when grading
capacity is low enough, welfare would be higher under monopoly than under
perfect competition! Our result provides yet another interesting example of the
general theory of second best. In the presence of one market imperfection i.e.
185

insufficiency of grading capacity (the situation when 1 ξ < ), introducing yet
another market imperfection (by granting the co-op local area monopoly status)
can actually deliver a welfare increase when the grading capacity is low enough!
In summary, our analysis contributes to the literature on market structure
and quality choice by considering the welfare implications of the equilibrium
choices of average quality of output by a monopoly cooperative that produces a
quality differentiated good of unobservable quality in the presence of grading
capacity constraints.
The analysis can be extended in several directions. We assumed that all
eligible farmers in the area of operation of the cooperative became members of the
cooperative. In reality, the membership of most cooperatives rarely exceeds 90%
of the eligible farmers in the area of operation. It would be interesting to
endogenize the decision to join the cooperative and then see how the existence of a
competitive fringe of farmers would alter the quality setting behavior of the
cooperative. Further, our analysis assumes that the membership of the cooperative
is homogeneous. One of the main sources of tension between members in
agricultural cooperatives is the fact that different members have differing costs of
provision of quality. If we allow for heterogeneity of the eligible farmers, this
raises questions of who will have incentive to join the cooperative, and how the
186

interests of these heterogeneous members can be reconciled through the internal
price quality schedule.
Finally, we have assumed that there are two qualities of graded output i.e.
high and low, as well as ungraded output which is a mix of high and low output in
a fixed proportion. It would be interesting to allow for the admission of more
grades of output through the mixing of high and low quality output in different
proportions at a fixed cost per grade creation. This would allow us to analyze the
issue of the equilibrium range of grades of output, which is another dimension
along which monopolies are known to distort.
187

References
Anderson, S.P., & Coate, S. (2005). Market provision of broadcasting: A welfare
analysis. The Review of Economic Studies, 72(4), 947-972. Retrieved from
http://www.jstor.org/stable/3700696
British Broadcasting Corporation. (2005). Review of the BBC's Royal Charter.
London, UK: Author. Retrieved from
http://www.bbc.co.uk/aboutthebbc/policies/pdf/green_paper_response.pdf
Bergès-Sennou, F., Bontems, P., & Réquillart, V. (2003). When monopoly
oversupplies quality (IDEI Working Paper). Toulouse, France: Institut D’
Economie Industrielle. Retrieved June 5, 2009, from
http://idei.fr/doc/wp/2003/monopoly.pdf
Block, P., Houseley, W., Nicholls, T., & Southwell, R. (2001). Managing in the
media. Boston: Focal Press.
Channel 4. (2008). Channel 4 Television Corporation Report and Financial
Statements 2008. Author. Retrieved from
http://www.channel4.com/about4/pdf/2008/CH4_Report2008_Full.pdf
Djankov, S., McLiesh, C., Nenova, T., & Shleifer, A. (2003). Who owns the
media? The Journal of Law and Economics, 46(2), 341-382.
http://dx.doi.org/10.1086/377116
European Commission. (2007). State Aid Decision on Public Service Broadcasting
in Germany. Brussels: Author. Retrieved from
http://ec.europa.eu/competition/state_aid/register/ii/doc/E-3-2005-WLWL-en-
24.04.2007.pdf
Eurovision Audiovisual Observatory. (2009). MAVISE database of TV companies
and TV channels in the European union and candidate countries. Retrieved
June 7, 2009, from http://mavise.obs.coe.int/
Gabszewicz, J. J., Laussel, D., & Sonnac, N. (2004). Programming and advertising
competition in the broadcasting industry. Journal of Economics &
Management Strategy, 13(4), 657-669. doi:10.1111/j.1430-9134.2004.00027.x
188

