Download The Efficiency-Enhancing Effects of Non

The Efficiency-Enhancing Effects of
Non-Horizontal Mergers
A report by
Simon Bishop
Andrea Lofaro
Francesco Rosati
Juliet Young
This report has been produced by RBB Economics as part of a project financed by the
European Commission (Enterprise and Industry Directorate-General). Its content
therefore represents RBB Economic’s own views on the subject matter.
The views expressed in this report, as well as the information included in it, do not
necessarily reflect the opinion or position of the European Commission and in no way
commit the institution.
Acknowledgments
The authors wish to express their gratitude to Miguel de la Mano, Rod Meiklejohn, Stefano
Vannini and the staff at the European Commission, as well as to Luke Froeb, Alistair Lindsay
and the FTC staff for their helpful comments and suggestions. They are also grateful to all their
colleagues at RBB Economics for the frequent discussions and comments, and in particular to
Markus Baldauf, Jenny Burrage and Leslie Neubecker for their help in the research and in
reviewing the draft.
For further information, contact
European Commission
Enterprise and Industry Directorate-General
Unit A4 - Competitiveness Aspects of Competition Policy
B-1049 Brussels
Fax: (32-2) 296 68 14
E-mail: [email protected]
The PDF Version of the study can be obtained on the link:
http://europa.eu.int/comm/enterprise/library/lib-competition/libr-competition.html
Luxembourg: Office for Official Publications of the European Communities, 2005
ISBN 92-894-9587-1
© European Communities, 2005
Reproduction is authorised provided the source is acknowledged
Printed in Belgium
Table of Contents
EXECUTIVE SUMMARY
I
1
INTRODUCTION
1
2
THE FUNDAMENTALS OF NON-HORIZONTAL MERGERS
3
2.1
Categorisation of Mergers
3
2.2
Assessment of Horizontal and Non-Horizontal Mergers
2.2.1 Mergers between producers of substitute products
2.2.2 Mergers between producers of complementary or unrelated products
2.2.3 Which efficiency gains benefit consumers?
6
6
9
11
2.3
Are Non-Horizontal Mergers Always Pro-Competitive?
2.3.1 Potential anti-competitive concerns in non-horizontal mergers
2.3.2 The Chicago School and “Internalizing Complementary Efficiencies”: a
presumption of innocence
2.3.3 Post-Chicago: exceptions to the rule
12
12
Summary
19
2.4
3
14
15
PRO-COMPETITIVE MOTIVATIONS FOR VERTICAL MERGERS
21
3.1
Introduction
21
3.2
Increasing Pricing Efficiency
3.2.1 Avoiding double marginalisation
3.2.2 Avoiding input substitution
3.2.3 Lowering rivals’ costs
3.2.4 Price discrimination
23
23
37
41
44
3.3
Increasing Productive Efficiency
3.3.1 Economies of scope
3.3.2 Supply assurance
3.3.3 Improved information flow and coordination
3.3.4 Managerial efficiency
46
46
48
51
54
3.4
Preventing Profit Expropriation
3.4.1 Free riding on pre-sale services
3.4.2 Internalisation of R&D benefits: horizontal spillovers
3.4.3 Internalisation of R&D benefits: vertical spillovers
55
55
59
61
3.5
Incomplete Contracts and Transaction Costs
3.5.1 Incomplete contracts
3.5.2 The hold-up problem
3.5.3 General transaction costs
64
65
69
72
4
5
6
PRO-COMPETITIVE MOTIVATIONS FOR CONGLOMERATE MERGERS
73
4.1
Introduction
73
4.2
Increasing Pricing Efficiency
4.2.1 Avoiding the Cournot effect
4.2.2 Price discrimination (metering)
4.2.3 Reducing customer heterogeneity
77
78
88
91
4.3
Increasing Productive Efficiency
4.3.1 Economies of scope and scale
4.3.2 Financial efficiency
4.3.3 Countervailing buying power
4.3.4 Managerial efficiency
93
93
97
99
100
4.4
Preventing Profit Expropriation
4.4.1 Quality standardisation
4.4.2 Internalisation of R&D benefits: horizontal spillovers
4.4.3 Internalisation of R&D benefits: complementary spillovers
100
100
101
102
4.5
Incomplete Contracts and Transaction Costs
4.5.1 Incomplete contracts and moral hazard
4.5.2 One-stop-shopping
103
103
103
POLICY RECOMMENDATIONS
105
5.1
Introduction
105
5.2
Efficiencies Arising From Non-Horizontal Mergers
5.2.1 Vertical mergers
5.2.2 Conglomerate mergers
107
107
113
5.3
Assessment of Non-Horizontal Merger Efficiencies
5.3.1 Introduction
5.3.2 An illustrative example
5.3.3 For many efficiencies it is not possible to separately assess the anticompetitive and pro-competitive effects
5.3.4 An economic presumption of pro-competitiveness
119
119
120
5.4
A Unified Approach to Assessing Competitive Effects
123
5.5
Summary and Conclusions
124
121
121
CASE STUDIES
125
6.1
Guinness/Grand Met
6.1.1 The merger
6.1.2 Treatment by antitrust authorities
6.1.3 Analysis
125
125
127
129
6.2
Boeing/Hughes
6.2.1 The merger
6.2.2 Treatment by antitrust authority
6.2.3 Analysis
131
132
132
135
7
6.3
AOL/Time Warner
6.3.1 The merger
6.3.2 Treatment by antitrust authorities
6.3.3 Analysis
136
136
138
140
6.4
Tetra-Laval/Sidel
6.4.1 The merger
6.4.2 Treatment by antitrust authorities
6.4.3 Analysis
143
143
145
146
BIBLIOGRAPHY
APPENDIX A : INTENSITY OF COMPETITION AND DOUBLE MARGINALISATION
147
155
A.1
Model
A.1.1 Assumptions
A.1.2 Time structure of the model
155
155
156
A.2
Solution in the case of non-integration
156
A.3
Solution in the case of vertical integration
157
A.4
Benefits from vertical integration
157
APPENDIX B: GLOSSARY
159
Executive Summary
This report has been prepared by RBB Economics on behalf of the Enterprise and Industry
Directorate-General of the European Commission. RBB has been commissioned to produce a
report considering the efficiency enhancing effects of non-horizontal (i.e., vertical and
conglomerate) mergers, in order to provide the Commission with economic guidance on the
models, tools and techniques relevant to the assessment of such efficiency claims.
Definition of Non-Horizontal Mergers
Non-horizontal mergers take place between firms which do not operate in the same market.
They include both vertical and conglomerate mergers. Vertical mergers involve firms at adjacent
levels of the supply chain. For example, a manufacturer may merge with its downstream
distributor or with its upstream input supplier. Conglomerate mergers take place between firms
that operate in different markets. Conglomerate mergers can be further broken down depending
on whether the merger involves complementary or unrelated products. Complementary products
are those such as bread and butter or staples and staplers. In either case, an increase in the
price of one good will decrease demand in the other. (This contrasts with substitute products
where an increase in the price of one good will induce an increase in demand for the other.)
There is a close analogy between vertical relationships and the relationship between producers
of complementary products. Vertically related products can be thought of as complements in
that an increase in the price of a downstream product will decrease the demand for upstream
inputs. The main difference between the two types of mergers is that in the case of the vertically
related goods it is the downstream firm that combines the two products, whereas in a
complementary relationship, the integration of the products is left to the consumers. Staples, for
example, could be thought of as an “input” of the stapler, but the combination of the products is
left to the consumer. Analogously, retailing can be considered as a service complementary to
manufacturing.
Distinction between Horizontal and Non-Horizontal Mergers
There are important differences in the competitive assessment of horizontal and non-horizontal
mergers. Horizontal mergers bring together manufacturers of substitute products, and so
remove a direct competitive constraint. Whether the removal of this pricing constraint leads to a
price increase requires a detailed assessment.1 But even to the extent to which the removal of
this direct competitive constraint does lead to an increase in price, this may be offset by
counterbalancing efficiencies so that the price might fall below the pre-merger level.
1
We do not subscribe to the view that all horizontal mergers necessarily give rise to an increase in price. Such
predictions are posited on simple theoretical models which, although providing a useful analytical framework, do not
provide a suitable basis for making real-world policy predictions.
Page i
In contrast, non-horizontal mergers bring together suppliers of complementary (or unrelated)
products, and so do not directly eliminate competition. Indeed, there is generally an incentive for
firms to lower prices, in line with consumer interests. This is because, in a complementary
relationship, a reduction in the price of one good will increase demand for the other. For
example, a decrease in a distributor’s margin would increase the sales of a manufacturer’s
product. Similarly, a decrease in the price of cars may increase the sales of complementary
products such as tyres. If firms are under separate ownership the beneficial impact of a price
decrease (or quality improvement) on the other party’s sales will not be taken into account.
However, if the two parties merge, such effects would be considered and would provide an
additional incentive to lower price.
Although such price reductions brought about by non-horizontal mergers are usually procompetitive, they may, under certain particular circumstances, give rise to anti-competitive
outcomes. The primary competitive concern in these cases often relies on the assumption that,
post merger, the ability of rival firms to compete will be reduced to such an extent that they are
marginalised or driven from the market altogether. If this applies, the merged firm will be able to
increase prices.2 However, the direct effect of non-horizontal mergers is generally procompetitive. Only if certain conditions hold, do such direct effects lead to anti-competitive
outcomes. This implies that, although concerns of foreclosure cannot be excluded a priori, in the
majority of cases non-horizontal mergers are pro-competitive.3 Therefore the analysis of such
anti-competitive effects must carefully specify the conditions that give rise to the purported
outcome, and must go beyond a mere theoretical assessment by accounting for observed
industry characteristics and behaviour.
Sources of Efficiency Gain
There is great scope for non-horizontal mergers to give rise to efficiency gains. As described
above, there is a direct incentive to reduce price or to improve quality post merger as this will
increase the sales of complementary products, as well as of the good itself. In addition,
however, as products manufactured by firms in vertical and complementary relationships are, by
definition, used together, non-horizontal mergers can give rise to efficiency gains arising from
improved interaction between the two products. Such efficiency gains are often not associated
with a direct impact on pricing behaviour but may improve quality, increase variety, or increase
levels of investment.
The potential sources of efficiency gains from vertical and conglomerate mergers result from a
wide variety of mechanisms and occur in a broad range of situations. However, they can be
broadly grouped under the following four headings:
•
Page ii
Increasing pricing efficiency: In many cases, charging a uniform unit price will not
be an optimal way of organising a supply chain or of providing complementary
2
Alternatively, entry or expansion may be deterred, allowing the firm to preserve its market power.
3
Nowhere in the extensive literature on the competitive effects of non-horizontal mergers is this view disputed. Rather
the dispute focuses on how many non-horizontal mergers fall outside this general presumption.
products. However, more sophisticated pricing mechanisms may be costly or
impossible to introduce. In such cases, by removing the need for external pricing,
non-horizontal integration can lead to a more efficient outcome.
•
Increasing productive efficiency: Non-horizontal integration may result in a more
efficient use of inputs and/or productive assets, due to the existence of economies
of scope or scale, improved managerial or financial efficiency, supply assurance or
increased buyer power.
•
Preventing profit expropriation: In some cases, a firm may be unable to appropriate
the full benefits of its investments, as these are partly reaped by competitors. This
can result in sub-optimal levels of investment. By “internalising” such externalities,
non-horizontal mergers can lead to significant efficiency gains.
•
Finally, market transactions may be subject to transaction costs and incomplete
contracts. Where contracts are not fully specified, a party may be able to exploit
contractual loopholes to the disadvantage of their vertical partners. To overcome
this problem, costly monitoring or incentive mechanisms may have to be used. In
certain situations, such as those that give rise to the so-called “hold-up” problem,
investment may be deterred altogether. Furthermore, at a general level, all
contractual relationships are subject to transaction costs, such as those incurred in
drawing up the contract or searching for an appropriate partner. By eliminating such
contractual problems, both vertical and conglomerate mergers can give rise to an
efficiency gain.
Efficiencies are thus an important motive for non-horizontal mergers. In some cases there may
be alternative means of achieving these benefits, such as non-linear pricing, vertical restraints,
other contractual obligations, commitment devices, or reputation considerations. However, such
mechanisms may only be applicable in specific circumstances, and may give rise to additional
costs and problems. Thus, in general, these alternative mechanisms do little to lessen the
power of efficiency rationales as a motive for non-horizontal integration. It should therefore not
be surprising that non-horizontally related firms from a wide range of sectors may wish to
integrate to a certain extent, in order to realise such gains.
Policy Implications
Based on the analysis presented in this report, we derive two main policy recommendations
regarding how efficiencies generated by non-horizontal mergers should be taken into account in
the competitive assessment of mergers.
Our first recommendation pertains to the current praxis of merger assessment. The typical
approach to assessing efficiencies as set out in the EC horizontal merger guidelines divides the
competitive assessment into two broad steps: first the potential anti-competitive effects of the
merger are examined. Then, where a merger is held to give rise to anti-competitive outcomes,
the extent to which efficiencies generated by the merger might offset such concerns is
Page iii
examined.4 Our first policy conclusion is that this two-step approach is both impractical and
inappropriate for assessing many of the efficiencies generated by non-horizontal mergers.
The focus of the EC horizontal merger guidelines’ assessment of merger efficiencies is on
marginal cost savings. The impact of such efficiencies on pricing incentives can readily be
assessed separately from any lessening of competition arising from the horizontal merger.5 But,
in practice, many of the most important efficiencies generated by non-horizontal mergers (and
some would argue also horizontal mergers) are the result of synergies arising from the
combination of complementary assets rather than direct cost savings as such.6 Incorporating
these types of efficiencies, to which non-horizontal mergers are most likely to give rise, into the
overall competitive assessment cannot easily be done as a separate stand-alone exercise.
Indeed, the two-step approach to assessing efficiencies is appropriate if and only if the following
two conditions hold.
•
The anti-competitive issues raised by the merger can be assessed separately from
the likely effects of any efficiency.
•
A direct trade-off comparison can be made (i.e., the efficiency can be readily
translated into an impact on pricing incentives).
Not only is it difficult to make a direct trade-off comparison between the pro-competitive
efficiency effects and the adverse effects on pricing and/or quality,7 but in many cases it is
simply not possible to assess the potential anti-competitive concerns separately from the procompetitive effects for the simple reason that both have the same source. Indeed, in many
instances, the theory of competitive harm posited in the appraisal of non-horizontal mergers
arises because the merger generates an efficiency. For example, in assessing the competitive
effects of a merger that is expected to give rise to bundling, the anti-competitive concern
(foreclosure of rivals who are not able to bundle) has the same source as the pro-competitive
benefits (the elimination of pricing inefficiencies).8 This implies that for many types of efficiency
generated by non-horizontal mergers, the assessment of efficiencies necessarily forms part of
an integrated unified competitive assessment.9
Second, there should be an economic presumption that non-horizontal mergers are procompetitive.10 This conclusion derives from a fundamental difference between horizontal and
non-horizontal mergers. By definition, a horizontal merger removes a direct competitive
Page iv
4
It should be stressed that in practice this approach to assessing the competitive effects of a horizontal merger is only
appropriate for (a) assessing unilateral effects concerns and (b) only applies to marginal cost efficiencies.
