Even though they use a different equilibrium concept (passive beliefs as in Katz and
Shapiro, 1985
), Anderson et al. (2015a) find similar incremental pricing results as
Ambrus, Calvano, and Reisinger.
9
In particular, they show that both the price per ad
and the price per ad per viewer may increase when merging platforms have some
multi-homing consumers. The model of
Anderson et al. (2015a) is presented in more
detail in
Peitz and Reisinger (2015).
Both
Anderson et al. (2015a) and Ambrus et al. (2015) assume that availability of more
content will increase total consumption. In contrast,
Athey et al. (2013) keep total media
consumption per capita fixed in order to isolate the effects of the observation that con-
sumers spend less time on traditional media and more time on online media. Importantly,
their model is able to explain the collapse of advertising revenue for printed newspapers.
The fraction of multi-homing consumers has increased with online platforms according
to Athey et al., and when consumer attention is scarce, consumers divide their fixed con-
sumption between more outlets. The advertisers may then reach the same person more
times than necessary if they place ads on multiple platforms. On the contrary, if they place
ads on just a few of the platforms, advertisers would not reach all consumers. Conse-
quently, bigger is better: A platform with a larger audience may increase its ad prices since
it reaches more consumers with a lower risk of duplication (i.e., reaching the same
consumer more than once). This may obviously induce mergers among ad-financed plat-
forms. Despite differences in assumptions, the incremental pricing mechanisms in Athey
et al. are similar to those in
Anderson et al. (2015a,b) and Ambrus et al. (2015).
Also two recent papers by
Anderson and Peitz (2014a, b) analyze competition for
advertisers.
Anderson and Peitz (2014a) introduce aggregative games into media eco-
nomics, i.e., games where each firm i’s payoff depends on its own actions (ψ
i
) and the
aggregate of all n players’ actions Ψ ¼
X
n
j¼1
ψ
j
!
.
10
Denoting profits for firm i as
Π
i
(ψ
i
, Ψ ), its first-order condition reads
Π
i
1
ψ
i
, ΨðÞ+ Π
i
2
ψ
i
, ΨðÞ¼0:
Under certain assumptions the best reply function of firm i, r
i
, is a function of all firms’
actions, r
i
¼r
i
ΨðÞ, with r
i
being upward-sloping if actions are strategic complements.
Anderson and Peitz model the advertiser side of the market in a fairly standard way,
but the consumer side has several interesting features. Specifically, they use a represen-
tative consumer model, normalize the aggregate time that consumers spend on media
consumption to 1, and label the fraction of the time spent on platform i as λ
i
, with
9
See Anderson and Jullien (2015, this volume) and Peitz and Reisinger (2015, this volume) for further
discussions of equilibrium concepts.
10
See Selton (1970).
234
Handbook of Media Economics