Galo-Or, E. (1985). First Mover and Second Mover Advantages. International
Economic Review, 26(3), 649-653. Retrieved from JSTOR at
http://www.jstor.org/pss/2526710
Gal-Or, E., & Dukes, A. (2003). Minimum Differentiation in Commercial Media
Markets. Journal of Economics & Management Strategy, 12(3), 291-325. doi:
10.1111/j.1430-9134.2003.00291.x
Heap, S. P. H. (2005). Television in a digital age: What role for public service
broadcasting?. Economic Policy, 20(41), 112-157. doi:10.1111/j.1468-
0327.2005.00134.x
Hollander, A., Monier-Dilhan, S., & Raynal, H., (2001), Grading and quality
upgrading: complements or substitutes? (Economics Working Paper Archive).
Toulouse, France: French Institute for Agronomy Research (INRA),
Economics Laboratory in Toulouse (ESR Toulouse). Retrieved July, 10, 2009,
from http://econpapers.repec.org/RePEc:rea:inrawp:14.
Hollander, A., Monier-Dilhan, S., & Ossard, H. (1999). Pleasures of cockaigne:
Quality gaps, market structure, and the amount of grading. American Journal
of Agricultural Economics, 81(3), 501-511. doi:10.2307/1244010
International Cooperative Alliance. (2009). Retrieved June, 21, 2009, from
http://www.ica.coop/coop/index.html
Kind, H. J., Nilssen, T., & Sørgard, L. (2007). Competition for viewers and
advertisers in a TV oligopoly. Journal of Media Economics, 20(3), 211.
Retrieved from http://www.informaworld.com/10.1080/08997760701290708
Lipsey, R. G., & Lancaster, K. (1956). The general theory of second best. The
Review of Economic Studies, 24(1), 11-32. Retrieved from JSTOR at
http://www.jstor.org/pss/2296233
Marette, S., & Crespi, J. M. (2003). Can quality certification lead to stable cartels?
Review of Industrial Organization, 23(1), 43-64.
doi:10.1023/B:REIO.0000005595.35570.1a
Mussa, M., & Rosen, S. (1978). Monopoly and product quality. Journal of
Economic Theory, 18(2), 301-317. doi:10.1016/0022-0531(78)90085-6
189

Nasser, S., Muller, E., & Assael, H. (2007). The effect of competition in television
broadcasting on the ratio of advertising to programming time. Retrieved from
New York University Stern School of Business Website:
http://homepages.nyu.edu/~szn200/Nasser,%20Muller%20and%20Assael%20
%282007%29.pdf
Norris, P., Pauling, B., Zanker, R., & Lealand, G. (2003). The Future of Public
Broadcasting: The Experience in Six Countries. Wellington, New Zealand:
New Zealand on Air. Retrieved July 10, 2009, from
http://www.nzonair.govt.nz/media/6956/pb%20exp%20in%206%20countries
%20nov%2003.pdf
Pan, H. (2009). Content and advertising: TV media competition in a mixed
duopoly market. The Journal of Economic Asymmetries, 6(2), 137-154.
Retrieved from http://www.apforum.org/JEAvol6-Sept-2009.pdf
Peitz, M., & Valletti, T. M. (2008). Content and advertising in the media: Pay-tv
versus free-to-air. International Journal of Industrial Organization, 26(4),
949-965. doi:10.1016/j.ijindorg.2007.08.003
Sappington, D. E. M., & Sidak, J. G. (2003). Incentives for anticompetitive
behavior by public enterprises. Review of Industrial Organization, 22(3), 183-
206. doi:10.1023/A:1023607223501
Shaw, C. (1999). Deciding what we watch: Taste, decency, and media ethics in the
UK and the USA. Oxford University Press, USA.
Shaw, V., & Wake, D. (2009, January 21). Ofcom pushes channel 4 merger. The
Independent. Retrieved July 10, 2009, from
http://www.independent.co.uk/news/media/tv-radio/ofcom-pushes-channel-4-
merger-1452447.html
Spence, A. M. (1975). Monopoly, quality, and regulation. The Bell Journal of
Economics, 6(2), 417-429. Retrieved from JSTOR at
http://www.jstor.org/pss/3003237

190

Appendix
Proofs of Propositions in Chapter II
Proposition 2.2 In the simultaneous setting, social welfare under ME is
always greater than social welfare under PE.
Proof.
We will split the parameter space into regions and prove the proposition for each
region of the parameter space. Consider first the region ( ) ( ) ( )
3 1
2 4
, , 1,2 α δ α ∈ × .
Evaluating the welfare expressions under PE and ME in this region, we have
welfare under ME exceeding welfare under PE if
13
14
δ
α < . This is true in this
region.
Consider next the region ( ) ( ) ( )
3 4
4 3
, ,1 1,
α
α δ ∈ × . Evaluating the welfare
expressions under PE and ME in this region, we have welfare under ME exceeding
welfare under PE if
( )
( )
2
5
4 2
0
α δ
δ α δ
−
−
> . This is true in this region.
Consider next the region ( ) ( ) ( )
3 4
4 3
, ,1 ,2
α
α δ α ∈ × . Evaluating the welfare
expressions under PE and ME in this region, we have welfare under ME exceeding
welfare under PE if
13
14
δ
α < . This is true in this region.
191