5
Indeed, it can be argued that the current approach to assessing efficiencies generated by horizontal mergers is itself
too restrictive due to its emphasis on marginal cost reductions.
6
This point has also been made by Kolasky (2004). See also Teece (1980, 1982), and Jorde and Teece (1990).
7
For example, how can one translate the efficiency generated by combining complementary assets into an impact on
pricing incentives?
8
See Church (2004) for a recent review of the potential anti-competitive effects of non-horizontal mergers.
9
Paradoxically, in such cases, the more modest the efficiency the less likely it is that the marginalisation of competitors
will take place.
10
See Section 2. In particular, most economists recognise that non-horizontal mergers are generally much more likely to
generate efficiencies than horizontal mergers. See for example Reiffen and Vita (1995), and Scheffman and Higgins
(2004).
constraint and this raises the possibility that post-merger prices will increase to the detriment of
consumers.11 In marked contrast, non-horizontal mergers do not, also by definition, remove a
direct competitive constraint. Moreover, the general impact of a non-horizontal merger on prices
is to reduce them as a result of eliminating externalities and other inefficiencies that might have
existed pre merger. While this is not to say that no non-horizontal merger is anti-competitive, it
does indicate an economic presumption (albeit rebuttable) that such mergers are procompetitive.12 This view is echoed by the European Court of Justice in its Tetra Laval
judgement.13
This economic presumption has important implications for the stance adopted in the integrated
assessment of the likely impact of a non-horizontal merger where the pro-competitive efficiency
effects of the merger cannot be disentangled from the potential anti-competitive effects. Namely,
in the competitive assessment of those commercial strategies that might give rise to both
efficiencies and potential anti-competitive concerns, it should be assumed that the procompetitive effects are predominant, i.e., any efficiency generated by a non-horizontal merger is
deemed to give rise to pro-competitive outcomes, unless it can be demonstrated otherwise. This
is in marked contrast to the approach, often labelled the efficiency offence, whereby efficiencies
tend to be viewed as inherently problematic.
11
We disagree with those commentators who argue that every horizontal merger gives rise to a price increase. Not only
are such claims not well-founded in theory, they are unsupported by empirical evidence.
12
This economic presumption does not necessarily imply a legal presumption. Nonetheless, it entails that compelling
evidence will be required to establish that the merger leads to a significant impediment of effective competition.
13
Judgment of the Court in case C-12/03 P, 15 February 2005, paragraph 44.
Page v
1 Introduction
This report has been prepared by RBB Economics on behalf of the Enterprise and Industry
Directorate-General of the European Commission. RBB has been commissioned to produce a
report considering the efficiency enhancing effects of non-horizontal mergers, in order to provide
the Commission with economic guidance on the models, tools and techniques relevant to the
assessment of such efficiency claims.
The report provides a detailed assessment of the type of efficiencies that non-horizontal
mergers generate, drawing on both relevant economic literature and recent antitrust cases. It
also recommends an approach for incorporating efficiencies into the overall competitive
assessment of non-horizontal mergers, considering key differences between the issues raised
by horizontal and non-horizontal mergers.
The report is organised as follows. Section 2 provides the background necessary for an
understanding of non-horizontal mergers. It defines and categorises such mergers and
considers their fundamental nature, contrasting them with horizontal mergers.
Section 3 provides a detailed description of the sources of efficiency gains that can arise from
vertical integration. Section 4 considers the efficiency gains that can arise from conglomerate
mergers. In both sections, these efficiencies are categorised into four groups: (i) increased
pricing efficiency; (ii) improved productive efficiency; (iii) the prevention of profit expropriation;
and (iv) overcoming problems arising from incomplete contracts and/or transaction costs.
Section 5 provides practical policy recommendations for the competitive assessment of nonhorizontal mergers and, in particular, how that assessment should incorporate any efficiencies
that might be generated by such mergers. It summarises the efficiencies to which vertical and
conglomerate mergers respectively give rise, and discusses the criteria for assessing such
efficiencies, contrasting them to the treatment given to efficiencies in horizontal mergers. It then
provides an analytical framework for assessing the competitive effects of non-horizontal
mergers.
Section 6 considers several case studies of non-horizontal mergers. The cases considered are
Guinness/Grand Met; Boeing/Hughes; AOL/Time Warner; and Tetra-Laval/Sidel. Each case,
and the treatment of the case by the antitrust authority, is reviewed in the light of the analysis
presented in the previous sections of this report.
Page 1
2 The Fundamentals of Non-Horizontal Mergers
This section sets out a number of important principles on the assessment of non-horizontal (i.e.,
vertical and conglomerate) mergers. It shows that non-horizontal mergers differ in a
fundamental manner from horizontal mergers, as they do not eliminate a direct competitive
constraint. This distinction suggests important differences in the treatment and the analysis of
efficiencies arising from horizontal and non-horizontal mergers. This section is organised as
follows.
Section 2.1 begins by describing the categorisation of mergers as horizontal, vertical or
conglomerate. Conglomerate mergers can be further categorised depending on whether the
merger involves unrelated or complementary products.
Section 2.2 highlights some important differences in the competitive assessment of horizontal
and non-horizontal concentrations. Horizontal mergers bring together suppliers of substitute
products and so remove a direct competitive constraint. Where the constraint is sufficiently
important, this can result in an increase in prices post merger. However, even if this is the case,
counterbalancing efficiencies may more than offset the lessening of competition and may
actually result in prices declining. In contrast, non-horizontal mergers bring together suppliers of
complementary (or unrelated) products. There is therefore usually no need to use efficiencies to
counterbalance a direct reduction in competition.
Section 2.2.3 considers whether non-horizontal mergers could nevertheless eliminate
competition indirectly, e.g., through tying, bundling, or vertical foreclosure. The Chicago School
theories, and the more recent concept of “ICE” (Internalizing Complementary Efficiencies)
developed by Farrell and Weiser (2003), suggest that in most cases, such concerns are
misplaced. In general, non-horizontal mergers will be pro-competitive and such exclusionary
behaviour can only occur under very stringent conditions. This implies that, although concerns
of foreclosure cannot be excluded a priori, in the vast majority of cases a non-horizontal merger
is likely to be pro-competitive. In short, anti-competitive non-horizontal mergers are the
exception rather than the rule.
2.1
Categorisation of Mergers
Mergers can be categorised as being horizontal, vertical or conglomerate.14 Of these categories,
this report concerns only non-horizontal (i.e., vertical and conglomerate) issues. However, it is
useful to understand the distinction between horizontal and non-horizontal mergers, as this
leads to important differences in their motivations and treatment. Below we give definitions and
examples of each type of concentration. Graphical illustrations are provided in Figure 2-1.
14
It is of course possible for a merger to have aspects that fall within more than one of these categories. For example, a
merger may affect several markets, with potential concerns being horizontal in some markets and vertical in others.
Page 3
Figure
2-1:
Examples
Horizontal
of
horizontal,
Vertical
vertical
and
conglomerate
mergers
Conglomerate
Horizontal mergers involve companies that operate at the same level of the supply chain,
producing substitute goods. Two products are termed substitutes if an increase in the price of
one good induces an increase in demand for the other. For example, if the price of butter
increases, consumers may consider buying alternative spreads, such as margarine, and the
demand for these alternative products will consequently rise. Horizontal mergers therefore bring
together two competing firms. A simple example of a horizontal merger is the recent case
considered by the European Commission (EC) involving Kimberly-Clark and Scott Paper, both
of which produce toilet tissue.15 The EC also considered the merger between Procter & Gamble,
manufacturing Always© sanitary towels, and Schickerdanz, producing ob© tampons.16 If sanitary
towels and tampons are substitute goods, this would be a horizontal merger.
Vertical mergers, in contrast, involve companies that operate at different levels of the supply
chain. A common example is a merger between a wholesaler and its retailer, or a manufacturer
and its input supplier. A recent example considered by the European Commission is the
proposed merger between Tetra Laval and Sidel.17 Tetra produced aseptic carton packaging
equipment, including PET bottles. Sidel produced the Stretch Blow Moulding (SBM) machines
which were used by Tetra to make its PET bottles.
Conglomerate concentrations involve companies that operate in different markets, which are not
horizontally or vertically related. The Virgin group, comprising records stores, aeroplanes, train
services, financial products, and much more, provides one example of a conglomerate firm.
Another example is General Electric whose markets include jet engines, power generation,
Page 4
15
Case No. M.623, 1996
16
Case No. M.430, 1994
17
Case No. M.2416, 2001
financial services, plastics, medical imaging and news and information. Conglomerate mergers
can be further sub-divided into two types:
•
Mergers involving producers of complementary goods: Two products are termed
complements if a rise in the price of one good induces a decrease in demand for
the other. A simple example of a complementary relationship might be bread and
butter. If the price of bread rises, less bread is purchased, and this may lead to a
reduction in the demand for butter. Another example might be staplers and staples.
These products are complements as an increase in the price of staplers will
decrease the demand for staples, and vice versa. Thus a merger between a firm
producing staples and another producing staplers would be classified as a
conglomerate merger involving complementary goods. A recent example of an EC
case involving complementary products is provided by the GE/Honeywell merger.
GE produced engines for large aircraft and Honeywell produced complementary
avionic and non-avionic equipment.18
•
Mergers involving producers of unrelated goods: Conglomerate mergers may bring
together products which have no obvious commonality. That is, they are neither
substitutes nor complements. Many of the Virgin group’s acquisitions, including its
purchase of a UK railway franchise, provide examples of conglomerate mergers
involving unrelated goods. In some cases, unrelated goods may use similar
distribution channels. For example banking services and financial services may be
neither complements nor substitutes. However, they may be aimed at the same
group of customers, and so utilise similar promotion strategies and distribution
channels. Such products are sometimes said to be in “neighbouring markets”.
Although the goods may be unrelated from the point of view of the consumer, and
thereby not complementary in a strict sense, they can nevertheless be considered
complements from the point of view of firms at a certain level of the supply chain.
(Some authors therefore categorise mergers between neighbouring markets
together with mergers between complementary products, or define them as a third,
separate category of conglomerate merger.) For example, whisky and vodka can be
unrelated for consumers, but a retailer (e.g. a bar) can perceive them as
complementary, because if it stores only one product the attractiveness of the bar
to consumers may be reduced. If tampons and sanitary towels are not considered
substitutes, the Procter & Gamble/Schickerdanz merger could be classed as a
conglomerate merger involving neighbouring markets.19
Note that there is a close analogy between vertical relationships and the relationship between
producers of complementary products. Indeed, vertically related products can also be thought of
as complements, in that an increase in the price of a downstream product will decrease the
demand for upstream inputs. Indeed, the main difference between the two types of mergers is
that in the case of the vertically related good it is the downstream firm that combines the two
18
Case No. M.2220, 2001
19
According to some authors, it is also useful to identify conglomerate mergers between buyers of the same input which
produce non-competing outputs. Although mergers of this type may give rise to specific effects in certain circumstances
Page 5
products, whereas in a complementary relationship, the integration of the products is left to the
consumers.20 Thus, staples, for example, could be thought of as an “input” of a stapler, where
the combination of the products is left to the consumer. Analogously, retailing can be considered
as a service complementary to manufacturing. This ambiguity is illustrated by the merger
between Boeing and Hughes.21 Hughes produces geostationary satellites and Boeing produces
the equipment which delivers such satellites to space orbit. The majority of customers purchase
satellites and launch services separately, and thus the merger could be classed as a
conglomerate merger involving complementary products. However, some customers buy the
system as a whole and from this perspective the integration would be classed as vertical.22
2.2
Assessment of Horizontal and Non-Horizontal Mergers
This sub-section compares the fundamental nature of horizontal and non-horizontal mergers. It
first considers horizontal mergers between substitute goods and discusses how the elimination
of an important competitive constraint provides an incentive to increase prices post merger.
However, this does not mean that a price increase is inevitable. In particular, the incentive to
raise prices may be counterbalanced by the creation of efficiencies.
We then consider non-horizontal mergers between producers of complementary (or unrelated)
goods and discuss why such mergers not only not eliminate an important competitive constraint
but often give rise to incentives to reduce prices.
Finally, in Section 2.2.3, we briefly consider which types of efficiency gains may benefit
consumers. In particular, we discuss to what extent fixed cost reductions and qualitative
improvements may increase consumer welfare.
2.2.1
Mergers between producers of substitute products
Horizontal mergers involve producers of substitute goods. Their relationship is therefore
characterised by a positive cross-price elasticity. If firm 1 raised the price of its product
unilaterally, that would cause some customers to choose substitute products instead, thus
increasing the demand faced by firm 2. The possible harmful effects of a merger between two
competing producers can be understood in the context of these cross-price effects.
(most notably buying power increases), it is not clear whether this justifies the introduction of an additional category of
conglomerate mergers.
20
Page 6
The theoretical economic literature often assumes the existence of other differences between vertical and
complementary relationships. The most notable are the following: (1) sometimes, in vertical settings, pricing
mechanisms more sophisticated than simple linear prices may be used; (2) differences may arise in the extent to which
commitment is possible; and (3) the upstream firms are assumed to set their price and make purchase decisions before
the downstream firms do. Although these differences may lead to different results (in particular points 1 and 2), many
economic predictions derived in a complementary setting also hold in a vertical industry, and vice-versa. In particular,
the impact of assumption (3) on the qualitative results of economic models should not be over-estimated.
21
Case No. M.1879, 2000
22
Another example is provided by the personal computer industry, where consumers can purchase a finished product
including all components (mother board, video card, hard-disk, etc) or, alternatively, can purchase the components
separately and assemble them themselves.
As is well established in economic theory, a profit maximising firm will produce output until
marginal cost is equal to marginal revenue. Any increase above this equilibrium level would be
detrimental to the firm in question, but would obviously benefit producers of substitute goods by
increasing demand for their products. Under separate ownership, firms do not take this benefit
to rival firms into account. However, if two producers of substitute goods come under common
ownership, they will consider the beneficial effect that their price increase has on the profits of
the other party. (More formally, the marginal revenue obtained from a price rise increases, as it
now includes the extra profit obtained by the other party.) There is therefore an additional
incentive to increase price which was not present prior to the merger. As this argument is
symmetric both of the merging parties have an incentive to increase price. The extent of this
incentive will depend upon the level of competition from other producers of substitute goods.
Where competition from these firms is intense, an increase in price by the merging parties will
divert sales to these rival firms rather than to each other.
Despite the incentive that the unified entity may have to increase prices, a post-merger price
increase is not inevitable. It can be prevented in a number of ways. First, post-merger
competition from existing producers of substitute goods may still be sufficiently strong. Second,
entry, repositioning and other dynamic effects may maintain the effectiveness of competition in
the market. Third, the merger may bring about efficiencies that provide an off-setting downward
impact on price. These efficiencies may more than counterbalance the post-merger incentive to
increase prices.
