platforms, consumers have become more prone to multi-home. For ad-financed online
newspapers, readers may easily browse news from multiple sources, whereas traditionally
they just read one newspaper. In addition to traditional media houses going online, we
have new ad-financed players, ranging from giants such as Google, Facebook, and You-
Tube to smaller bloggers, which increase competition for advertisers. Furthermore, the
noteworthy increase in new ways to access content, in particular by smartphones, implies
that consumers simultaneously access several platforms.
Ofcom (2014, p. 5) states that:
“UK adults squeeze over 11 hours’ worth of communications and media activity into less than nine
hours. The total volume of media and communications activities undertaken by an individual each
day equate to 11 hours 7 minutes. But as some media activities are conducted simultaneously, this is
squeezed into 8 hours 41 minutes per day.” As multi-homing becomes more widespread (see
Peitz and Reisinger, 2015, this volume), it becomes more important to understand how
competition for advertisers works, also when evaluating mergers.
Even if consumers have increased their consumption as the availability of multiple
platforms has increased, there is also a limit to attention. This may give ad-financed plat-
forms incentives to merge. As emphasized by Joshua Gans (Washington Post, September
17, 2013): “Space on the internet is infinite, but consumer attention is scarce, and now divided
between many outlets. So companies that want to reach consumers run the risk of hitting the same
person more than once if they advertise with multiple publications, or only hitting a slice of them if
they advertise on just one. That allows publications with a wider reach to command higher prices for
their ads, since they can reach more people with a lower risk of duplication.” (Formally shown in
Athey et al., 2013.)
The effects of mergers are also paid attention to in other chapters.
Sweeting (2015)
provides a discussion of the effects of merger in radio markets. Chandra and Kaiser (2015,
this volume) discuss mergers among printed media platforms (newspapers and maga-
zines), whereas
Armstrong and Crawford (2015, this volume) discuss mergers within
television.
6.2. PRICE AND QUANTITY EFFECTS OF ME RGERS
IN TWO-SIDED MARKETS
In two-sided markets two groups of users are mediated by a platform. At least one of the
groups imposes a positive network externality on the other group (see
Anderson and
Jullien, 2015
, this volume, for the definition of a two-sided market). In media markets,
the two groups are typically advertisers and consumers (viewers/readers/listeners), and
the platform is a media firm which delivers content to consumers. Ad-financed media
platforms then offer eyeballs (consumers) to the advertisers. There is a positive network
effect from the consumer side to the advertisers, since advertisers prefer to reach as many
consumers as possible. Advertisers’ impact on consumers depends on consumers’ attitude
to ads. If consumers dislike ads, as they typically do for television and radio (see
Sweeting,
228 Handbook of Media Economics
2015, this volume, on radio and Armstrong and Crawford, 2015, this volume, on tele-
vision), there is a negative externality from advertisers to consumers. In contrast, for mag-
azines, readers may like ads (see
Chandra and Kaiser, 2015, this volume), in which case
there is also a positive externality from advertisers to consumers.
Anderson and Jullien
(2015
, this volume) provide a complete treatment of the economics of two-sided mar-
kets. We restrict our attention to the effects of mergers. In
Section 6.3, we analyze how a
merger affects media platforms’ choice of genre. First, in this section, we hold product
characteristics and the number of goods fixed. We do this in order to focus on pure price
(and quantity) effects. General insight tells us that increased market power on one side of
the market tends to reduce prices on the other side (see, e.g.,
Rochet and Tirole, 2006;
Weyl, 2010
).
2
6.2.1 Backdrop: Price and Quantity Effects of Mergers in One-Sided
Markets
Holding factors like product characteristics, costs, and the number of goods fixed, the pure
price effects of mergers in one-sided markets are typically straightforward. To set the
scene, consider two newspapers that are financed by subscription fees alone (no ads). Stan-
dard economic theory predicts that the closer substitutes the newspapers are in the eyes of
the consumers, the less profit they will make if they compete, other things being equal.
Each of the newspapers has incentives to set a low subscription price in order to steal busi-
ness from the rival, and more so the more prone the consumers are to shift from one news-
paper to the other. If the newspapers merge (or set prices cooperatively), the owners will
internalize these business-stealing effects. Prices will thus unambiguously increase.
Farrell and Shapiro (1990) analyze possible welfare effects of mergers in a one-sided
market.
3
First, they consider a potential price increase subsequent to a merger. If there are
no changes in costs, they show that a merger between firms that produce substitutes will
always lead to higher prices. While a reduction in fixed costs will not change this, a reduc-
tion in marginal costs will matter. In a market with Cournot competition, Farrell and
Shapiro derive the criteria for how large the reduction in marginal costs must be to pre-
vent upward price pressure after the merger. Works that are more recent have extended
their analyses to other market structures, and in particular to Bertrand competition with
differentiated products.