Consider next the region ( ) ( ) ( )
4
3
, 1, 1,
α
α δ ∈ ∞ × . Evaluating the welfare
expressions under PE and ME in this region, we have welfare under ME exceeding
welfare under PE if
( )
( )
2
5
4 2
0
α δ
δ α δ
−
−
> . This is true in this region.
Consider next the region ( ) ( ) ( )
4
3
, 1, ,2
α
α δ α ∈ ∞ × . Evaluating the welfare
expressions under PE and ME in this region, we have welfare under ME exceeding
welfare under PE if
13
14
δ
α < . This is true in this region.
Combining results for both regions, we have the desired result that social
welfare is always higher under ME than under PE. ▲

Proposition 2.3 In the simultaneous setting, PRITV’s profit is higher (lower)
under ME than under PE when ( ) δ α < > .
Proof.
We will split the parameter space into regions and prove the proposition for each
region of the parameter space. Consider first the region ( ) ( ) ( )
3 1
2 4
, , 1,2 α δ α ∈ × .
Evaluating PRITV’s profit expressions under PE and ME in this region, we have
profit under PE exceeding profit under ME if
3
0
α
δ
− < . This is true in this region.
192

Consider next the region ( ) ( ) ( )
3 4
4 3
, ,1 1,
α
α δ ∈ × . Evaluating the profit
expressions under PE and ME in this region, we have profit under PE exceeding
profit under ME if
( )( )
( )
2
5 3
2 2
0
α α δ α δ
δ α δ
− −
−
< i.e. if
( )
5
3
,
α
δ α ∈ . This is true in this region.
Consider next the region ( ) ( ) ( )
3 4
4 3
, ,1 ,2
α
α δ α ∈ × . Evaluating the profit
expressions under PE and ME in this region, we have profit under PE exceeding
profit under ME if
3
8
0
α
δ
− < . This is true in this region.
Consider next the region ( ) ( ) ( ) , 1, 1, α δ α ∈ ∞ × . Evaluating the profit
expressions under PE and ME in this region, we have profit under ME exceeding
profit under PE if
( )( )
( )
2
5 3
2 2
0
α α δ α δ
δ α δ
− −
−
> , i.e. if δ α < . This is true in this region.
Consider next the region ( ) ( ) ( )
4
3
, 1, ,
α
α δ α ∈ ∞ × . Evaluating the profit
expressions under PE and ME in this region, we have profit under PE exceeding
profit under ME if
( )( )
( )
2
5 3
2 2
0
α α δ α δ
δ α δ
− −
−
< , i.e. if
( )
5
3
,
α
δ α ∈ . This is true in this region.
Consider next the region ( ) ( ) ( )
4
3
, 1, ,2
α
α δ α ∈ ∞ × . Evaluating the profit
expressions under PE and ME in this region, we have profit under PE exceeding
profit under ME if
3
8
0
α
δ
− < . This is true in this region.
Combining results for both regions, we have the desired result regarding
PRITV’s profits. ▲

193

Proposition 2.4 In the simultaneous setting, PUBTV’s profit is higher (lower)
under ME than under PE when ( ) δ α < > .
Proof.
We will split the parameter space into regions and prove the proposition for each
region of the parameter space. Consider first the region ( ) ( ) ( )
3 1
2 4
, , 1,2 α δ α ∈ × .
Evaluating PUBTV’s profit expressions under PE and ME in this region, we have
profit under PE exceeding profit under ME if
2
0
α
δ
− < . This is true in this region.
Consider next the region ( ) ( ) ( )
3 4
4 3
, ,1 1,
α
α δ ∈ × . Evaluating the profit
expressions under PE and ME in this region, we have profit under PE exceeding
profit under ME if
( )
( )
2
2 2
0
α α δ
α δ
−
−
< i.e. if δ α > . This is true in this region.
Consider next the region ( ) ( ) ( )
3 4
4 3
, ,1 ,2
α
α δ α ∈ × . Evaluating the profit
expressions under PE and ME in this region, we have profit under PE exceeding
profit under ME if
2
0
α
δ
− < . This is true in this region.
Consider next the region ( ) ( ) ( ) , 1, 1, α δ α ∈ ∞ × . Evaluating the profit
expressions under PE and ME in this region, we have profit under ME exceeding
profit under PE if
( )
( )
2
2 2
0
α α δ
α δ
−
−
> , i.e. if δ α < . This is true in this region.
194