In assessing the extent to which efficiencies lower prices, it is important to distinguish between
marginal cost reductions and fixed cost reductions. Reductions in marginal costs will lead to
lower prices and so represent a consumer benefit. In general, reductions in fixed costs will not
feed directly through to consumer prices and it is therefore more difficult to assess the impact of
these dynamic efficiencies on consumer welfare (see Section 2.2.3).
An example of how marginal cost reductions can outweigh a post-merger incentive to reduce
prices is given in Box 2.1. This example illustrates how, in principle, the assessment of cost
efficiencies can play such an important role in the evaluation of horizontal mergers. Even if the
merger could eliminate an important competitive constraint, this could be offset by a reduction in
marginal cost brought about by the efficiency gain. As we show in the next sub-section, this is
not the case in a non-horizontal merger.
Page 7
Box 2.1: A simple numerical example of a merger between substitutable products
The incentive that firms producing substitutable products have to increase prices following a
merger can be illustrated by reference to the following simple example. Suppose that in an
industry there are only two firms each producing one variety of a differentiated commodity.
Inverse demand functions are:
p1 = a − q1 − dq 2
p 2 = a − q 2 − dq1
[
where a >0, pi is firm i’s price, qi is firm i’s output and d ∈ 0,1) indicates to what extent
the two products are differentiated. If d=0, products are unrelated while if d → 1 , products
tend to be perfectly homogeneous. Both firms produce at a marginal cost equal to c and
face no fixed costs. This implies that each firm’s profit can be simply written as follows:
Π i = ( pi − c ) qi = (a − qi − dq j − c ) qi
i,j = 1,2 and i ≠ j
If the two firms compete in price à la Bertrand, then equilibrium prices are:
p1 = p2 =
a(1 − d ) + c
2−d
Following a merger between the two firms, the new entity will maximise joint profits, resulting
in the following new price:
p1 = p2 =
a+c
2
It is straightforward to verify that the prices of both products increase after the merger
whatever the degree of substitutability between the two products. For instance, if a =7, c=3
and d=0.5, then post merger the equilibrium price would increase from p1 = p 2 = 4.3 to
p1 = p 2 = 5 . Suppose now that the merger brings about a reduction in the marginal cost of
the post-merger entity. In particular, we assume that after the merger the marginal cost
decreases from c=3 to c=0. This implies that post merger the new equilibrium price will be:
p1 = p2 =
a+c 7+0
=
= 3.5
2
2
In other words, despite the elimination of the competition between firms 1 and 2, due to the
reduction in the marginal cost of production, after the merger prices are lower than pre
merger.
Page 8
2.2.2
Mergers between producers of complementary or unrelated products
Vertical and conglomerate concentrations involve firms producing complementary (or unrelated)
products, rather than substitute goods. The relationship between producers of complementary
goods is defined by a negative cross-price elasticity. If a firm reduces the price of its product,
demand for its own product would increase. Consequently, sales of complementary goods
would also increase. This distinction between positive and negative cross-price elasticities gives
rise to a fundamental difference in the relationship between horizontal and non-horizontal
mergers.
As discussed above, in cases where products are substitutes each firm would benefit if the
other increased prices and this can give rise to a post-merger incentive to raise prices. This
effect is not present between producers of complementary goods. Moreover, in complementary
relationships, each firm would benefit if the other lowered the price of its product. When
producers of complementary goods are under separate ownership, the benefit that a price
decrease has on each other’s sales is not taken into account. However, if the firms were to
come under common ownership, the increase in the other party’s profits which comes about
following a decrease in price would be taken into account. This gives an incentive to reduce
prices post merger (even when both parties have market power), raising consumer welfare. This
difference in the incentives of firms engaged in a non-horizontal relationship from those
engaged in a horizontal relationship is acknowledged in the Commission’s Guidelines on
Vertical Restraints.23 A simple numerical example showing the post-merger incentive to lower
price in a non-horizontal setting is given in Box 2.2.
This effect can be interpreted as an “externality”. An externality refers to an adverse or positive
effect of one firm’s action on another firm. If two producers are in a complementary relationship,
it consists in a firm’s gain from a price reduction by a rival. If a firm in a vertical relationship
reduces its price, it has also a positive effect on other parts of the supply chain by increasing
their sales. This “link” between firms in a vertical relationship is described in the economic
literature and throughout this report as the “vertical externality”. The externality effect is not
restricted to prices. Improvements in, for example, quality, promotion or distribution will also
have a positive effect on other firms in the supply chain. Non-price externalities raise significant
complications in attempts to apply the “net price test” that is put forward in the competitive
assessment of horizontal mergers.
This externality effect has fundamental implications for the assessment of cost efficiencies.
Because a non-horizontal merger is likely to lead to a price reduction, it is clear that an
assessment of efficiencies which may give rise to further price decreases is in most cases
unnecessary. However, although these price reductions are fundamentally pro-competitive in
the short term, they may, under certain circumstances, raise anti-competitive concerns in the
longer term. It is to a discussion of the potential for non-horizontal mergers to harm competition
that we now turn.
23
European Commission (2000).
Page 9
Box 2.2 : A simple example of a merger between complementary products
The incentive that firms producing complementary products have to decrease prices
following a merger can be illustrated by reference to the following simple example. Suppose
that in an industry there are only two firms each producing one variety of a complementary
commodity. Demand functions are
q1 = a − p1 − dp 2
q 2 = a − p 2 − dp1
where a >0, pi is firm i’s price, qi is firm i’s output and d>0 indicates to what extent the two
products are complementary. If d=0, products are unrelated while the higher the value of d,
the higher the degree of complementarity between the two products. Both firms produce at a
marginal cost equal to c and face no fixed costs. This implies that each firm’s profit can be
simply written as follows:
Π i = ( pi − c )(a − pi − dp j ) i=1, 2.
If the two firms compete in price à la Bertrand, then equilibrium prices are:
p1 = p2 =
a+c
2+d
Following a merger between the two firms, the new entity will maximise joint profits, resulting
in the following new price:
p1 = p 2 =
a + (1 + d )c
2(1 + d )
It is straightforward to verify that the prices of both products decrease after the merger
whatever the degree of complementarity between the two products. For instance, if a =7,
c=3 and d=0.5, then post merger the equilibrium price would decrease from p1 = p 2 = 4
to p1 = p 2 = 3.8 .
Suppose now that the merger brings about a reduction in the marginal cost of the postmerger entity. In particular, we assume that after the merger the marginal cost decreases
from c=3 to c=0. This implies that post merger the new equilibrium price will be:
p1 = p2 =
a+c
7+0
=
= 2.3 .
2(1 + d ) 2(1 + 0.5)
In other words, due to the reduction in the marginal cost of production, after the merger
prices would decrease even further.
Page 10
2.2.3
Which efficiency gains benefit consumers?
The relevant benchmark in assessing efficiency claims is that consumers will not be worse off
as a result of the merger.24 In the Horizontal Merger Guidelines, the Commission states that,
“[f]or that purpose, efficiencies should be substantial and timely, and should, in principle, benefit
consumers in those relevant markets where it is otherwise likely that competition concerns
would occur.”25 This means that cost reductions which merely result from anti-competitive
reductions in output cannot be considered as efficiencies benefiting consumers.
In order to assess whether efficiencies can be expected to benefit consumers, a distinction
needs to be made between efficiencies that reduce variable costs and those that affect fixed
costs. From an economic perspective, a reduction in variable (or marginal) costs is most likely to
benefit customers in the short term as this will have a downward effect on price.26 Even a
monopolist would react to a reduction in marginal cost by reducing price. In contrast, fixed-cost
savings do not affect marginal production costs and are therefore unlikely to lead to lower prices
that directly benefit consumers in the short term.
The Commission (also following section 4 of the US Horizontal Merger Guidelines) clearly
acknowledges this in the Guidelines, in paragraph 80:
“… In line with the need to ascertain whether efficiencies will lead to a net
benefit to consumers, cost efficiencies that lead to reductions in variable or
marginal costs are more likely to be relevant to the assessment of efficiencies
than reductions in fixed costs; the former are, in principle, more likely to result
in lower prices for consumers.”
Note, however, that whether an efficiency is classed as a reduction in variable cost or a fixed
cost saving depends on the time frame considered. For example, efficiencies resulting in
reductions in the cost of building new machinery do not affect the short run variable cost of units
produced on old machines. However, in the longer term, they affect the cost of expanding
production onto a new machine, and so may be thought of as variable costs over a longer time
horizon. In some cases, therefore, such cost reductions could also lower consumer prices. In
particular, for industries where short run marginal costs are close to zero, the proposition that
long run marginal costs are not taken into account in price setting does not appear to accord
with common sense. The more pragmatic, common sense view would be that reductions in
costs that are avoidable, or fluctuate with output in the medium term (say over 12 to 24 months),
do affect the price formation process and should therefore be taken into account.
24
For a discussion of welfare standards see, for example, Lyons (2002).
25
European Commission (2004a), paragraph 79. The US guidelines, by contrast, require that the agencies will also
consider efficiencies in other markets that are “inextricably linked” to the effects in the relevant market. This implies that,
in the US, mergers which deliver significant efficiencies in large markets, but have potential anti-competitive effects in a
smaller market (in which the consumers are broadly the same as the first), can nevertheless be cleared (see Alfter,
Bishop and Mahr, 2004).
26
Although strictly speaking not identical, average variable costs are in practice often used as an acceptable
approximation of marginal costs. For the purposes of this report it is assumed that variable and marginal costs are
indeed a good approximation of marginal costs. Both are therefore treated as equivalent.
Page 11
Efficiency benefits are not restricted to savings in production costs. The Commission’s
Guidelines state that efficiencies that lead to new or improved products may also directly benefit
consumers. Indeed, paragraph 81 of the EC Merger Guidelines states:
“A joint venture company set up in order to develop a new product may bring
about the type of efficiencies that the Commission can take into account.”
Such benefits are, by their nature, passed on to consumers, although the benefit may be offset
to some extent by increases in price.
2.3
Are Non-Horizontal Mergers Always Pro-Competitive?
2.3.1
Potential anti-competitive concerns in non-horizontal mergers
As discussed above, given that non-horizontal mergers bring together complements or
unrelated products, rather than substitutes, such mergers are unlikely to have a direct adverse
impact on price. However, it is possible that such mergers might give rise to anti-competitive
outcomes, for example if they permit post-merger behaviour that is able to exclude rivals and
thereby reduces the overall effectiveness of competition.
The main potential anti-competitive concern regarding conglomerate mergers is that by tying the
sales of the products together, a firm enjoying significant market power in one market (the
“tying” market) might be able to leverage this market power into another market (the “tied”
market).
In the case of vertical integration, it is sometimes claimed that the merged firm will deny its
horizontal competitors access to the vertically related good. Such “foreclosure” is closely
analogous to a "tie" between two complementary products (e.g. manufacture and distribution).
In both cases, by excluding its rivals in this way, the intensity of competition may be reduced. In
a less extreme case, firms may allow access to the related goods, but on terms which
marginalise its horizontal competitors. For example, the integrated firm may charge high prices
for access to its distribution network, or produce upstream components which are less
compatible with the products of rivals in the downstream market.
The concept of foreclosure was used during the 1950s in the US to prohibit a number of nonhorizontal mergers such as Brown Shoe/Kinney and General Motors/DuPont. The arguments
put forward were based on an intuitive notion that extending control over the supply chain would
enhance market power. These simplistic arguments were later refuted by academics of the
“Chicago School” line of thought. The success of these more formalised economic theories is
confirmed by recent works, such as Farrell and Weiser’s (2003) concept of “ICE”. These
theories are explained in Section 2.3.2 below. Nonetheless, more sophisticated “Post-Chicago”
theories of foreclosure have been put forward which show that exclusionary practices can
Page 12
indeed succeed in certain specific circumstances. These arguments are expanded upon in
Section 2.3.3.
Foreclosure (or marginalisation) is the primary concern regarding anti-competitive nonhorizontal mergers, both in theory and in practice. In addition, however, it has been suggested
that non-horizontal mergers may, in some cases, have other anti-competitive effects. In
particular, they may increase the chance of collusion or reduce incentives to engage in R&D
because of fears of “spillovers”.
Several mechanisms for the increased chance of collusion have been put forward:
•
Increasing visibility: In order for collusion to be successful, firms have to be able to
monitor each other’s behaviour. In general, the prices of intermediate goods tend to
be less visible than retail prices, and this will necessarily undermine attempts at
tacit or explicit collusion. However, if firms are vertically integrated, then they can
collude over the retail price, which tends to be more visible. Attempts at engaging in
collusion may therefore be more stable in a vertically integrated environment than in
a non-integrated environment.27 However, it is not clear how the integration of a
single firm affects the ability to collude. If collusion is successful in a vertically
separated upstream market, then the vertical integration of a firm may destabilise
arrangements. This is because transfer price will generally be hard to observe (or
infer), and the integrating firm can therefore undercut its non-integrated competitors
by charging a lower price to its downstream divisions.28
•
Allowing information exchange: Vertical mergers may also facilitate the monitoring
of collusive agreements if integrated firms continue to supply other downstream
firms. If this allows the exchange of pricing and other competitively sensitive
information between competing input suppliers, it may facilitate collusion. However,
it is not clear how far such information will be reliable.29
•
Eliminating a disruptive buyer: Where one buyer is much larger than the others in
the market, firms may be tempted to deviate from an agreement to secure the high
one-shot profit from the large buyer’s order. It has been suggested that, in certain
circumstances, vertical integration with the buyer may overcome such problems.30
•
Multi-market contact: Non-horizontal mergers may result in the same firms
competing in a number of different markets. Such multi-market contact can facilitate
collusion as agreements can be sustained in a market where punishment is not
possible, because it is possible for punishment to occur in another market.
However, such strategies can only be successful under very special circumstances.
In particular, it usually requires a larger market where collusion can be easily
27
See, for example, US Non-Horizontal Guidelines, originally issued as part of "US Department of Justice Merger
Guidelines", June 14, 1984.
28
See Lipczynski and Wilson (2001), page 82.
29
See Riordan and Salop (1995).
30
See Lipczynski and Wilson (2001), page 82.
Page 13
sustained, and a much smaller market, where conditions are less conducive to
collusion. In this case, collusion may be sustained in the smaller market, despite
there being no punishment mechanism, because of fears that collusion will break
down in the larger market.31
The impact of non-horizontal integration on collusion seems rather ambiguous. Further, many
arguments are mainly intuitive, and lack economic rigour. Indeed some theories seem to run
directly counter to each other (for example, by eliminating a disruptive buyer, the vertically
integrated firm can also undercut its rivals by charging a lower internal transfer price).
The fears regarding “spillover” effects concern the possibility for non-horizontal mergers to
reduce consumer welfare by increasing the potential for confidential information to be leaked.32
The idea is that non-integrated firms have to disclose confidential information in order to
cooperate with producers of complementary or vertically related goods. If these partner firms
are integrated, they may pass this information on to subsidiaries which compete with the nonintegrated firm. This may deter non-integrated firms from cooperating with the integrated body,
and can result in a reduction in R&D from the part of non-integrated firms.33 However, economic
theory has shown that integrated firms may not have the incentive to follow such strategies.