4
This approach has become quite important in antitrust author-
ities’ analysis of mergers (see
Section 6.4 for more details). One reason is that antitrust
2
Weyl (2010) also provides an informal discussion of the effects of mergers in media markets.
3
Their model builds on the more informal model in Williamson (1968), where he shows the tradeoff in
welfare between reduction in competition and cost savings.
4
See Werden (1996) for an extension to Bertrand competition and differentiated products, which was
further developed in
Farrell and Shapiro (2010).
229
Merger Policy and Regulation in Media Industries
authorities apply a consumer welfare standard and the price effect on final consumers then
becomes very important.
The second issue
Farrell and Shapiro (1990) analyze is how a merger could lead to a
reallocation of production between firms. If a reallocation takes place, it might change the
cost level in the industry. To see this, consider a merger between two firms that have high
market shares due to low marginal costs. After the merger, they increase prices and cut
down on production, while the non-merging firms produce more. This is a reallocation
of production from firms with low costs to firms with high costs in production. This
explains why mergers between large firms can reduce total welfare, while mergers
between small firms can increase total welfare even if they lead to higher prices.
6.2.2 Two-Sided Markets and Single-Homing Consumers
In the seminal paper on two-sided media markets (Anderson and Coate, 2005), consumers
are restricted to attend a single ad-financed platform. This is termed single-homing in the
literature although advertisers multi-home and place ads on all platforms. The assumption of
single-homing closes down price competition for advertisers; each platform has monopoly
power in delivering its consumers to advertisers. This gives rise to the “competitive
bottleneck” problem identified by
Armstrong (2002, 2006). When considering the effects
of mergers among platforms, the single-homing assumption gives rise to a puzzle when
consumers dislike ads. As mentioned above, TV viewers typically dislike advertising.
Anderson and Coate (2005) consider competition between two advertising-financed
TV channels, assuming that advertising is a nuisance to viewers. In their model, the TV
channels compete by having few ads, thereby attracting viewers. They find that a shift to
monopoly—a merger between two TV channels—leads to more advertising and lower
advertising prices per viewer.
5
As in corresponding one-sided markets, competition
between the merging firms is eliminated. In contrast to traditional markets, elimination
of competition leads to lower advertising prices and correspondingly higher advertising
volumes. The consumers on the other side of the market—the viewers—will be worse
off, since they dislike advertising and their only “payment” is to incur the nuisance cost of
watching ads. The total welfare effect of the merger will depend on whether there was
under- or overprovision of advertising and programming (number of programs) before
the merger took place.
6
5
See Anderson and Coate (2005, Proposition 5), where they consider the regime where the programming
(number of programs being offered in the market) is not affected by such a merger. The counterintuitive
effect of tougher competition—lower prices in the advertising market—is also shown in
Barros et al.
(2004)
.
6
See Anderson and Coate (2005, Proposition 6). Kind et al. (2007) also discuss under- or overprovision of
advertising. They compare competition with collusion, where the latter can be interpreted as monopoly.
Ambrus et al. (2015) show that multi-homing will make it less likely that there is underprovision of ads in
the competitive equilibrium (see Footnote
3).
230
Handbook of Media Economics
Consumers’ attitude toward ads—whether they like or dislike advertising—can be of
importance for the effects of a merger in a two-sided market. In particular, with ad lovers,
the advertising market is expected to behave in the “normal” way, i.e., higher advertising
prices after a merger. Unfortunately, from the existing literature we cannot draw any
clear and robust conclusions on how consumers’ attitude toward advertising affects
welfare consequences of a merger in a two-sided market.
6.2.3 Two-Sided Markets and Multi-Homing Consumers
Several recent papers show that introducing competition for advertisers may resolve the
puzzling prediction that mergers might reduce advertising prices (
Ambrus et al., 2015;
Anderson and Peitz, 2014a,b; Anderson et al., 2015a
; Athey et al., 2015).
Ambrus et al. (2015) supersede Ambrus and Reisinger (2006). They consider a gen-
eral consumer demand function and allow (a share of) the consumers to multi-home.
They also relax the standard assumption that all media consumers necessarily notice all
ads to which they are exposed. More reasonably, they assume that any given consumer
becomes aware of an ad only with a certain probability. This has the important implica-
tion that it might be optimal to expose a consumer to the same ad several times. In
contrast to
Anderson et al. (2015a), which will be discussed later in this chapter, a
profit-maximizing advertiser may thus choose to advertise on several platforms even
in the extreme case where all consumers multi-home.