Consider next the region ( ) ( ) ( )
4
3
, 1, ,
α
α δ α ∈ ∞ × . Evaluating the profit
expressions under PE and ME in this region, we have profit under PE exceeding
profit under ME if
( )
( )
2
2 2
0
α α δ
α δ
−
−
< , i.e. if δ α > . This is true in this region.
Consider next the region ( ) ( ) ( )
4
3
, 1, ,2
α
α δ α ∈ ∞ × . Evaluating the profit
expressions under PE and ME in this region, we have profit under PE exceeding
profit under ME if
2
0
α
δ
− < . This is true in this region.
Combining results for both regions, we have the desired result regarding
PUBTV’s profits. ▲

Proposition 2.6 Both broadcasters’ make lower profits under ME when
α δ < .
Proof.
We will split the parameter space into regions and prove the proposition for each
region of the parameter space. Consider first the region ( ) ( ) ( )
4 3
5 2
, , 1,
δ
α δ δ ∈ × .
PRITV’s profit is lower under ME than under PE since
( )( )
( )
2
1 2
5 4
4
δ α δ
α δ
+ −
−
 
>
 
. PUBTV’s
profit is lower under ME than under PE since
( )( )
( )
2
2 8 7 2
2
3 0
α δ α δ
δ α δ
δ
− −
−
− − < .
195

Consider next the region ( ) ( ) ( )
4 3
5 2
, , ,
δ
α δ δ ∈ × ∞ . PRITV’s profit is lower
under ME than under PE since ( )( ) 10 11 10 7 0 α δ α δ − − < . PUBTV’s profit is
lower under ME than under PE since
( )
( )
2
2
10 11
16 2
0
α α δ
δ α δ
−
−
− < .
Consider next the region ( ) ( ) ( )
4 3
2 5 2
, , 1,
δ δ
α δ ∈ × . PRITV’s profit is lower
under ME than under PE since
( ) 2
8
0
α δ +
− < . PUBTV’s profit is lower under ME
than under PE since
( ) 3
4
0
α δ−
< .
Consider next the region ( ) ( ) ( )
4 3
2 5 2
, , ,
δ δ
α δ ∈ × ∞ . PRITV’s profit is lower
under ME than under PE since
21
32
0
α
δ
− < . PUBTV’s profit is lower under ME than
under PE since
16
0
α
δ
− < .
Combining results for both regions, we have the desired result regarding
broadcaster profits. ▲

Proposition 2.7 Social welfare in the sequential setting is equal or higher
under ME than under PE except when ( ) ( ) ( )
3
2
, 2 , ,3 α δ δ ∈ ∞ × .
Proof.
We will split the parameter space into regions and prove the proposition for each
region of the parameter space. Consider first the region ( ) ( ) ( )
3
2
, 2 , 1, α δ δ ∈ ∞ × .
196

Evaluating social welfare expressions under PE and ME in this region, we observe
that social welfare under ME equals social welfare under PE.
Consider next the region ( ) ( ) ( )
3
2
, 2 , ,3 α δ δ ∈ ∞ × . Evaluating social welfare
expressions under PE and ME in this region, we observe that social welfare under
ME is lower than social welfare under PE if
( )
( )
25 2
26 4
δ δ
δ
α
−
−
> , which is true in this
region.
Consider next the region ( ) ( ) ( )
4 3
5 2
, ,2 1,
δ
α δ δ ∈ × . Evaluating social welfare
expressions under PE and ME in this region, we observe that social welfare under
ME is higher than social welfare under PE since
( ) ( )
( )
2
2 4 7
16 2
0
α δ δ δ
δ α δ
− − −  
 
−
> .
Consider next the region ( ) ( ) ( )
4 3
5 2
, ,2 ,
δ
α δ δ ∈ × ∞ . Evaluating social
welfare expressions under PE and ME in this region, we observe that social
welfare under ME is higher than social welfare under PE since
( )
( )
2
10 11
64 2
0
α δ
δ α δ
−
−
> .
Consider next the region ( ) ( ) ( )
4 3
2 5 2
, , 1,
δ δ
α δ ∈ × . Evaluating social welfare
expressions under PE and ME in this region, we observe that social welfare under
ME is higher than social welfare under PE since
( )
( )
14
2 8
δ δ
δ
α
−
−
> in this region.
Consider next the region ( ) ( ) ( )
4 3
2 5 2
, , ,
δ δ
α δ ∈ × ∞ . Evaluating social welfare
expressions under PE and ME in this region, we observe that social welfare under
ME is higher than social welfare under PE since
25
26
δ
α < in this region.
197