Further, it has shown that consumers are not necessarily harmed as a result of this type of
behaviour, for example because the integrated firm may tend to increase its R&D investment.34
2.3.2
The Chicago School and
presumption of innocence
“Internalizing
Complementary
Efficiencies”:
a
The idea that exclusionary behaviour, such as tying or input foreclosure, can be used to
increase market power by “leveraging” it from one market to another came under close scrutiny
during the 1970’s by “Chicago School” economists. These authors (see e.g. Bork, 1978, and
Posner, 1976) demonstrated that economic theory does not support such naïve “leveraging”
mechanisms. They showed that, in a number of cases, tying and vertical integration cannot
have anti-competitive effects. When there is a monopoly in the tying (or input) good, a
competitive market in the tied (or output) good and tying and tied goods (or inputs) are used in
fixed proportions, tying and vertical integration cannot have anti-competitive effects.35 They
argued that "there is only one monopoly profit" to be earned and that the monopolist can gain
this profit purely through its pricing in the monopoly (tying good) market (see Section 3.2.1).36
Having done so, imposing restrictions on consumer choice through bundling or tying will, unless
Page 14
31
Bernheim and Whinston (1990). However, also see Bishop and Walker (2002) and Bishop and Lofaro (2004) for a
counterview on this issue.
32
See, for example, Milliou (2004).
33
Competitive harm may also result from the leaking of confidential information regarding the bids made by competing
firms, their demand for inputs and other commercial information. There concerns have been raised in the recent EC
merger case M.3440 EDP/ENI/GDP (4064).
34
See Section 3.3.3.
35
A further necessary condition is that the tied goods have no secondary uses, i.e., they are useless if not combined
together to form a “composite” good.
36
In short, under the circumstances described, there is only one monopoly profit because there is effectively only one
(composite) product!
there are strong efficiency reasons for doing so, simply harm the firm's commercial interests.
Hence, if firms wish to engage in non-horizontal mergers, it will be for pro-competitive reasons.37
As noted by several authors (see e.g. Farrell and Weiser, 2003), this argument can be taken
one step further. The crucial point is that, because of the incentive structure resulting from the
vertical externality discussed in Section 2.2.2, a monopolist will generally have the economic
incentive to actively promote the most efficient industry organization at other levels of the supply
chain. That is, the objectives of the monopolist and the consumers are aligned. In many cases,
realising an efficient organisation of the supply chain will require vertical (conglomerate)
integration. However, integration may not always be appropriate, as competition may foster
efficiency better than common ownership. Therefore, in some cases, the monopolist may prefer
to operate in a disintegrated structure. For example, this may explain “outsourcing” decisions.
Further, Farrell and Weiser argue that even if the monopolist does choose to integrate into other
markets that are vertically related or complementary, it will still welcome value-added innovation
by its competitors in these industries.
For example, Palm first launched its personal digital assistant as an integrated system including
proprietary hardware and operating system, presumably recognising that there were benefits of
such vertical integration. As the industry evolved, however, the company moved away from
such integration. In response, Palm introduced an “open licensing” system, and separated its
hardware and operating system divisions into two independent companies.38 In this way, Palm
expected to benefit from the contribution of the highly innovative independent firms typical of the
computer hardware industry; in other words, effective competition in one market (hardware) was
expected to bring higher profits to the monopolist of the complementary product (the operating
system).39
Farrell and Weiser (2003) refer to this principle as “Internalizing Complementary Efficiencies” or
“ICE”. The concept does not rely on a specific theoretical model, but only depends on the
complementary or vertical relationship between the two goods. The message of ICE is that
antitrust authorities should, in general, trust the market to provide the most efficient organization
for the industry, and refrain from intervening in firms’ decisions on vertical or complementary
integration, mirroring the earlier policy conclusions of the Chicago School.
2.3.3
Post-Chicago: exceptions to the rule
Although the ICE principle is very general, in certain specific cases the incentive to promote
efficiency may be counterbalanced by other considerations, and the outcome may not be
socially efficient. Such exceptions encompass many of the “post-Chicago” theories, which
counter the sweeping predictions of the Chicago School arguments, claiming they are
dependent on the tied market being competitive, or the monopolist being able to extract the full
37
The principle of the “one monopoly profit” is described in detail in section 3.2.1.
38
Fried and Kawamoto (2002)
39
The video games industry provides another example. Each of the three largest suppliers of hardware consoles –
Microsoft, Sony and Nintendo – all license independent third party developers and publishers of video games. See
Mock (2005) for an example in mobile telephone technology.
Page 15
monopoly rent. In this final subsection, we discuss a number of conditions under which ICE
(and, a fortiori, the ”one monopoly profit” principle) does not hold. When these exceptions apply,
the industry structure that emerges through market mechanisms may be inefficient, and
therefore potentially, but not certainly, raising the prospect that intervention on the part of
competition authorities would be appropriate.
In consequence, when assessing real-world examples of non-horizontal mergers, these
exceptions should not be used to motivate a presumption of guilt as soon as market conditions
meet some of the stated conditions. The fact that the ICE principle may not hold does not
automatically imply that firms will behave anti-competitively. Rather, it means that authorities
should take a closer look at the case to assess, in addition to likely efficiencies, the potential
anti-competitive concerns. Indeed, it is one thing to develop a mathematical model that shows
exclusion could happen, and quite another to use that model to derive real life predictions that
are sufficiently robust to justify a decision to prohibit a merger. To make the transition from the
theory to a decision to prohibit a merger, the exclusionary effects theories need to be tested
against the facts of the industry.
The exceptions to the ICE principle that can be found in the economic literature are briefly
explained in the following. These include the cases when the products have a secondary use,
the possible incentives to restore monopoly power or to achieve price discrimination, the aim of
protecting the monopoly against entry, strategic considerations with respect to possible future
regulation, and situations where the managers fail to implement ICE, either because of
incompetence, or because their incentives are not aligned with the firms’ interests.
Secondary use
Winston’s (1990) classic article points out that a monopolist can profitably enter a
complementary market if the complement has a positive value without the monopolized
good. In such cases, a firm which extends its monopoly power to the complementary
market can enjoy a profit that is higher than the one-monopoly profit. A similar
mechanism holds in a vertical setting. For example, a hypothetical monopolist of steel
manufacturing can profitably enter the iron extraction industry, because iron has other
uses than making steel. Clearly, while this provides an incentive to extend market
power, it does not in itself suffice to show that leveraging would be possible.
Restoring monopoly power
The logic of Internalizing Complementary Efficiencies (ICE) can fail if the monopolist is
unable to extract the full monopoly rent. This may happen because of price regulation,40
contractual problems (namely, some inability to commit, mainly related to the degree of
downstream competition, as discussed below), or other reasons. In such cases, the
incentive to restore monopoly power may be stronger than the ICE incentive to promote
40
Page 16
A general discussion of this issue can be found in Farrell and Weiser (2003). Recent surveys of the access-price
literature include: Laffont and Tirole (1999), and Armstrong and Sappington (2003).
efficiency: the monopolist would be willing to tolerate a certain degree of inefficiency in
order to recover the full monopoly rent. More generally, a monopolist’s objective and the
authority’s objective of increasing consumer welfare are no longer aligned. One
possible strategy for the monopolist to extract a higher surplus from consumers is to
attempt to extend its market power to the complementary market, and/or to engage in
foreclosing activities.
A specific instance of monopolists being unable to extract full monopoly profits, put
forward by Rey and Tirole (2003)41, relates to a firm’s “inability to commit”. Their results
are based on the hypothesis that the monopolised good is not sold on the marketplace,
but through private contracts. As a result, there is not a single unit price for all buyers
and all quantities. Instead, prices and quantities are negotiated for each transaction.
The authors claim that this is also the central failure of the “one-monopoly profit” line of
reasoning because the inability to commit is directly related to the degree of competition
in the downstream market, which in turn is central for the ICE logic.
Their argument can be explained with a simple example. Consider the case where the
monopolist is the patent-holder of a unique technology.42 The monopolist licences the
use of the patent, for a fixed fee, to downstream competitive firms. In the first period, it
would want to charge a high fee to a limited number of firms in order to extract the full
monopoly profit. But at a later time (ex-post), it has an incentive to exploit the residual
demand by selling other licences at a lower fee. In the absence of a credible
commitment not to do so, the first licensors only agree to pay a smaller fee, and the
patent-holder cannot extract the full monopoly profit.43 Economists refer to such effects
as a “commitment problem”. The authors claim that “the commitment problem is very
general and extends to situations in which downstream firms purchase variable
amounts of the monopolised input”. More precisely, the authors suggest the following
criteria.44
•
The loss of monopoly profit is larger the higher the intensity of competition in the
downstream competitive sector. Therefore, the monopolist’s profit is smaller the
larger the number of downstream firms and, for a given number of firms, it is higher
when the final products are more differentiated. Thus anti-competitive vertical
integration may be more likely where the downstream industry is more competitive.
•
It matters whether the monopoly is upstream or downstream, since a downstream
monopoly does not face any commitment problem. Thus, in general, anticompetitive integration can only avoid commitment problems when the upstream
firm is a monopoly.
41
See also Hart and Tirole (1990), and McAfee and Schwartz (1994).
42
An identical example could be given where the monopolist is the franchisor of a unique brand.
43
A similar problem is faced by any monopolist who sells a durable good. Consumers with a moderate preference for this
good anticipate that, after initial sales at a high price, the monopolist has an incentive to satisfy a part of the residual
demand at a lower price. Therefore, they wait for the price to fall, in the same way as licensees in the previous
example. Consequently, the demand – and, thus the price – is initially lower than in the case of a non-durable good.
44
Hart and Tirole (1990).
Page 17
Price discrimination/metering
Although price discrimination generally allows firms to extract a higher surplus from their
customers, it is not necessarily anti-competitive since generally price discrimination
increases output, and assists in enlarging the customer base. However, the payoff to
the firm of such practices can be high enough to induce the monopolist to integrate,
even when this is a socially inefficient structure. Thus, as Farrell and Weiser point out, it
constitutes an exception to ICE. A more extensive discussion of price discrimination and
metering as a potentially efficient motivation for vertical and conglomerate integration is
provided in sections 3 and 4.
Protecting the monopoly against entry
There is a relatively large literature on foreclosure in the complementary or vertically
related market as a means of protecting against entry in the monopolized market.45
First, if the monopolist dominates the two complementary markets, a potential entrant
must simultaneously enter both markets in order to compete. This generally raises the
costs of entry and reduces the probability of success.46 Even if the two goods are
valuable independently, the incumbent can use bundling to lower the potential profits of
the (single-component) entrant without having to price low itself.47 Nalebuff (2003) cites
the case of Microsoft Office as a notable possible case of bundling to deter entry; the
Office package includes Word, Excel, PowerPoint and Outlook at a much lower price
than the sum of individual applications.
Second, if the monopolist dominates only one of the two complementary (or vertically
related) markets (say, market A), competition in the market for the complementary
product (market B) may, over time, facilitate competition in the monopolistic market.
Thus, from monopolist’s viewpoint, there is a clear incentive to marginalise competitors
in the complementary market B, in order to provide long term protection for its primary
market A. This motivation appeared in the judgments of US courts in the Microsoft case.
The reasoning goes as follows. Microsoft’s monopoly power in operating systems (OSs)
is based, at least in part, on the fact that application software designed for MS Windows
cannot be easily modified to run on alternative OSs. As a result, customers who want a
large choice of software would not purchase a hypothetical new operating system for
which few applications exist. This incompatibility problem constitutes a formidable
barrier to entry in the OS market. In the short-run, the Netscape web browser was just a
complementary product to MS Windows operating system. However, in the long-run,
Netscape technology could have allowed software manufacturers to produce
Page 18
45
We do not describe theories relevant to tying or bundling independent goods, such as in Section 1 of Whinston (1990).
46
Choi and Stefanadis (2001) show that when entry is risky, tying reduces the expected returns of entry in each market,
since only successful entry in both would be profitable; Carlton and Waldman (2002) analyse a situation where entry in
both markets gives rise to economies of scope.
47
Nalebuff (1999).
applications that can run on multiple OSs.48 Since it could lower entry barriers in the OS
market, Netscape posed a threat to MS Windows monopoly in the long term. The courts
suggested that, for this reason, Microsoft undertook a campaign to undermine
Netscape’s browser.49
Managerial incentives and managerial incompetence
Farrell and Weiser also suggest that a firm may fail to follow ICE because the benefits
to some of the managers do not coincide with the benefits to the firm.50 When managers
derive private benefit from their actions, they will not always seek the most efficient
outcome for the firm. A stream of economic and financial literature has analysed these
issues.51 Alternatively, a firm may not behave according to ICE because its managers
fail to appreciate its benefits.
2.4
Summary
This section has considered the fundamental nature of non-horizontal mergers. It first described
the categorisation of concentrations as horizontal, vertical or conglomerate. Conglomerate
mergers can be further divided into those involving complementary and unrelated goods. There
is a strong commonality between vertical mergers and mergers involving producers of
complementary goods.
It then highlighted the fundamental difference between horizontal and non-horizontal mergers.
In particular, unlike horizontal mergers, vertical and conglomerate concentrations do not remove
a direct competitive constraint. This is because the firms in a non-horizontal relationship
produce complementary (or unrelated), rather than substitute, products. This means that, in
strong contrast to horizontal firms, each party would generally like the other to lower its price.
Thus, there is a post-merger incentive to reduce, rather than increase, prices.
Nevertheless, it is possible under certain specific circumstances for non-horizontal mergers to
give rise to anti-competitive effects through exclusionary behaviour. However, although such
cases of foreclosure cannot be dismissed out of hand and a closer look may be advisable, the
very specific conditions on which they are based suggest a particular caution in their application
to justify a negative decision. Such a transition from exclusionary effects theories to a decision
to block a merger would need to be tested against the facts of the industry.
48
Netscape offered software manufacturers a set of middleware application programming interfaces (APIs) alternative to
Windows’ APIs.
49
As found by the district court and confirmed by the court of appeals. See United States v. Microsoft Corp., 87 F.Supp.2d
30,38 (D.D.C. 2000) and 253 F.3d 34 (D.C.Cir 2001); Bresnahan (2001); see also Farrell and Weiser (2003) and the
references therein.
50
Farrell and Weiser (2003).
51
See, e.g., Milgrom and Roberts (1992).
Page 19
3 Pro-Competitive Motivations for Vertical Mergers
“If you want something done right, do it yourself!”
– A theory of vertical integration –
M. K. Perry, 1989
3.1
Introduction
This section considers the sources of efficiency gains that can arise from vertical integration. In
addition to cost and price reductions, these efficiency gains also include improvements in
quality, increased variety, and innovations leading to new products.
The potential sources of efficiency gains from vertical mergers result from a wide variety of
mechanisms and occur in a broad range of situations. As such, the potential for pro-competitive
motivations for vertical integration is very large. In this section we have grouped the sources of
potential gain into four broad categories, according to the nature of the economic mechanism:
•
Increasing pricing efficiency: In many cases, charging a uniform unit price will not
be an optimal way of organising a supply chain. However, more sophisticated
pricing mechanisms may be costly or impossible to introduce. In such cases,
vertical integration can lead to a more efficient outcome by removing the need for
external pricing.
