that these two firms operate in a two-sided market, with publishers and advertisers as two
distinctly different groups of consumers. They apply a calibration technique to reveal
critical and actual loss, and therefore do not have any information about demand. In line
with what we explained when comparing Equations
(6.8) and (6.4), they find that apply-
ing a Lerner index from one side of the market could be quite misleading. They claim that
in this particular case it would lead to a too narrow market definition to only consider one
of the sides of the two-sided market. Given the competition authorities’ focus on market
definition and market shares, this might lead to type I errors (banning welfare-enhancing
mergers).
Google’s acquisition of DoubleClick was investigated in detail by both the European
Commission and the US Federal Trade Commission. Eventually both cleared the acqui-
sition, and both accepted that ad networks constitute two-sided platforms serving both
publishers and advertisers. For example, the European Commission stated the following:
“Ad networks generate revenues (paid by advertisers for access to publishers’ ad space inventory)
that are shared between the network manager (as intermediation fee) and publishers.”
In 2008, the Dutch Competition Authority (NMa) investigated an acquisition in the
Dutch yellow pages market. The two companies, Truvo Netherlands and European
Directories, published the only two nationwide print directories in the Netherlands.
The parties argued that this was a two-sided market, where the firms provide “usage”
to advertisers (look-ups or eyeballs) and provide users with information about businesses.
The evidence consisted of an econometric study inspired by
Rysman (2004) to detect
cross price effects and two-sidedness, and two surveys among users and advertisers to
detect substitution, i.e., diversion ratios among merging parties’ products. According
to
Camesasca et al. (2009), this is one of the first merger cases where the competition
authorities explicitly acknowledged the role of two-sidedness. The NMa wrote in their
decision:
“The supply of directories is thus marked by two-sidedness. It is accepted that this two-sidedness
can have a certain effect. Although a limited fall in the number of users will not lead to a com-
parable drop in the price of advertisements, the [NMa] Board does accept that (certainly in time) a
strong increase or decrease in usage will lead to a reaction from advertisers.” (para 108)
The effects of the two-sidedness were decisive for NMa’s decision to clear this merger.
At the same time, the European Commission investigated the US firm Travelport’s
acquisition of Worldspan Technologies (another US firm). This transaction was cleared
by the Commission in August 2007.
29
Both merging parties provide travel distribution
services, in particular through their respective “global distribution systems” (GDS)
Worldspan and Galileo (Travelport’s brand). These technical platforms match travel
29
The merger is discussed in Vannini (2008) and Rosati (2008), and briefly described in Filistrucchi et al.
(2010)
.
254
Handbook of Media Economics
content provided by airlines, hotel chains, car rental services, etc. on the one side, and the
demand for such content as conveyed by travel agents on the other side. The Commis-
sion acknowledged that GDS was a two-sided market, and wrote:
“… [t]he existence of the GDS platform is justified by the added value it creates. A GDS coordinates
the demand of [travel agents], thereby generating a positive network externality which is inter-
nalised by the [travel service providers].”
Another example is A-pressen’s acquisition of Edda Media, two media firms in Norway
which each owned several newspapers. The Norwegian Competition Authority did take
into account the two-sidedness by investigating both the advertising and the reader mar-
ket. They focused on two local areas with respect to a potential anticompetitive effect of
the acquisition, and made a survey among both advertisers and readers in both areas to
detect the diversion ratios. In one of the areas, they concluded from the surveys that the
two newspapers were close substitutes for both readers and advertisers, and they imposed
a remedy (to sell out one of the parties’ newspapers). In the other area, they found that the
two newspapers were close substitutes only on the advertising side, but also here they
imposed a remedy (to sell out one of the parties’ newspapers). In this case, they failed
to take into account the fact that increased market power on the advertising side could
lead to lower prices or higher quality on the reader side (see
Kind and Sørgard, 2013).
Filistrucchi et al. (2014) refer to several fairly recent merger cases in the TV industry.
They argue that competition authorities often fail to recognize that TV channels can be
differentiated not only on content as such but also on the amount of advertising they air.
One example is BSkyB’s acquisition of shares in ITV.
30
The UK Competition Commis-
sion did define the two sides of the market, and discussed whether viewers would regard
content of those two firms as close substitutes. But they failed to recognize that the firms
could be differentiated also on the amount of advertising. As a consequence, they did not
take into account any changes in the behavior in the advertiser market following the
acquisition. According to
Filistrucchi et al. (2014), they should have discussed whether
viewers dislike advertising on TV or not. If they dislike advertising, then some theory
models would predict that the merger would lead to more advertising. This could be
harmful for the viewers. However, as explained in the theory part, it is not at all trivial
to conclude on the welfare effect in such cases.
6.4.2.2 Empirical Studies
Let us contrast the approach we have seen in merger control with the approach used by
researchers who have analyzed the effect of media mergers. One first observation is that
surveys among consumers (example: readers and advertisers) are often used to analyze
30
BSkyB acquired in 2007 17.9% of the shares in ITV. For details, see Filistrucchi et al. (2014).