Combining results for both regions, we have the desired result regarding
social welfare. ▲

Proofs of Propositions in Chapter III
Proposition 3.2 Social welfare is the same or higher under ME than under PE
when ( ) ( ) ( )
2
, , 3,
δ
α δ ∈ ∞ × ∞ .
Proof.
We will split the parameter space into regions and prove the proposition for each
region of the parameter space. Consider first the region ( ) ( ) ( )
2
2 3
, , 3,
δ δ
α δ ∈ × ∞ . In
this region, the unique PE equilibrium involves social welfare
7
12 2
2
P
w v k
α
δ
= − − + . In the same region, there are two ME equilibria: one
equilibrium is identical to the PE equilibrium and thus has identical social welfare;
and the other equilibrium involves social welfare
7
48
2
M
w v k = − − . The latter ME
equilibrium is associated with higher social welfare than the PE equilibrium if
7
8
α δ < , which is true in this region.
Consider next the region ( ) ( ) ( )
2
3
, , 3,
δ
α δ ∈ ∞ × ∞ . In this region, the unique
PE equilibrium involves social welfare
7
12 2
2
P
w v k
α
δ
= − − + . In the same region,
the unique ME equilibrium is identical to the PE equilibrium, and thus has
198

identical social welfare. Combining results for both regions, we have the desired
result that social welfare is the same or higher under ME than under PE. ▲

Proposition 3.3 PRITV profit is the same or lower under ME than under PE
when ( ) ( ) ( )
2
, , 3,
δ
α δ ∈ ∞ × ∞ .
Proof.
We will split the parameter space into regions and prove the proposition for each
region of the parameter space. Consider first the region ( ) ( ) ( )
2
2 3
, , 3,
δ δ
α δ ∈ × ∞ . In
this region, the unique PE equilibrium involves PRITV making a profit
1
2 1
P
k π = − . In the same region, there are two ME equilibria: one equilibrium is
identical to the PE equilibrium and therefore PRITV makes identical profit; the
other equilibrium has PRITV making a profit
1
8 1
M
k π = − . Clearly, the profit of
PRITV in this equilibrium is lower than that under PE.
Now consider the region ( ) ( ) ( )
2
3
, , 3,
δ
α δ ∈ ∞ × ∞ . In this region, the unique
ME equilibrium is identical to the unique PE equilibrium, and hence PRITV
makes the same profit under ME and PE. Combining results for both regions, we
have the desired result that PRITV’s profit is the same or lower under ME than
under PE. ▲

199

Proposition 3.4 PUBTV profit is the same or lower under ME than under PE
when ( ) ( ) ( )
2
, , 3,
δ
α δ ∈ ∞ × ∞ .
Proof.
We will split the parameter space into regions and prove the proposition for each
region of the parameter space. Consider first the region ( ) ( ) ( )
2
2 3
, , 3,
δ δ
α δ ∈ × ∞ . In
this region, the unique PE equilibrium involves PUBTV making a profit
2 2
P
k
α
δ
π = − . In the same region, there are two ME equilibria: one equilibrium is
identical to the PE equilibrium and therefore PUBTV makes identical profit; the
other equilibrium has PUBTV making a profit
2
M
k π =− . Clearly, the profit of
PUBTV in this equilibrium is lower than that under PE.
Now consider the region ( ) ( ) ( )
2
3
, , 3,
δ
α δ ∈ ∞ × ∞ . In this region, the unique
ME equilibrium is identical to the unique PE equilibrium, and hence PUBTV
makes the same profit under ME and PE. Combining results for both regions, we
have the desired result that PUBTV’s profit is the same or lower under ME than
under PE. ▲

Proposition 3.6 Social welfare in the sequential setting is the same or higher
under ME than under PE if
5
6
δ
α > , else it is lower.
200