•
Improvements in productive efficiency: Vertical integration may result in a more
efficient use of inputs and/or productive assets, thanks to economies of scope,
supply assurance, improved information flow and coordination, or better
management.
•
Preventing profit expropriation: In some cases, a firm may be unable to appropriate
the full benefits of its investments, as this is partly reaped by competitors or firms at
other levels of the supply chain. This can result in sub-optimal levels of investment.
By “internalising” such external benefits, non-horizontal mergers can lead to
significant efficiency gains.
•
Finally, market transactions may be subject to transaction costs and incomplete
contracts. Where contracts are not fully specified, a party may be able to exploit
contractual loopholes to the disadvantage of their vertical partners. To overcome
this problem, costly monitoring or incentive mechanisms may have to be used. In
certain situations, such as those that give rise to the so-called “hold-up” problem,
investment may be deterred altogether. At a very general level, all contractual
relationships are subject to transaction costs, such as those incurred in drawing up
the contract or searching for an appropriate partner. By minimising such contractual
problems, vertical integration may provide an efficiency gain.
Page 21
Figure 3-1 below summarises the efficiencies that are considered under each of these headings
in the remainder of this section. Note that these categories are principally intended for
organizational purposes; as will become clear in the discussion, several potential benefits of
vertical integration could appropriately be included under more than one heading. The
discussion of incomplete contracts and transaction costs, in particular, is relevant to the analysis
of all other efficiencies described in this section. This is because contractual agreements
between firms can sometimes be employed instead of mergers to realise the efficiencies
described in the first three categories described above. Such agreements, however, are not
always possible or, if they are, they may be less efficient solutions than full integration.
Therefore, what ultimately motivates a vertical merger is the fact that contractual alternatives
are inefficient. For this reason, the matter of incomplete contracts and transaction costs is
essential to the analysis of all other efficiencies described in this section.
Figure 3-1: Classification of Efficiencies
Increasing Pricing Efficiency
Increasing Productive Efficiency
Avoiding double marginalisation
Economies of scope
Avoiding input substitution
Supply assurance
Lowering rivals’ costs
Improved information flow and coordination
Price discrimination
Managerial efficiency
Preventing Profit Expropriation
Incomplete Contracts & Transaction Costs
Free riding on promotional activity
Incomplete contracts
Internalisation of R&D benefits: horizontal spillovers
The hold-up problem
Internalisation of R&D benefits: vertical spillovers
General transaction costs
In the following sections we consider each efficiency in turn, drawing heavily on the relevant
economic literature as well as recent antitrust cases. Section 3.2 deals with improvements to
pricing efficiency; Section 3.3 considers improvements to productive efficiency; Section 3.4
considers how benefits can arise through overcoming externalities; finally, Section 3.5 considers
how the inefficiencies arising from incomplete contracts and transaction costs can be resolved
through vertical integration.
Our discussion of the various forms of efficiencies is organised as follows. For each efficiency,
we first set out how a non-integrated environment can produce suboptimal results. We then
discuss the impact of vertical integration, showing how it can eliminate this inefficiency, and
consider the impact this has on firms and consumers. We also discuss the situations in which
the efficiency gain occurs, highlighting the factors which affect the size of the benefit, and
consider whether the results hold under more general scenarios. Finally, we discuss whether
each efficiency could be achieved without resorting to full vertical integration, and the factors
affecting the likelihood that the benefits would be passed on to consumers. Each subsection
presents, where relevant, the specific contractual agreements that can be employed to realise
the efficiency. However, such contractual alternatives should always be assessed in the general
framework of incomplete contracts presented in Section 3.5.
Page 22
3.2
Increasing Pricing Efficiency
This section considers inefficiencies in vertical relationships relating to the pricing mechanism. It
is organised as follows:
3.2.1
•
Section 3.2.1 looks at the problem of double marginalisation;
•
Section 3.2.2 discusses input substitution;
•
Section 3.2.3 shows how integration can lower (rather than raise) rivals’ costs;
•
Section 3.2.4 discusses price discrimination.
Avoiding double marginalisation
Source of inefficiency
Where a firm has some market power, it will price at a mark-up over the short-run marginal cost
of production. Mark-ups may also be needed in competitive industries where firms have to cover
significant fixed costs.52 Where two firms in a supply chain sell at a unit price that includes a
mark-up, the price of the final product is generally inefficiently high. This is a consequence of
the vertical externality described in Section 2.2.2, whereby each firm overlooks the impact of its
pricing decision on the profits of firms operating at different levels of the supply chain. This is
usually known as the ‘double mark-up’ or ‘double marginalisation’ effect. If either firm were to
reduce its mark-up, the price of the final good would decrease, leading to higher output and
increased consumer surplus. Total industry profits would also increase, but the firm’s own profits
would be reduced; for this reason firms lack the incentive to unilaterally reduce prices. A
coordinated reduction of the mark-ups of two vertically related firms, instead, would be profitable
for both. Thus, vertical integration can improve pricing efficiency by enhancing coordination
between the merging firms.
An analogous source of inefficiency arises between suppliers of complementary goods or
services, since the externality present between such firms is identical to the vertical externality
described here.53 In the context of complements this effect is known as “Cournot effect”, and is
described in Section 4.2.1.
52
If prices were driven down to marginal cost, market players would be unable to cover their fixed costs, and
consequently would be forced, in the long term, to exit the market.
53
For example, retailing can be considered as a service complementary to manufacturing. Also, the distinction between
vertical integration and the integration of complements followed by pure bundling or tying disappears post merger: in
both cases the merged firm ends up supplying a finished system.
Page 23
Effect of vertical integration
Figure 3-2 below provides a graphical representation of a simple market structure that could
give rise to the double marginalisation problem. It shows an upstream monopolist (e.g. a
manufacturer) that produces at a marginal cost of production equal to cu and a downstream
monopolist (e.g. a distributor) which faces a marginal cost of cd, over and above the cost of
purchasing the manufacturers product, for which it pays a price pu.54 This price is higher than
the upstream firm’s marginal cost of production; i.e., the manufacturer adds a mark-up, Mu. The
monopolistic distributor also adds its own mark up, Md, to the price that it charges to the end
consumers, pd.
Figure 3-2: Upstream and downstream monopolists
Manufacturer
Cost = Cu
Pu = Cu + Mu
Distributor
Cost = Cd
Pd = Pu + Cd + Md
Consumers
Both the manufacturer and the distributor would like the other to reduce its mark-up. If the
distributor reduced Md, the price to consumers, pd, would fall and sales would increase. Thus,
for the same intermediate price, pu, the manufacturer would sell more of its good and its profits
increase. Similarly, if the manufacturer decreased its mark-up, Mu, the distributor would obtain
its inputs more cheaply, increasing its own profits, and part of this cost reduction would also be
passed through, lowering the price of the final good. However, when the firms are vertically
separated, neither firm takes account of this beneficial effect of a price decrease on the other
firm. In contrast, a single vertically integrated firm (as shown in Figure 3-3) would consider these
effects in its decisions. This gives it an incentive to reduce the mark-ups. In consequence, the
mark-up, M, that the vertically integrated firm charges is lower than the aggregate mark-up (Mu
+ Md) charged by the vertically separated firm. The final consumer therefore faces a lower price
than in the vertically separated environment. Appendix A and Box 3.1 present two simple
theoretical models of this effect.
54
Page 24
Assume that each unit of the final good requires the input of one unit of the intermediate good.
Figure 3-3: Vertical integration between upstream and downstream sectors
Manufacturer
Cost = Cu
Distributor
Cost = Cd
P = Cu + Cd + M
Consumers
Under what conditions the inefficiency occurs
Double marginalisation can arise whenever there are mark-ups at both stages of the supply
chain. However, the problem becomes less serious, the lower the mark-ups applied. In the
limiting case where there is perfect competition and no fixed costs at any level of the supply
chain, that firm can impose no mark-up. Thus there is only one mark-up charged (that of the
vertically integrated firm), and vertical integration confers no further benefits.
This result is illustrated graphically in Figure 3-4. This figure shows the results of a simple
economic model in which there are an upstream monopolist and a number of downstream firms
which compete in output (à la Cournot). The details of this model are set out in Appendix A. The
vertical axis shows the increase in output which arises when an upstream monopolist acquires
all downstream firms, thus eliminating the double mark-up problem. On the horizontal axis, the
number of firms in the downstream sector is varied, to reflect differing levels of competition
(more precisely, the higher the number of firms, the higher the intensity of competition). The
chart shows that as competition in the downstream market becomes more intense, the benefits
from vertical integration decrease. This implies that a single, vertically integrated monopoly is no
more and no less efficient than an industry with an upstream monopoly and perfect competition
downstream: price and output would be exactly the same.
Page 25
Figure
3-4:
Efficiency
gain
from
elimination
of
double
marginalisation
12.0
Additional Output from Vertical Integration
10.0
8.0
6.0
4.0
2.0
0.0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
Number of firms
In addition to demonstrating how double marginalisation arises with imperfect competition,
Figure 3-4 also explains the Chicago School's principle of the “one monopoly profit” already
described in Section 2.3.2. This implies that prices and output with a single, vertically integrated
monopoly would be exactly the same as in an industry with an upstream monopoly and perfect
competition downstream or perfect competition upstream and monopoly downstream.
Therefore, the monopolist’s profit is also the same. In other words, if a monopolist is supplying a
competitive downstream market, it cannot gain any additional profits by vertically integrating
with the downstream firms and selling at the monopoly price to the final consumer. Therefore, a
monopolist who seeks to integrate vertically with a firm that operates in a perfectly competitive
market cannot be motivated by the anti-competitive objective of foreclosing competitors.
Instead, its motivation is very likely to be a pro-competitive efficiency gain. Indeed, if the merger
conferred other kinds of efficiencies (e.g. in production, transaction costs, etc.), consumers may
face lower prices and higher output.
Page 26
Box 3.1: A simple model of a multi-firm supply chain
The example below illustrates how the double marginalisation problem
tends to get worse the longer the supply chain and the smaller the
price elasticity of demand. This model complements the analysis for a
horizontally differentiated good conducted in Appendix A.
c
Firm n
We consider an industry where the supply chain is constituted of n
successive monopolies, as shown in the figure. This structure is a
generalisation of the one illustrated in Figure 3-2. We assume for all
firms that one unit of input produces one unit of output, and marginal
cost of production is only given by input cost. We assume that
pn
p3
Firm 2
p2
consumer demand shows constant elasticity ε ,
Firm 1
p
ε > 1:
D ( p ) = p −ε .
Consumers
The timing of the model and the solution method are as in Appendix A.
Output and prices can be computed as follows:
Firm 1:
q1 = p −ε ,
output
monopoly price p =
… Firm n:
ε
ε −1
p2 ,
−ε
q n = ( εε−1 ) ε ( n −1) p n ,
output
monopoly price p n =
ε
ε −1
c.
With an isoelastic demand, all firms charge the same mark-up over the cost of inputs. The
price to consumers is thus exponentially increasing in the number of firms n, and output is
exponentially decreasing:
p = ( εε−1 ) n c ,
q1 = q n = ( εε−1 ) εn c −ε .
To provide an example, the chart below illustrates the price to consumers as a function of
the number of firms in the supply chain, in the case ε = 2 and c = 1 . Note that the
marginal cost of production is the same for all numbers of firms, so that the different
prices are only a consequence of the different industry organisation. With two firms the
price to consumers is twice as large as with only one integrated firm (4 instead of 2, in the
example). Further, note that larger demand elasticity implies smaller mark-ups and,
therefore, also a smaller efficiency loss.
35
30
price to consumers
25
20
15
10
5
0
1
2
3
4
5
number of firms
Page 27
The basic models of double marginalisation, such as the one set out above, assume an
upstream monopolist, and downstream firms that produce identical goods. In that situation,
social and consumer welfare are unambiguously increased by the merger. However, imperfect
competition often exists both upstream and downstream, i.e., the industry is indeed
concentrated, but not monopolised at any level of the supply chain. In addition, most industries
are characterised to a greater or lesser extent by product differentiation. Although in such
industries the nature of the efficiency gain is the same as described above, the different
competitive interaction between rival firms may change firms’ incentives and the consequences
for competitors and consumers. A large amount of attention has recently been paid to modelling
such industry characteristics (e.g. Salinger, 1988; Gaudet and Long, 1996; Higgins, 1999). A
benchmark model is based on the duopoly industry structure illustrated in Figure 3-5. This
shows an integrated firm A1B1, which also buys input from a non-integrated upstream firm (A2)
and sells input to a non-integrated downstream firm (B2). Consumers can choose between four
differentiated composite goods (A1B1, A2B1, etc.) Thick/dashed lines indicate possible price
decreases/increases with respect to pre-merger prices (price changes in the input market are
not indicated).
Figure 3-5: Effect of vertical integration in a differentiated market
Post Merger
B1
B2
Consumers
A2B2
B2
A1B2
B1
A2B1
A2
A1B1
A1
A2B2
A2
A1B2
A1
A2B1
A1B1
Pre-Merger
Consumers
Note: Thick/dashed lines indicate possible post-merger reductions/increases in prices to final
consumers with respect to pre merger prices (see also Section 4.2.1).
Page 28
Very interesting results on vertical industries can also be derived from similar models describing
the structure of an industry which produces complements.55 (The details of these latter models
are presented in Section 4.2.1.) The basic lessons that one can derive from these papers are
the following:
•
First, vertical integration makes the pricing of the internally-produced product (i.e.,
A1B1 in Figure 3-5) more efficient, because it eliminates the double marginalisation
effect for the merged firm. This is clearly a pro-competitive effect since it tends to
increase social welfare. However, in differentiated industries, the pricing of the
mixed-components products (A1B2 and A2B1) may become less efficient, to the
extent that the firms may have the incentive to increase the price of these products,
thus making double marginalisation worse for these products (see below).56
•
Second, the merger also usually reduces the profitability of competitors. This is the
result of two different effects. On the one side, the merged firm may sell to final
consumers at a price lower than its competitors simply because it has become
more efficient than its competitors. This outcome is pro-competitive, unless rival
firms are marginalised and the market power of the integrated firm increases
significantly. On the other side, the merged firm may have an incentive to buy less
input from non-integrated rivals, and to sell inputs at a higher price. Such behaviour
may harm competitors by reducing their revenues (customer foreclosure) or by
increasing their costs (input foreclosure), although not all rivals are necessarily
harmed.57 We note, however, that such effects do not imply a strategic objective to
marginalise competitors, but are simply the result of the static profit maximising
decision in the new industry structure. The size of these effects and their overall
impact on consumer welfare are also discussed below.
•
Third, in many situations rival firms have the possibility and the incentive to respond
with a counter-merger. When this occurs, the final outcome is a single oligopolistic
market where all players are integrated. In such markets, all firms internalise the
price externality (at least for the goods that are entirely internally-produced), but the
benefits of the increased efficiency are “competed away” and passed on to
consumers. All firms end up making smaller profits than before the “merger wave”.