In their baseline,
Ambrus et al. (2015) consider a context with two advertising-
financed TV channels (i ¼1, 2), and initially make the non-critical assumption that
the advertisers are homogeneous. The consumers differ in their preferences for the
two channels, which are described by their types q :¼ q
1
, q
2
ðÞ
. Ambrus et al. analyze a
two-stage game where at stage one the platforms first non-cooperatively offer contracts
to the advertisers, before the advertisers and consumers at stage two decide whether to
join the platform(s).
7
Starting with the last stage, a consumer watches channel i if and only if q
i
γn
i
> 0,
where γ >0 is the disutility of ads and n
i
> 0 is the advertising intensity on the platform.
This implies that a consumer will single-home on channel 1, say, if q
1
γn
1
> 0 and
q
2
γn
2
< 0, while he will multi-home if q
i
γn
i
> 0 for i¼1, 2. If q
i
γn
i
< 0 for
i ¼1, 2, then the consumer will not watch either of the channels (zero-homing). Note
that this formulation implies that the decision of whether to watch channel i is indepen-
dent of the advertising volume on channel j. This reflects the fact that the incremental
7
The results we focus on below hold in a setting where media firms raise revenue only through the ad mar-
ket, as well as in one where they also raise revenue through subscription fees. We will only discuss the
version with pure ad financing.
231
Merger Policy and Regulation in Media Industries
utility of watching a given channel is independent of the utility of watching another
channel. Contrary to single-homing frameworks (like
Anderson and Coate, 2005),
the size of a channel’s audience i is thus not increasing in the rival’s advertising volume.
The value for a producer of informing a consumer of a product equals ω. The prob-
ability that a consumer who single-homes on platform i becomes aware of a given ad is
ϕ
i
(n
i
), while the analog probability for a multi-homing consumer equals ϕ
12
(n
1
, n
2
). The
probability functions are increasing and concave in their arguments.
The expected gross payoff for a firm which advertises on both platforms is equal to
un
1
, n
2
ðÞ¼ω
X
2
i¼1
D
i
n
1
, n
2
ðÞϕ
i
n
i
ðÞ+ D
12
n
1
, n
2
ðÞϕ
12
n
1
, n
2
ðÞ
"#
,
where D
i
is the number of single-homing consumers on platform i and D
12
is the number
of multi-homing consumers. With a fixed cost of advertising equal to t
i
, net payoff for the
advertiser equals u(n
1
,n
2
) t
1
t
2
. In equilibrium, each platform can extract the incre-
mental value it brings over its rival’s offer, which means that
t
d
1
¼un
d
1
, n
d
2
u 0, n
d
2
and t
d
2
¼un
d
1
, n
d
2
un
d
1
,0
:
By advertising at both channels, the probability of reaching a multi-homer increases by
ϕ
12
ϕ
j
units. Platform i can therefore charge ωD
12
ϕ
12
ϕ
j
for the multi-homers
and ωD
i
ϕ
i
for its exclusive viewers. The mass of advertisers is set to 1, which will also be
the number of advertisers in equilibrium (since they are assumed to be homogeneous).
Profits for channel i are thus equal to
Π
d
i
¼t
d
i
¼ω D
i
ϕ
i
+ D
12
ϕ
12
ϕ
j
hi
:
If channel i increases the advertising intensity by one unit, the probability that an exclu-
sive or a multi-homing consumer becomes aware of an ad increases by @ϕ
i
=@n
i
and
@ϕ
12
=@n
i
units, respectively. The marginal value of this for the channel is
ω D
i
@ϕ
i
=@n
i
ðÞ+ D
12
@ϕ
12
=@n
i
ðÞ½. However, since the consumers dislike ads, the channel
will also lose some viewers (@D
i
=@n
i
exclusive and @D
12
=@n
i
multi-homers). The mar-
ginal costs of increasing the advertising intensity thus equal MC
d
i
¼ωϕ
i
@D
i
=@n
i
ðÞ½
+ ϕ
12
ϕ
j
@ϕ
12
=@n
i
ðÞ. The equilibrium advertising intensity is consequently charac-
terized by
D
i
@ϕ
i
@n
i
+ D
12
@ϕ
12
@n
i
¼ ϕ
i
@D
i
@n
i
+ ϕ
12
ϕ
j
@ϕ
12
@n
i
:
Now suppose that the two channels merge, forming a two-channel monopoly. Since the
advertisers are homogeneous, the platforms will set the fixed advertising rates such that
they extract all surplus. Profits for the merged company are consequently equal to
232 Handbook of Media Economics
..................Content has been hidden....................
You can't read the all page of ebook, please click here login for view all page.