Proof.
We will split the parameter space into regions and prove the proposition for each
region of the parameter space. Consider first the region ( ) ( ) ( )
3 5 3
4 6 2
, , 1,
δ δ
α δ ∈ × . In
this region, the unique PE equilibrium is associated with a welfare of
( ) ( ) 3 10 3
7
48 4 16
2
P
w v k
α δ δ δ − −
= − − + − . In the same region, the unique ME equilibrium is
associated with a welfare of
7
48
2
M
w v k = − − . Welfare is higher under ME if
( )
( )
10 3
4 3
δ δ
δ
α
−
−
< , which is not true in the given region.
Now consider the region ( )
( )
( )
( )
5 3
5 3
6 6 4 2
, , 1,
δ δ
δ
δ
α δ
−
−
∈ × . In this region, the
unique PE equilibrium is associated with a welfare of
( ) ( ) 3 10 3
7
48 4 16
2
P
w v k
α δ δ δ − −
= − − + − . In the same region, the unique ME equilibrium is
associated with a welfare of [ ]
27 2 1
3 48 4 3
2
M
w v k
α
α δ δ −
= − − + + . Welfare is higher
under ME if
( ) ( )
( )
2
5 3 6 4
16 4 3
0
δ δ α δ
δ α δ
− − −  
 
−
> , which is true in the given region.
Now consider the region ( )
( )
( )
( )
5 3
3
6 4 2
, , 1,
δ δ
δ
α δ
−
−
∈ ∞ × . In this region, the
unique PE equilibrium is associated with a welfare of
( ) ( ) 3 10 3
7
48 4 16
2
P
w v k
α δ δ δ − −
= − − + − . In the same region, the unique ME equilibrium is
associated with a welfare of
( ) ( ) 3 10 3
7
48 4 16
2
M
w v k
α δ δ δ − −
= − − + − , which is the same as
the welfare in the ME equilibrium.
201

Now consider the region ( ) ( ) ( )
3 5 3
4 6 2
, , ,
δ δ
α δ ∈ × ∞ . In this region, the unique
PE equilibrium is associated with a welfare of
127 9
192 16
2
P
w v k
α
δ
= − − + . In the same
region, the unique ME equilibrium is associated with a welfare of
7
48
2
M
w v k = − − . Welfare under ME is higher if
( )
( )
10 3
4 3
δ δ
δ
α
−
−
< , which is not true in
this region.
Now consider the region ( ) ( ) ( )
5 3
6 2
, , ,
δ
α δ ∈ ∞ × ∞ . In this region, the unique
PE equilibrium is associated with a welfare of
127 9
192 16
2
P
w v k
α
δ
= − − + . In the same
region, the unique ME equilibrium is associated with a welfare of
[ ]
27 2 1
3 48 4 3
2
M
w v k
α
α δ δ −
= − − + + . Welfare under ME is higher if
256 192
0
δ
α δ −
> , which
is true in this region.
Combining the results from the various parameter regions we have the
required result that social welfare in the sequential setting is the same or higher
under ME than under PE if
5
6
δ
α > , else it is lower. ▲

Proposition 3.7 PRITV’s profit in the sequential setting is the same or lower
under ME than under PE.
Proof.
We will split the parameter space into regions and prove the proposition for each
region of the parameter space. Consider first the region ( ) ( ) ( )
3 5 3
4 6 2
, , 1,
δ δ
α δ ∈ × . In
202

this region, the unique PE equilibrium has PRITV’s profit as
( )
2
1
8 1
P
k
δ
π
+
= − . In the
same region, the unique ME equilibrium has PRITV’s profit as
1
8 1
M
k π = − . Since
viewer nuisance per ad is strictly positive in this region, clearly PRITV’s profit is
lower under ME than under PE.
Now consider the region ( )
( )
( )
( )
5 3
5 3
6 6 4 2
, , 1,
δ δ
δ
δ
α δ
−
−
∈ × . In this region, the
unique PE equilibrium has PRITV’s profit as
( )
2
1
8 1
P
k
δ
π
+
= − . In the same region,
the unique ME equilibrium has PRITV’s profit as
( )
( )
2
2
5 4
1
2 4 3
M
k
α δ
α δ
π
−
−
= − . PRITV profit
is lower under ME if ( )
( )
( )
2
2
2
4 5 4
4 3
1 0
α δ
α δ
δ
−
−
+ − > . Now, this expression is monotone
decreasing in α and monotone increasing in δ . So, it just needs to be evaluated at
the lowest value of α and the highest value of δ . Therefore, evaluating it at the
lowest value of α and the highest value of δ reveals that PRITV’s profit is lower
under ME.
Now consider the region ( )
( )
( )
( )
5 3
3
6 4 2
, , 1,
δ δ
δ
α δ
−
−
∈ ∞ × . In this region, the
unique PE equilibrium has PRITV’s profit as
( )
2
1
8 1
P
k
δ
π
+
= − . In the same region,
the unique ME equilibrium has PRITV’s profit as
( )
2
1
8 1
M
k
δ
π
+
= − . Clearly,
PRITV’s profit is the same under ME and PE for this region.
203