Several theoretical models illustrate this situation as a typical prisoner’s dilemma.58
However, it has been shown in the context of complementary-products mergers
that, in certain situations, rivals of the merging parties may find it more profitable to
55
Some of the relevant literature was at least partly motivated by the GE/Honeywell merger case; for example Nalebuff
and Lu (2001), Choi (2001 and 2003). In that case, Nalebuff and Lu (among other academics) worked for the merging
parties, while Choi worked for a complainant (Rolls Royce). See also the earlier works of Economides and Salop
(1992), Economides (1993), and Matutes and Regibau (1988).
56
However, the impact of this effect on social welfare is mitigated by the fact that output is likely to increase for the
efficiently priced goods, while it decreases for products that are priced less efficiently. The relative magnitude of these
effects also depends on the degree of product differentiation.
57
For example, if there is input foreclosure, remaining input suppliers may increase their market power, and hence may
also be better off (at least in the short term). Analogously in the case of customer foreclosure for remaining buyers.
58
See for example Gaudet and Long (1996) and, in the context of complementary products mergers, Economides (1993)
and Matutes and Regibau (1988). However, note that Gaudet and Long find that non-integration can also be equilibrium
when there is a sufficiently large number of firms in the market.
Page 29
remain non-integrated (although they still make less profits than when all firms were
non-integrated). 59 These results are discussed in more detail in Section 4.2.1.
Impact of vertical integration on consumer welfare
The resulting prices and the overall impact on consumers critically depend on a number of
factors. For example, the following results hold, depending on the degree of product
differentiation, and whether or not the integrated firm continues to participate in the market for
intermediate goods:60
Page 30
•
Where intermediate and final goods are homogeneous, the price of the input may
rise or fall. This is for the following reason. Assuming that the vertically integrated
firm no longer sells input to, or buys input from, rival firms, the number of buyers
and sellers in the intermediate goods market is reduced. On the one hand, the
reduction in the number of upstream suppliers suggests that prices will rise; on the
other hand, demand to the independent suppliers falls, because of the reduction in
the number of buyers, and also because the remaining independent downstream
firms lose market shares versus the integrated firm. However, even if the input price
increases, the price of the final good may still decrease because of the downward
pressure on retail prices arising from the higher pricing efficiency achieved by the
integrated firm (see second point above).61
•
Where final goods are differentiated, and the integrated firm continues to participate
in the input market, the impact on consumers can be ambiguous, not only in terms
of whether consumers gain or not (as in the previous case), but also so that some
gain and others lose. This effect is well illustrated by Choi’s model described in
Section 4.2.1 (see Figure 4-4). Although the paper is concerned with conglomerate
mergers, its lessons are useful to understand vertical integration issues.62 We
describe these results with respect to the hypothetical industry structure shown in
Figure 3-5. Integration has a positive impact on consumers who, pre-merger, used
to buy the products of the merging parties (A1B1), as they can now enjoy a
discounted price. Consumers who used to buy from rivals (A2B2) are also betteroff. However, consumers who used to buy mixed versions (A1B2 and A2B1) are
worse off, because the price of these products is higher than pre merger.63 Thus the
effect on consumer welfare depends on their willingness to substitute between
products as well as on the relative importance of each group of consumers. More
generally, vertical (or complementary) integration may reduce the number of
59
Nalebuff (2000 and 2003)
60
Other important factors include the number of firms upstream and downstream, the possible existence of inferior
substitutes, and characteristics of demand such as own- and cross-price elasticity and market elasticity. The impact of
some of these factors is described qualitatively in the remainder of this section. However, the overall impact on prices of
all factors together is in general very difficult to assess, since it depends on the relative magnitude of opposing effects.
61
Salinger (1988). We note, however, that in Salinger’s model most vertical mergers are not profitable.
62
Some modelling differences may arise, for example because in a vertical setting the upstream firms may set the prices
of inputs before the downstream firms set the price for the finished goods. See also footnote 20.
63
These net effects on prices hold in equilibrium, and take account of the shifts in demand. Therefore, the magnitude of
the price effect crucially depends on the details of consumers’ demand.
product varieties available to consumers, or it may reduce the attractiveness of
some of those (because of the new price differential). On the other hand, the
merger also leads to lower prices for the remaining varieties. The total impact on
consumer welfare depends to a great extent on consumer preferences.
Moreover, the increased efficiency of the merged firm is likely to induce rival firms to increase
their own competitiveness, thus generating further benefits for consumers. For example, if
competitors respond with a counter-merger, this may lead to the additional customer benefits
illustrated in point three above.
Factors affecting the magnitude of the efficiency gain
Double-marginalisation occurs in a rather wide range of settings: it applies whenever there are
mark-ups at two stages of the supply chain, and includes situations when such mark-ups result
from market power conferred by a degree of product differentiation. The magnitude of the
efficiency loss, and thus of the possible gain depends primarily on the size of the mark-ups
charged on the goods. (The size of the gain is not affected by whether the cause of the high
mark-ups is needed to recover large fixed or sunk costs, or for another reason.)
However, despite the general validity of the economic mechanism, there are a number of other
factors that contribute to determine the size of the inefficiency. In some extreme cases, the
inefficiency does not arise in the first place, and a merger would bring no benefits:
Internal transfer price
In order to eliminate fully the double marginalisation effect, internal transactions within the
same company should be priced at marginal cost. Where this is the case, there is the
maximum efficiency gain. However, considerations other than pricing efficiency can lead
to internal transfer prices above marginal cost, resulting in partly inefficient outcomes.
These include the following:
•
Incentives to managers: Where managers are rewarded according to the profitability
of their division, they may have a private incentive to price above marginal cost in
order to extract some profit from another division.
•
Tax regulation: Where internal transactions are cross-border, tax regulation requires
firms to set internal transfer prices equal to market prices for the same good;
alternatively, multinational firms may tend to distort transfer prices in cross-border
transactions in order to avoid taxes by transferring profits to jurisdictions with lower
taxation levels.
•
Strategic objectives: When transfer prices are publicly observable, firms in
oligopolistic markets may set transfer prices above market prices in order to “induce
Page 31
their marketing managers to behave as softer competitors in the final product
market”.64
Note that these are all specific examples of the more general discussion on when a
monopolist’s choice regarding integration may be inefficient, presented in Section 2.3.3.
Price discrimination (between final consumers)
The models described above assumed that the same unit price is set for all final
customers. For most goods, this is no more than approximately true. Most consumables,
as well as durables, at the retail level, are offered with volume discounts, coupons, and
other kinds of price discrimination devices. Non-consumer products are often purchased
through tenders, implying prices may differ between customers, particularly where
suppliers have information on the customer’s valuation of the various bids.
Under perfect price discrimination, double marginalisation does not occur. In fact, a seller
who knows exactly the customer’s valuation for the good would charge a price equal to
that valuation, whether it is vertically integrated or not. Therefore, it is possible to tailor
prices to each customer, to match their willingness to pay exactly. Then, the pre-merger
price is not socially inefficient, and integration does not increase social welfare, nor does
it decrease consumer welfare.
Assume, for instance, that the buyer prefers to buy the product produced solely from the
two firms which propose to vertically integrate (i.e., A1B1 in Figure 3-5), and is willing to
pay a higher price for this than for products involving components of other firms. Then, in
a bidding competition where preferences are observable, the downstream firm would set
its price equal to competitors’ marginal costs plus a “preference premium” (and would
always win) regardless of whether or not it was integrated with the upstream firm. On the
other hand, the buyer may, instead, like only one of the components produced by the
integrated firm, preferring the rivals’ version for the second component (i.e., A1B2 or
A2B1 in Figure 3-5). In this case, if the vertically integrated firm were to deny its
competitors access to its individual components it would reduce its own profits. (Note that
this assumes consumer preferences are observable; the relevance of the pricing
externality under imperfect information, where perfect price discrimination is not possible,
is still an open question.)
Secondary use
The double marginalisation models discussed above assume that the intermediate good
has no secondary use. Relaxing this assumption can reduce the importance of the
‘double mark-up’ inefficiency. Consider the example of a copper producer supplying a
downstream brass monopolist. If brass was the only use of copper, then the double
marginalisation effect would go as described in the previous section. However, copper is
also employed for other purposes, such as producing electric wires and cables. Assuming
price discrimination is not possible, we conclude that the price of copper is determined by
64
Page 32
Gox (1998).
the total demand for copper, of which the demand of the brass manufacturer is only a part
(and, likely, a small part). Indeed, total demand may differ to a great extent from the
demand of the brass producer regarding elasticity, shape, volatility, etc. It is therefore
possible that the price charged to the brass manufacturer is already below the one
predicted by the models of the previous section, and may even be lower than the
“optimal” level. In that case, vertical integration between the copper and brass
manufacturers would not lead to lower prices for final consumers. Nonetheless,
integration may indeed enhance efficiency in the case where the price is suboptimally
high for the brass manufacturer. Further, gains obtained through the elimination of double
marginalisation would add to gains from price discrimination obtained by effectively
charging different prices to the brass manufacturing division and to downstream firms
using copper for other purposes (see the discussion in Section 3.2.4).65.
Complements do not have the same value
Results further depend on the relative values of input and output. In some cases there
may be a large difference in the value of the components; for example, a car may be
valued at thousands of euros, but the value of some of the furnishings provided by
upstream firms may be small relative to the total value of the car. A reduction of the markup on such components would have a negligible impact on the total price and, thus, is
unlikely to increase output. Nalebuff and Lu (2001) show (in a complementary product
setting) that in such conditions the pricing efficiencies from integration are minimal (and
also have little impact on competitors). However, if several component manufacturers
integrate vertically (and/or laterally), the total efficiency gain may become significant even
where the contribution from each component is negligible (see next point).
Small number of inputs / short supply chain
The double marginalisation inefficiency is more significant the larger the number of inputs
considered. If the final good is made of n components, each component’s price has an
impact on the sales of the other n-1. Thus, the size of the possible efficiency gain per
component is amplified and can become significant overall in cases where several
components are object to vertical integration (see previous point). In a purely vertical
setting, this amounts to vertical integration of n firms operating at different levels of the
supply chain (see Box 3.1). A more general example involves the integration of two or
more upstream producers of complementary inputs and a downstream integrator. In both
cases, the inefficiency increases at an increasing rate in the number of components.
However, if some of the components were sold as a bundle before the merger, these
should be considered as a single product, and integration is unlikely to bring further
benefits.66
65
Here, “price discrimination” refers to the possibility to sell copper at different prices to the brass producer and to cable
and wires manufacturers.
66
Since bundling is of greater importance for mergers between producers of unrelated or complementary goods, this
issue is discussed in detail in the following chapter (Section 4.2.1).
Page 33
Figure 3-6 below summarises the analysis presented in this subsection. It lists some
criteria that are useful to determine the likely importance of the double marginalisation
inefficiency – and thus the size of the potential efficiency gain from its elimination.
Figure 3-6: Criteria to assess the likely relative magnitude of the firm’s efficiency gain
Larger efficiency gain
Smaller efficiency gain
Pre merger: Large mark-ups at both stages of the supply chain
Pre merger: Small mark-ups at any stage of the supply chain
Post merger: Internal transfer price close to marginal cost
Post merger: Internal transfer price above marginal cost
One unit price for all customers
Sufficient degree of price discrimination
No secondary use
Important secondary use
Components have similar values
Components have very different values
Larger number of complementary inputs
Smaller number of complementary inputs
Can the efficiency be achieved through other means?
The double marginalisation problem can potentially be alleviated by a number of vertical
restraints, such as two-part tariffs, retail price maintenance (RPM), or quantity forcing. These
tools are indeed used in many industries. However, contractual arrangements should not be
considered to be as perfect and costless solutions to the double marginalisation problem. First,
real-world contracts can be relatively complex, inducing substantial transaction costs. Secondly,
while eliminating the pricing externality, such contractual solutions may generate other
problems, leading to a partly inefficient outcome.
We analyse below some of the most common vertical restraints that are used to eliminate the
double marginalisation inefficiency, and we review the limits of the applicability of each of
them.67 We also discuss which of the suggested contracts is most appropriate (if any),
depending on the details of the business environment.
In an idealised business environment where marginal costs are constant, there is no uncertainty
about demand or costs, all information is public and contracts are complete, optimal pricing can
be achieved for example through a two-part tariff, maximum RPM, or quantity forcing.68 A twopart tariff works as follows: the input is sold at a unit price equal to its marginal cost, thus
inducing the retailer to set the optimal price; the profit is transferred to the upstream firm through
a franchise fee. With RPM, either in the form of a definite price or of a price ceiling, the
upstream firm directly sets the optimal retail price, and can extract all profits from the
downstream firm by using an input unit price arbitrarily close to the retail price. Similarly, forcing
the downstream firm to sell a given quantity implicitly sets the downstream price.
Page 34
67
For a more detailed analysis of the economic theory of vertical restraints see e.g. Tirole (1988) and Tan (2001).
68
Where retailer shelf-space considerations are introduced, a multi-product manufacturer can eliminate double
marginalisation and capture strategic rents by using full-line forcing, brand discounts, aggregate rebates and maximum
RPM (Shaffer, 1991).
In more realistic business environments such contractual means may be unfeasible, or
inefficient. The main problems are generated by the presence of uncertainty, incomplete
contracts and asymmetric information. We describe below the main potential problems
addressed by the economic literature.
Uncertainty
Assume, for example, that consumer demand and/or retail costs are uncertain (i.e.,
subject to important and unpredictable fluctuations), and the retailer is risk averse. In this
case RPM or quantity forcing cannot achieve the optimal level of profits, since the
contract is written before the uncertainty is resolved. Under a two-part tariff as described
above, the retailer bears too much risk, and may be unwilling to accept the contract. A
better contract would in this case allow upstream and downstream firms to share the risk;
for example, a two-part tariff with unit price above marginal cost and reduced franchise
fee would achieve the goal, but then double marginalisation would not be fully eliminated.
Adverse selection
An adverse selection problem may arise when the downstream firm has more information
about final demand or about its own efficiency in selling the product.69 For example,
where demand for a certain good is uncertain, contracts between the manufacturer and
the retailers may be contingent on demand: e.g. in low-demand periods, the retailer pays
a lower franchise fee. Since a certain degree of demand uncertainty is present in most
markets, such contracts are very common in distribution. However, where the
manufacturer has poor information on actual demand faced by the retailer, this type of
contract is open to manipulation, since the retailer would have an incentive to underreport demand. This generates agency costs and is thus a source of inefficiency. In such
situations vertical integration may be more efficient (Gal-Or, 1999). Alternatively, the
upstream firm may sometimes be able to reduce the inefficiency through price
discrimination, by offering retailers a menu of contracts specifying different combinations
of price, quantity and/or quality (Tan, 2001).