255
Merger Policy and Regulation in Media Industries
substitution between merging firms’ products, while this is seldom the case in empirical
studies carried out by researchers. In contrast, researchers often estimate demand from
observed quantities and prices on both sides of the market. One obvious reason for this
distinction is the time constraint in merger control. Antitrust authorities will be facing a
time constraint, and it may then be more feasible to conduct a survey among the con-
sumers than to undertake a full-fledged estimation of a demand system.
Rysman (2004) is one of the first empirical studies that analyze a two-sided market
(yellow pages).
31
Although this is not a study of a particular merger, he used the estimated
demand system to compare monopoly and oligopoly. He found that a more competitive
market is preferable. This study was the starting point for an econometric study that was
applied in an acquisition in the Dutch yellow pages market (see above). There are other
studies, though, that focus directly on mergers.
Chandra and Collard-Wexler (2009) study the effect of numerous mergers in the
newspaper market in Canada in the late 1990s. They apply a difference-in-difference
approach by comparing merging newspapers with non-merging newspapers. They find
that those newspapers that were involved in mergers did not have any different price
changes on either advertising prices or reader prices than the newspapers in the control
groups. However, they are careful in their interpretation of their results. It is possible that
the lack of price effect could be due to cost reductions following the mergers. It could also
be the result of a self-selection bias, where the newspapers that merged had great market
power and would have experienced even lower prices without a merger. Finally, they
claim that there is some evidence of mergers having more to do with political motives
and empire building than with profits.
Alternatively, one could estimate a structural model of a two-sided market in order to
investigate the interplay between the two sides of the market. Examples of some early
empirical studies along these lines are
Rysman (2004), Kaiser and Wright (2006), and
Argentesi and Fillistruchi (2007). However, there are only a few empirical studies that
estimate a structural model in order to investigate the possible price effects of a merger
in a two-sided market.
Both
Affeldt et al. (2013) and Filistrucchi et al. (2012) estimate a structural model for
the newspaper market in the Netherlands, and derive possible demand interdependencies
between the advertiser and the reader side in addition to the traditional one-sided
demand effects. They apply the estimated demand system to simulate possible price
effects of various mergers between newspapers, and compare the results with and without
incorporating the two-sidedness. They find in both studies that a one-sided approach
31
See also Kaiser and Wright (2006), who test empirically the price-cost margin in the magazine market on
the advertising and the reader side, respectively. They find that in this market the readers are subsidized
and the magazines make most of their profits from the advertising side.
Argentesi and Fillistruchi (2007)
follow a similar approach by estimating a structural model for the Italian newspaper market.
256
Handbook of Media Economics
might underestimate the price increase on advertising. The intuition is that the one-sided
approach does not take into account the diversion on the reader side among the merging
parties following a change in the prices of advertising. If readers like advertising, they will
divert from the newspaper that raises advertising prices and thereby reduces the volume
of advertising, and some of them will be recaptured by the other merging newspaper.
After the merger, this indirect network effect is taken into account by the merging firm,
leading to a stronger incentive to raise the advertising prices.
Fan (2013) also estimates a structural model for the newspaper market. She uses the
model to simulate the effects of a merger between two newspapers in the Minneapolis
area in the US that was blocked by the US Department of Justice. Contrary to
Affeldt
et al. (2013)
and Filistrucchi et al. (2012), she takes into account product characteristics
and the potential for a change in these following the merger. It is found that the merger
would have led to a reduced quality of content, higher prices to readers, and lower cir-
culation. She finds that ignoring the quality change could substantially underestimate the
welfare effect of such a merger.
Song (2013) estimates a structural model to investigate the price effects of a merger in
the market for TV magazines in Germany. It is a market with payment on both sides,
advertisers and readers, and, in line with
Kaiser and Wright (2006), he finds that the firms
set prices below marginal costs on the reader side of the market and earn profits on the
advertising side. He simulates the price effect of a merger to monopoly, and finds that
advertising prices would generally increase, while the reader prices in some cases would
drop after the merger.
Finally,
Jeziorski (2011) investigates the effect of the merger wave in the US radio
market that occurred between 1996 and 2006. He estimates a structural model and takes
into account the two-sidedness of this market (listeners and advertisers). Radio is free-to-
air, so the main effects of the mergers come through repositioning and changes in adver-
tising prices and quantities. It is found that the mergers led to more product variety for the
benefit of listeners, and less advertising both in small markets (with more market power)
and in large markets. On average, it is found that listeners benefit from the merger wave
while advertisers are worse off.
6.4.3 The Possible Non-Price Effect of Media Mergers
Although some of the empirical studies we have referred to also discuss some non-price
effects, such as product quality, most of the discussion so far has focused on the possible
price effect.
As explained earlier in this chapter, a merger might lead to more diversity. If media
firms have been competing fiercely on location/differentiation, they will in some
instances offer rather similar products and, in the extreme case, identical products. If
two firms then merge, they might have incentives to reposition their products and
257Merger Policy and Regulation in Media Industries
..................Content has been hidden....................
You can't read the all page of ebook, please click here login for view all page.