Now consider the region ( ) ( ) ( )
3 5 3
4 6 2
, , ,
δ δ
α δ ∈ × ∞ . In this region, the unique
PE equilibrium has PRITV’s profit as
25
32 1
P
k π = − . In the same region, the unique
ME equilibrium has PRITV’s profit as
1
8 1
M
k π = − . Clearly, PRITV’s profit is
lower under ME in this region.
Now consider the region ( ) ( ) ( )
5 3
6 2
, , ,
δ
α δ ∈ ∞ × ∞ . In this region, the unique
PE equilibrium has PRITV’s profit as
25
32 1
P
k π = − . In the same region, the unique
ME equilibrium has PRITV’s profit as
( )
( )
2
2
5 4
1
2 4 3
M
k
α δ
α δ
π
−
−
= − . PRITV’s profits are
lower under ME if
31
40
δ
α > , which is true in this region.
Combining the results from the various parameter regions we have the
required result that PRITV’s profit in the sequential setting is the same or lower
under ME than under PE. ▲

Proposition 3.8 PUBTV’s profit in the sequential setting is the same or lower
under ME than under PE.
Proof.
We will split the parameter space into regions and prove the proposition for each
region of the parameter space. Consider first the region ( ) ( ) ( )
3 5 3
4 6 2
, , 1,
δ δ
α δ ∈ × . In
this region, the unique PE equilibrium has PUBTV’s profit as
( ) 3
4 2
P
k
α δ
π
−
= − . In
204

the same region, the unique ME equilibrium has PUBTV’s profit as
2
M
k π =− .
Clearly, PUBTV’s profit is lower under ME than under PE.
Now consider the region ( )
( )
( )
( )
5 3
5 3
6 6 4 2
, , 1,
δ δ
δ
δ
α δ
−
−
∈ × . In this region, the
unique PE equilibrium has PUBTV’s profit as
( ) 3
4 2
P
k
α δ
π
−
= − . In the same region,
the unique ME equilibrium has PUBTV’s profit as
( )( )
( )
2
6 5 3 2
2
2 4 3
M
k
α α δ α δ
δ α δ
π
− −
−
= − . It can be
shown that PUBTV’s profit is lower under ME over this region.
Now consider the region ( )
( )
( )
( )
5 3
3
6 4 2
, , 1,
δ δ
δ
α δ
−
−
∈ ∞ × . In this region, the
unique PE equilibrium has PUBTV’s profit as
( ) 3
4 2
P
k
α δ
π
−
= − . In the same region,
the unique ME equilibrium has PUBTV’s profit as
( ) 3
4 2
M
k
α δ
π
−
= − . Clearly,
PUBTV’s profit is the same under ME and PE for this region.
Now consider the region ( ) ( ) ( )
3 5 3
4 6 2
, , ,
δ δ
α δ ∈ × ∞ . In this region, the unique
PE equilibrium has PRITV’s profit as
9
16 2
P
k
α
δ
π = − . In the same region, the unique
ME equilibrium has PRITV’s profit as
2
M
k π =− . Clearly, PRITV’s profit is lower
under ME in this region.
Now consider the region ( ) ( ) ( )
5 3
6 2
, , ,
δ
α δ ∈ ∞ × ∞ . In this region, the unique
PE equilibrium has PUBTV’s profit as
9
16 2
P
k
α
δ
π = − . In the same region, the
205

unique ME equilibrium has PUBTV’s profit as
( )( )
( )
2
6 5 3 2
2
2 4 3
M
k
α α δ α δ
δ α δ
π
− −
−
= − . PUBTV’s
profits are lower under ME if
( )
2
16 4 3
0
αδ
α δ −
> , which is true in this region.
Combining the results from the various parameter regions we have the
required result that PUBTV’s profit in the sequential setting is the same or lower
under ME than under PE. ▲

Proofs of Propositions in Chapter IV
Proposition 4.3 Social welfare under monopoly exceeds social welfare under
perfect competition whenever
* *
M C
λ λ >
Proof.
We have social welfare under monopoly exceeding social welfare under perfect
competition i.e.
M C
W W > if ABC > 0, where the terms A, B, and C are given
below.