Promotional activities
When promotional activity is important a problem of “moral hazard” may affect the
relationship between manufacturer and retailer. Moral hazard occurs when the behaviour
of parties cannot be easily verified, in which case a party may have an incentive to
deviate from any agreement made. Services such as pre-sale information, free delivery,
free alterations, credit, etc., are particularly prone to such problems as they are hard to
monitor and to specify in a contract. In particular, since such activities are costly for the
retailer it would tend to undersupply unless it received a specific incentive. In these
cases, the contractual agreement has to provide a solution to both the double
69
See, e.g. Gal-Or (1991, 1999), Tan (2001).
Page 35
marginalisation and the moral hazard problems. In a simple environment 70 a two-part
tariff or quantity forcing are generally sufficient to achieve the vertically integrated
optimum (maximum RPM, instead, is not suitable).71 However, when the manufacturer
also provides hard-to-verify services such as brand advertising and product quality, a
double moral hazard problem has to be addressed.72 Finally, promotional activity may be
undersupplied if there is a risk of free riding (see Section 3.4.1). We note here that while
double marginalisation would be solved by increased horizontal competition, free-riding
inefficiency would be made worse. For example, the former would call for a maximum
RPM, while the latter requires minimum RPM. Section 3.5 of this report discusses moral
hazard issues in more detail.
Commitment problem
When contracts are unobservable by third parties or can be privately renegotiated, a
monopolist willing to charge a fixed fee (for example in a two-part tariff contract) incurs a
commitment problem. As explained by Hart and Tirole (1990), once the monopolist has
sold a given number of contracts, it has the economic incentive to sell more contracts at a
lower price, thus diluting the value of all contracts. The risk of such opportunistic
behaviour reduces the monopoly profit ex-ante, unless specific mechanisms are devised
to eliminate the moral hazard problem.73 In practical circumstances, therefore, two-part
tariffs are likely to be affected by this commitment problem, and require additional
clauses. Possible solutions include: (a) exclusive dealing, ensuring that the monopolist
can sell only a determined number of contracts; (b) a market-wise minimum RPM,
together with a return option; (c) tariffs contingent on total output of all downstream
licensees; (d) reputation building (implicit exclusive dealing).74 Therefore, in these cases,
contractual solutions alternative to a merger would be complex (thus involving significant
transaction costs) and may include exclusionary clauses (raising anti-competitive
concerns). The problem is also solved if the monopolist integrates vertically and
forecloses independent firms from the input market. In the theoretical models analysed by
the authors, the presence of this commitment problem prevents the firms from extracting
the full monopoly rent; hence, integration generally results in higher prices, and lower
social and consumer welfare.75
The assessment of whether a merger is the most efficient solution to the double marginalisation
problem or, instead, whether a contractual alternative that would be viable and equally efficient,
should be performed on a case by case basis. In general, some contractual alternative is often
(but not always) available. However, depending on the particular circumstances and business
Page 36
70
I.e., one that is subject to moral hazard but not to additional complications such as uncertainty or “double-sided” moral
hazard.
71
Tirole (1988). However, in the presence of uncertainty and retailer risk aversion, optimal contracts need to trade-off the
incentive to behave accordingly with an appropriate level of risk sharing.
72
See, e.g. Romano (1994).
73
This economic mechanism is also described in Section 2.3.3 (under the heading “Restoring monopoly power”).
74
Rey and Tirole (2003), and references therein.
75
Note that here the comparison is between two situations where double marginalisation is automatically eliminated: twopart tariff and vertical integration. However, we also note that if the non-integrated monopolist is not allowed the
exclusionary practices necessary to solve the commitment problem, it might find it more profitable to charge a uniform
unit price, even when this is socially less efficient because it results in double marginalisation. In that case, consumer
welfare would be lower than it would be if the exclusionary practices were permitted.
environment, as described above, such contracts may involve significant transaction costs.
These include the costs associated with writing the contracts, the costs of monitoring and
enforcing, agency costs, etc. (see Section 3.5). Transaction costs are also likely to be higher
where the parties intend to realise other efficiencies besides the elimination of double
marginalisation. Moreover, the practical implementation of these kinds of vertical restraints often
includes clauses that reduce competition (e.g. exclusive dealing).
Summarising, where a merger raises significant anti-competitive concerns, alternative
contractual solutions may not be feasible (if transaction costs are excessively high) or not less
anti-competitive (for instance because of some contractual clauses necessary for
implementation). Nonetheless, where a viable contractual solution has been identified, there
may still be some merger-specific efficiency gains. These would materialise as transaction costs
associated with such a solution (including agency costs), that are likely to be saved, at least in
part, when the two entities are brought under common control (see Section 3.5).
3.2.2
Avoiding input substitution
Source of inefficiency
Input substitution occurs when a mark-up applied by an upstream firm causes downstream firms
to switch to an alternative input which is less efficient for production but is sold under more
competitive conditions.76
For example a manufacturer of fabric may be able to substitute between cotton and polyester. If
both inputs were provided at marginal cost, it may be more cost effective to use polyester.
However, it may be that polyester is subject to a mark-up, for example because it is
manufactured by a monopolist,77 so that it is priced above the marginal costs of production,
while cotton is produced under competitive conditions and supplied at cost. In this case the
downstream manufacturer will tend to use less polyester and more cotton. This distortion of
production, termed “input substitution”, leads to an inefficient use of inputs; i.e., production is
“wasteful” as the same output could be produced at lower social cost.
Inefficient substitution can also occur when a downstream firm applies a mark-up. For example,
if a distributor applies a mark-up to its services, a manufacturer may be pushed to use a less
efficient distributor, or to find an alternative means of distributing its products.78
Figure 3-7 illustrates this effect. The axes represent the quantities of cotton and polyester inputs
required by the downstream firm. The chart shows the socially optimal production plan and the
“distorted”, sub-optimal production plan. Both plans correspond to the same output (i.e., the
76
Riordan and Salop (1995), McKenzie (1951), Vernon and Graham (1971).
77
Note that the term “monopolist” refers to the fact that there is only one polyester manufacturer; however, this does not
imply full monopoly power, because of the competitive constraint exerted by cotton.
78
However, this can be seen as a cost of distribution. See, e.g. Krattenmaker and Salop (1986); Salinger (1989).
Page 37
points belong to the same “isoquant” curve), but require different proportions of the two inputs:
the distorted plan employs more cotton and less polyester. The production plans are chosen in
order to maximise the (downstream) firm’s profits, given the prices of the inputs, the required
amount of output, and the available technology. These conditions are satisfied when the slope
of the isoquant curve is equal (in absolute value) to the ratio of input prices.79 When the ratio of
input prices is different from the ratio of the marginal costs of producing those inputs, the
production plan chosen by the downstream firm is not socially optimal.
Figure 3-7: Inefficient input substitution
q cotton
“Distorted” production plan
Optimal production plan
Isoquant
Slope= -Ppoly / Cco
Slope = -Cpoly / Cco
q polyester
Effect of vertical integration
Assume now the downstream manufacturer was to integrate with the upstream polyester
monopolist, as shown in Figure 3-8. The integrated firm will generally base its production
decisions on the marginal cost of production; if it is more cost effective to produce the fabric by
manufacturing more polyester rather than buying in cotton, it will do so. Given that cotton is
produced in a competitive environment, the fabric manufacturer essentially bases its input
decisions upon marginal costs, which results in efficient production. This increased efficiency
implies that the combined profits of the monopolist and the manufacturer are higher than in the
case of vertical separation, which is very much akin to the double marginalisation result
discussed in Section 3.2.1.
79
Page 38
I.e., the marginal rate of technical substitution is equal to the ratio of input prices. (This criterion only holds for interior
solutions.)
Figure 3-8:
Integration to eliminate input substitution
Pre-merger
Competitive
Firms
Monopolist
Po
l
ye
s
Post-merger
Co
te
r
n
tto
Manufacturer
Competitive
Firms
Monopolist
Po
lye
s
n
tto
Co
te
r
Manufacturer
Fabric
Fabric
Impact of vertical integration on consumer welfare
Although the elimination of input substitution always increases productive efficiency, since it
directly reduces the marginal cost of production, the impact on social and consumer welfare can
be ambiguous. The reason for this is that it can increase the market power of the upstream
“monopolist”, by effectively eliminating the competitive constraint exerted by the alternative input
(e.g. cotton). For example, assume that before the merger the fabric manufacturer operates
under perfect competition, while post merger the mixed-fabric industry is monopolised by the
vertically integrated firm. Also assume that the final good (mixed-fibre fabric) does not admit any
significant substitute, so that a monopolistic manufacturer would be able to extract the full
monopoly profit. In the vertically separated environment, the “formal” monopolist of polyester
manufacturing does not have full monopoly power on its product, because the downstream
competitive firms would partly switch to cotton if it tried to raise the price of polyester. On the
contrary, the vertically integrated firm, once it has monopolised the downstream market for
mixed-fibre fabric by marginalising other downstream competitors (e.g. through input
foreclosure), is in the position to exert full monopoly power on the downstream market. The
Chicago School principle that there is only “one monopoly profit” in this case does not hold,
because we dropped the crucial assumption of fixed proportions between input and output.80
We also note that, in this example, vertical integration (with foreclosure) can increase market
power even where the downstream (i.e., mixed-fibre fabric) product does not constitute a
relevant market but only forms one segment of a broader, differentiated product market (e.g.
fabric). In that case, the integrated firm would not achieve monopoly power, but may
nevertheless increase its market power.
Price to final consumers therefore can either increase or decrease, depending on the relative
magnitude of efficiency gain and the increase in market power. In the theoretical models, the
80
See Section 2.3.2.
Page 39
outcome is determined by the precise assumptions made on the production and demand
functions, the competitive environment, and the numerical value of the various parameters.81
Under what conditions the inefficiency occurs
For this basic model of input substitution to apply, firms must be able to use multiple inputs,
some of which are produced more competitively than others, in the sense that the products are
sold at different mark-ups. These inputs must be substitutable; where the inputs are used in
fixed proportions such as nuts and bolts, there is no opportunity for substitution between the
inputs, and consequently no efficiency loss. On the other hand, the inputs cannot be perfect
substitutes: where products are perfectly substitutable production cannot be sensitive to the
input mix. The basic result is also based on the assumption that goods are sold at a constant
unit price.
Can the efficiency be achieved through other means?
The input substitution inefficiency is similar to the double marginalisation problem. In both cases
the mark-up charged by one firm induces inefficiencies at other levels of the supply chain. As
with the double marginalisation problem, input substitution can be overcome by adopting more
complex pricing mechanisms, such as a franchise-fee, RPM, or quantity forcing discussed
above.82
In addition, the input substitution problem can also be overcome if the polyester manufacturer
ties all of its sales with sales of cotton (i.e., pure bundling), as depicted in Figure 3-9. It amounts
to forcing the downstream firm to buy both inputs from the monopolist. By doing this, the
monopolist can induce the manufacturer to use the inputs in the efficient proportions. A
complication is that where several inputs are substitutable, the upstream firm has to supply all of
them in order to eliminate the inefficiency. Further, if the downstream firm also charges a markup, additional instruments must be used in order to avoid double marginalisation (e.g. RPM).
Finally, pure bundling may not dissipate potential anti-competitive concerns: in situations where
vertical integration may increase market power, as discussed above, pure bundling would allow
the firm to achieve the same result.
Page 40
81
See, for example, Riordan and Salop (1995), Burstein (1960), Schmalensee (1973), Hay (1973), Warren-Boulton
(1974), Mallela and Nahata (1980), Westfield (1982), Waterson (1982) and Quirmbach (1986).
82
See Section 3.2.1.
Figure 3-9: Pure bundling (without integration) to avoid input substitution
Cotton
Monopolist
Competitive
Firms
r
te
es
l y + on
Po ott
C
Manufacturer
Fabric
3.2.3
Lowering rivals’ costs
One frequently cited concern regarding vertical mergers is that the merged firm may increase
the price of the input to its competitors in the downstream market, and thereby increase the
costs of these rivals (see Section 2.2.3). Although this can be a natural (and not necessarily
anti-competitive)83 outcome of some theoretical models of vertical integration, several authors
have shown that a vertical merger may also lead to lower prices in the input market. Indeed,
such considerations were part of the raising rivals’ costs (RRC) theory from the very beginning.
For example, Section 3.2.1 discussed some of the results obtained by Salinger (1988) on
homogeneous product markets. In this section, we present a more recent contribution by
McAfee (1999) on differentiated products industries.84
Source of inefficiency
The source of inefficiency described in this section is, again, double marginalisation. The
difference with the discussion presented in Section 3.2.1 is in the fact that instead of analysing
vertical integration as a means of increasing the efficiency of the merging firms, we discuss how
integration can reduce double marginalisation inefficiency of rival firms – those that remain
vertically non-integrated.
Effect of vertical integration
To appreciate why rivals’ costs might decrease post merger consider Figure 3-10. Two
upstream firms (U1 and U2) produce differentiated intermediate goods (X1 and X2). Two
83
See the effect of vertical integration in Section 3.2.1.
84
The model by Choi (2003) described in section 3.2.1 produced similar results: the prices of the final good decrease for
products supplied by the integrated firm as well as for those supplied by rivals. However, the price of “mix and match”
systems tends to increase. See also Reiffen and Vita (1995).
Page 41
downstream firms (D1 and D2) employ both products as inputs: proportions, however, can be
changed.
Figure 3-10: Lowering rivals costs
U2
U1
X1
X2
X1
X2
D2
D1
Consumers
For the sake of simplicity, first assume that downstream firms do not compete; namely, assume
that final goods form separate markets. When a downstream firm (i.e., D1) merges with a
supplier of one of the inputs (i.e., U1) previously sold at a mark-up over the competitive price, it
can eliminate this mark-up and thereby reduce its costs of using that input. It would therefore
tend to use more of the internal input and less of the externally supplied one. Other input
suppliers (i.e., U2), faced with a reduction in the demand for their products, may have to
respond by pricing more aggressively. This in turn tends to lower the price that the integrated
firm (U1D1) charges for its intermediate product (X1). As a result, non-integrated downstream
firms (such as D2) gain lower prices for all their inputs. The net effect of the merger is that the
double marginalisation problem is also reduced for the non-integrated firms.
This outcome – opposite to the prediction of the RRC mechanism – stems from the crucial
assumption that downstream firms do not compete with each other. Since D2 is not a rival of the
integrated firm, the latter has no reason to try and raise its costs. In fact, damaging the
profitability of a customer would only damage its own business.
The relevance of this result (which in itself is not particularly surprising) lies in the fact that it
extends also to situations where the products of the downstream firms are imperfect substitutes
and, hence, firms D1 and D2 are indeed competing with each other. However, in these cases,
other incentives are present, and it is no longer guaranteed that prices will fall.
If there is competition between D1 and D2 in the final good market, a second effect operates
against the one described above – the RRC effect. On one side input prices would tend to fall
because of the reduced demand faced by U2; on the other side, the integrated firm has an
incentive to increase its rivals’ costs by raising input prices. Depending on which effect
Page 42
dominates, input prices will rise or fall. In the theoretical models, the precise outcome is
determined by the numerical value of the various parameters. In general, we note that the
“lowering rivals’ costs” effect is more likely to dominate the lower the degree of substitution
between the final goods (i.e., the less intense the competition between downstream firms).