( )
( )
( )
( ) ( )( )
1
8 2
2
3 2
2
3 2
2 1 1 2
H
H
s
nc
s H
H
A
B s
C s nc
ξ ξ
ξ ξ ξ
ξ
ξ
ξ ξ ξ ξ
−
− +
−
=
 
= − −
 
  = − + + + −
 

206

Clearly, A > 0 always. When
1
2
ξ < , term C is always positive. When
1
2
ξ > , we
have C > 0 when
( )
2
2 1
2 1
H
nc
s
ξ ξ ξ
ξ
− + +
−
> . Since the LHS is always greater than 1, and the
RHS is always less than 1 by virtue of the interior conditions for
*
C
λ and
*
M
λ , it is
always true that C > 0. Thus,
M C
W W > if B > 0, and vice versa. Clearly, B > 0 if
* *
M C
λ λ > , and vice versa. Therefore, welfare under monopoly exceeds welfare
under perfect competition whenever
* *
M C
λ λ > . ▲

Abstract (if available)

Abstract This dissertation examines the economics of non-profit institutions in the television industry and agriculture. We analyze two kinds of non-profit institutions: (1) state owned welfare maximizing television broadcasters, and (2) agricultural cooperatives. With regard to the television industry, our main results are as follows. Broadcasters’ decisions on advertising levels and viewer subscription fees are not always more optimal under mixed economy. Social welfare is always higher under mixed economy except in situations where the state broadcaster carries 0% advertising level (e.g. BBC, France TV etc.), or 100% advertising level. Private broadcaster profits are mostly lower under mixed economy, thereby lending credence to their complaints of suffering due to "anti-competitive" actions of state broadcasters. However, no regulatory action is required unless the state broadcaster carries 0% or 100% advertising level, since lower private broadcaster profits are consistent with higher welfare under mixed economy. The state broadcaster usually makes lower profits than its private counterpart, and sometimes even makes losses while fulfilling its public service broadcasting mandate. However, a welfare argument can be made for the continued existence and operation of loss making state broadcasters.

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

Essays in political economics

PDF

Essays on the economics of radio spectrum

PDF

Essays on human capital accumulation -- health and education

PDF

Essays on political economy of privatization

PDF

Essays on bundling and discounts

PDF

Pricing strategy in a duopolistic market: aviation industry in Middle East

PDF

The divergence of the economic fortunes of Hindus and Muslims in British India: a comparative institutional analysis

PDF

Essays on child labor and poverty in the context of a conditional cash transfer program in Nicaragua

PDF

Adaptation, assets, and aspiration. Three essays on the economics of subjective well-being

PDF

Essays in international economics

PDF

Essays on the econometrics of program evaluation

PDF

The underdevelopment of India's Muslim minority: an institutional analysis

PDF

Costly quality, moral hazard and two-sided markets

PDF

Community forest management in Nepal: saving the forest at the expense of the poorest

PDF

Essays on labor and development economics

PDF

Essays on the economics of subjective well-being in transition countries

PDF

Essays in the study of institutions and development

PDF

Essays in empirical health economics

PDF

The decisions of migration and remittances in rural China

PDF

Essays on the empirics of risk and time preferences in Indonesia

Asset Metadata

Creator Nilakantan, Rahul (author)

Core Title Essays on the economics of non-profit institutions

School College of Letters, Arts and Sciences

Degree Doctor of Philosophy

Degree Program Economics

Publication Date 06/07/2010

Defense Date 05/28/2010

Publisher University of Southern California (original), University of Southern California. Libraries (digital)

Tag Economics,non-profit institutions,OAI-PMH Harvest

Language English

Contributor Electronically uploaded by the author (provenance)

Advisor Wilkie, Simon J. (committee chair), Alonso, Ricardo (committee member), Nugent, Jeffrey B. (committee member)

Creator Email nilakant@usc.edu,rahulnilakantan@gmail.com

Permanent Link (DOI) https://doi.org/10.25549/usctheses-m3116

Unique identifier UC1150833

Identifier etd-Nilakantan-3823 (filename),usctheses-m40 (legacy collection record id),usctheses-c127-335059 (legacy record id),usctheses-m3116 (legacy record id)

Legacy Identifier etd-Nilakantan-3823.pdf

Dmrecord 335059

Document Type Dissertation

Rights Nilakantan, Rahul

Type texts

Source University of Southern California (contributing entity), University of Southern California Dissertations and Theses (collection)

Repository Name Libraries, University of Southern California

Repository Location Los Angeles, California

Repository Email cisadmin@lib.usc.edu