The main lesson of these models is that the impact of a vertical merger on competitors’ costs is
a non-trivial question. Although, under certain conditions regarding the structure of industry and
demand, a vertical merger may lead to an increase of input prices, the opposite result may hold
under other conditions. In particular, input prices tend to be lower when input buyers are not
close competitors to the merging parties.
Impact of vertical integration on consumer welfare
In general, a reduction of input costs is partly passed through to consumers in the form of lower
prices for the final goods. The degree to which this happens depends on the intensity of
competition in the downstream market: in a perfectly competitive environment, the additional
surplus would be fully passed on, while a monopolist would reap part of it and increase its
profits. In the model described by McAfee, the surplus gained by consumers and downstream
firm(s) corresponds to lower profits of the independent upstream firm (U2).
Although McAfee’s model describes an industry with duopolies both upstream and downstream,
the results generalise to the situation where the integrated firm has several rivals, which
compete in output among themselves (in so called Cournot competition). However, it is then
necessary to assume that the downstream division of the integrated firms accounts for a
significant proportion of input sales. In that case, a greater part of the reduction of input prices is
passed on to consumers.
Under what conditions the inefficiency occurs
The discussion shows that the inefficiency is large if the non-integrated downstream firms do not
compete with the integrated firm in the final good market. The results also extend to the case
where the firms compete weakly, i.e., the products are partly substitutable.
The mechanism described by McAfee also requires the inputs to be imperfect substitutes, used
in variable proportions. Further, the inefficiency is not reduced by a vertical merger if the
independent upstream firm is able to price discriminate, since in that case it would charge a
lower price to downstream division of the integrated firm, and a higher price to non-integrated
customers.
Page 43
3.2.4
Price discrimination
Source of the efficiency gain
A firm enjoying sufficient market power may increase its profits if it can charge different prices to
different customers (for the same good).85 In particular, it can increase the price to customers
that have a high valuation for the good, while still selling the product at a low price to those
consumers who value it less. However, in order for price discrimination to be feasible, two
conditions must be met.
First, if the firm is to sell at a higher price to some customers than to others, there must be
something that prevents the customers who benefit from the low price from reselling to those
customers that are charged a higher price. If such resale is possible, arbitrage between lowprice and high-price customers will eventually break down the discriminatory scheme.
Second, the possibility of price discrimination depends on the ability of producers to separate
customers with different valuations of the product. This may be done directly, using some
information on the customer available to the producer (e.g. age), or indirectly, proposing a menu
of offers and leaving the customers self-select. If it is possible to do this, the firm may be able to
charge a higher price to those customers who have a higher willingness to pay and a lower
price to the more price sensitive customers.
It is useful to make a distinction between two types of price discrimination.86 The first is based
on the buyers having different valuations for the good. In this case, customers in one group are
willing to pay less for the good than customers in the second group. The different sensitiveness
to prices is represented by different price elasticities of demand for the two groups: for the same
price, the price-sensitive group would have a higher elasticity of demand. The second type of
price discrimination is based on buyers having different ability or willingness to switch seller, e.g.
to search for the shop offering the best price. In this case, the two groups are characterised by a
different cross-price elasticity of demand. Quite clearly, this second type of price discrimination
is possible only under oligopolistic competition, and not under monopoly.
Effect of vertical integration
In a vertical industry structure, we must distinguish two cases: (1) when upstream firms price
discriminate between different downstream firms; (2) when downstream firms price discriminate
between different final consumers. Vertical integration may improve the firms’ ability to engage
in either of these practices. We discuss each of these in turn.
Page 44
85
In perfectly competitive industries, where price is equal to marginal cost, price discrimination is impossible, since price
is equal to marginal cost.
86
See, e.g. Winter (1997), Chen (1999). Stole (2003) presents a recent survey of the literature on price discrimination and
imperfect competition.
•
Upstream price discrimination: In this framework, vertical integration is a means of
preventing arbitrage.87 Consider an upstream manufacturer that sells the same
input (e.g. virgin-aluminium ingots) to several downstream firms, some of which are
more price-sensitive than others. For example, airplane manufacturers and
saucepan manufacturers both use aluminium, but saucepan manufacturers may
tend to be more sensitive to the price of the input than the airplane manufacturer.
This may happen, for example, because their respective final customers show
different sensitiveness to prices, or because alternative inputs exist for saucepans
but not for airplanes (e.g. steel). In this case, the aluminium manufacturer would like
to price discriminate between the downstream firms. However, this may not be
possible if the downstream firms are able to trade between themselves. In this
case, the bottleneck firm can overcome the problem by acquiring price sensitive
firms. In this way, it can easily charge different prices to its (price sensitive)
downstream division, and to the independent (price insensitive) firms. It has been
suggested that this effect was seen when Alcoa, the upstream aluminium producer,
integrated downstream into a manufacture of aluminium rolled sheets (a sector with
a large price elasticity of demand).88
•
Downstream price discrimination: Chen (1999) considers an industry where an
upstream manufacturer, with some degree of market power, sells a good to
downstream retailers who compete oligopolistically. For simplicity assume the firm
with market power is a monopolist. Assume further that consumers are
heterogeneous in both their valuation of the good, and in their willingness to shop
around for the best price.89 The key insight here is that the oligopolistic retailers
may engage in both types of price discrimination described previously, but, while
price discrimination based on different valuations (i.e., direct price elasticity)
increases industry profits, the second type, based on willingness to compare prices
(i.e., cross-price elasticity) reduces industry profits, as it induces retailers to
compete more intensely. Therefore the upstream monopolist would seek ways to
prevent it (besides, Chen shows that seeking to prevent this kind of price
discrimination may also increase consumer welfare). Further, the monopolist must
carefully choose the instruments to achieve its goal, in order to avoid the double
marginalisation problem that naturally arises when market power is present at both
levels of the supply chain. Chen shows that RPM, together with a two-part tariff, is
one such instrument. The author also argues that this motivation can explain the
long history of RPM practices in the US gasoline retail market. We note that forward
integration into retailing (together with the exclusion of rival retailers) is another
possible means to realise the same objective.90
87
This mechanism is described by Tirole (1988), p.141.
88
Tirole (1988), Chapter 3.
89
To be more specific, some customers are assumed to be “captive” to each retailer, while others always buy at the shop
with the lowest price in the market. Chen further assumes that (1) captive shoppers are those with the highest valuation
of the good, and (2) the retailers know the type of the customer, i.e., captive or comparison shopper, but not their
valuation.
90
The potential weaknesses of RPM as an instrument to achieve optimal retail price are discussed in Section 3.2.1. For
example, RPM may be suboptimal in the presence of uncertainty.
Page 45
Impact of vertical integration on consumer welfare
In general, price discrimination has an ambiguous impact on consumer welfare. In a
monopolistic market, high valuation customers lose, but low valuation customers normally gain.
In particular, there may be some low valuation customers who, under a uniform price, would not
buy the good, while if price discrimination is possible they would be offered a price that they
would be willing to pay. The welfare effects of price discrimination are discussed in more detail
in Section 4.2.2.
Regarding upstream price discrimination, two possible effects should be considered. First, the
two downstream sectors produce different final products (e.g. saucepans and airplanes, in the
example above), so that lower prices in one market can be accompanied by higher prices in the
other market. Antitrust policy in the EU generally does not allow losses in one market to be
offset against gains in another.91
Second, where more than one firm operated pre merger in the price sensitive sector (e.g.
aluminium rolled sheets in the Alcoa case), vertical integration could result in a price squeeze
on the independent firms (input foreclosure). This does not need be motivated by a strategy to
raise rivals’ costs, but is simply the result of the lower elasticity of market demand resulting from
the merger. Nevertheless, independent firms may be marginalised, and monopoly may result in
the price sensitive segment.
3.3
Increasing Productive Efficiency
Non-horizontal mergers may often allow the production process for one or both goods to
become more efficient, i.e., for the same amount of input a greater amount of output, or a good
of better quality, can be produced. This section considers various ways in which vertical
integration and conglomerate mergers can be used to improve productive efficiency: Section
3.3.1 discusses economies of scope; Section 3.3.2 discusses supply assurance; Section 3.3.3
analyses the benefits of improved information flow and coordination; and Section 3.3.4 looks at
managerial efficiency.
3.3.1
Economies of scope
Source of efficiency
Economies of scope refer to reductions in costs arising from combining the production of
different products. (This can be contrasted with economies of scale which arise by expanding
91
Page 46
A notable exception seems to be the UK: the Enterprise Act 2002, Section 30(1)(a), states that “a benefit to relevant
customers” can take the form of “lower prices, higher quality or greater choice of goods or services in any market in the
United Kingdom (whether or not the market or markets in which the substantial lessening of competition concerned has,
or may have, occurred or (as the case may be) may occur)”.
output of the same product.) Such cost savings can arise from technological synergies, or from
shared indivisible or hard-to-trade assets.
Under what conditions the efficiency occurs
Technological synergies arise when the technical conditions required to produce two goods are
similar. The classic example of this is the vertical integration of pig iron and steel manufacturers.
Pig iron is impure raw iron resulting from the smelting of iron ore; it requires further treatment in
order to produce steel. By undertaking the production processes in the same factory, the pig
iron can be used while still molten, saving on the cost of reheating the input. A similar rationale
can be given for the merger of timber pulp and paper makers. The general conditions for
technological synergies to arise are summarised as follows:
•
The good cannot be easily traded in the market unless it receives some special
treatment;
•
The treatment is unnecessary to the productive process: i.e., it would not be
undertaken if the whole production process was integrated within the same
company;
•
The treatment is costly, either to the seller to perform, or to the buyer to reverse, or
both.
The merged entity may also be able to share indivisible, hard to trade assets.92 Human capital is
a typical example of an input that can be used to produce more than one good, yet is hard to
trade between firms. When productive assets are shared, the total production costs of the
combined entity are lower than the sum of the production costs of the merging parties. This
amounts to a net saving.93
Impact of vertical integration on consumer welfare
The efficiency gains resulting from economies of scope can take the form of lower marginal
costs and/or lower fixed costs. Generally, technological synergies give rise to lower marginal
costs, while shared assets result in fixed costs savings, although this is not always the case.
Marginal cost savings generally give rise to immediate price reductions. Fixed costs savings
may, in certain circumstances, also lead to lower prices.94
92
Teece (1980, 1982).
93
Since sharing of indivisible, specialised assets is more relevant for conglomerate mergers, the detailed discussion is
referred to Section 4.3.1 on economies of scope.
94
The impact of reductions in marginal and fixed costs on consumer welfare is discussed in Section 2.2.3.
Page 47
Can the efficiency be achieved through other means?
In some cases it may be possible, at least in principle, to achieve these benefits through
contractual means. For instance, the pig iron manufacturer in the example above could supply
molten product by locating its facilities next to the steel plant. However, such contractual
solutions may create other problems associated with a vertically separated relationship. For
instance, the steel manufacturer in the example above may not want to become too dependent
on the pig iron producer. We refer the reader to Section 3.5 and 4.5 for a general discussion of
the costs associated with contractual agreements that might replace a merger.
3.3.2
Supply assurance
Source of efficiency
The need to assure the supply of inputs or the distribution of output has traditionally been cited
in the business literature as a major motivation for vertical upstream integration.95 For example,
it has been claimed that airlines merged with petroleum refiners in order to obtain more stable
deliveries and prices of jet fuel.96 Similarly, Paramount Pictures claimed they were merging with
their film theatres in order to guarantee outlets for their films.97
The basic idea is that a downstream manufacturer may fear that its raw material supplier may
not be able to supply the input during periods of high demand. Of course, such failure to supply
will only occur in the short term; in the long term, a shortage of raw material will increase prices,
until demand is brought into line with supply again. However, if the input supplier is not
expecting demand to increase, it may be “caught short” and have to ration the available supply.
Conversely, in times of decreasing demand for the upstream good, the input supplier may find
itself over-producing, and have to bear the costs of storing the extra inventory. Similar problems
may also be created where the demand for the input is relatively constant, but the input supply
is variable.
A recent, sophisticated version of supply/demand coordination is known as Just-In-Time (JIT)
production and distribution. JIT is based upon the principle that no products should be made,
and no components ordered, until there is downstream demand.98 It involves precise monitoring
and forecasting of demand, fine tuning of inventory levels and delivery schedules, and flexible
and timely production. JIT allows firms to save on the cost of inventories and reduces the risk of
shortages. (It also gives rise to other competitive advantages – see Section 3.3.3).
This form of organisation, driven by innovations in information technology, automation, laser
scanning, etc., is increasingly important in both manufacturing and distribution, and is fostering
Page 48
95
Chandler (1969), Porter (1980).
96
Business week, November 17, 1980.
97
Case 334 US 131 (1948); also cited in Mason and Phillips (2000).
98
European Commission (1997).
increased vertical coordination of firms. As pointed out by the Green Paper on Vertical
Restraints, “The adoption of JIT […] in distribution involve[s] a shift from the traditional arms’
length relationship between manufacturers, wholesalers and retailers towards a relationship of
cooperation”.
Note that Bolton and Whinston (1993) show that vertical integration can alleviate supply
concerns for the merging parties, but may also be accompanied by a shift towards internal
supply. This may exacerbate supply assurance concerns for other downstream firms, leading to
a form of input foreclosure.
Effect of vertical integration
Broadly speaking, vertical integration may give a firm more “control” over its inputs, thus
increasing the chance that supply and derived demand will remain in step. Thus, together, the
two firms should be better able to synchronise the production of inputs and outputs. This
reduces the chance that supply will have to be rationed, or extra inventory stored. However,
despite the appeal and apparent simplicity of the argument, economists have resorted to a
number of rather complex models in order to explain the mechanism. We discuss one of these
contributions in Box 3.2.99,
Impact on consumer welfare
As noted by Carlton (see Box 3.2), in markets characterised by uncertainty equilibrium price
tends to be higher. This effect may be due, for example, to the extra costs of unsold production,
or to the risk aversion of some firms in the supply chain (e.g. retailers). If vertical integration can
help minimise these inefficiencies, prices may fall.100 Note that it might often be appropriate to
treat these costs as fixed costs.
Alternatively, or additionally, consumers may benefit from improved supply assurance when it
results in increased product variety. For example, we can expect that new product varieties
would be introduced if the fixed cost of introducing one additional variety were reduced.101 This
cost is likely to fall if the risk of wasteful production is reduced (for example through JIT; see
also Section 3.3.3).
Finally, social welfare (and in certain cases consumer welfare) may increase when vertical
integration raises a firm’s incentive to invest. For example, Bolton and Whinston (1993) analyse
the problem of an upstream monopolist with random capacity in the context of incomplete
99
Early works on this subject include Adelman (1955), Arrow (1975), Green (1986).
100
Note that in Carlton’s (1979) model, vertical integration is not socially optimal, because upstream firms are more
efficient absorbers of risk than downstream firms. However, this particular aspect of his model does not invalidate the
idea that vertical integration can be socially optimal when it is motivated by the need to assure supply or demand.
Indeed, Carlton’s result that integration is suboptimal appears to us as highly dependent on the specific assumptions of
the model, and therefore not sufficiently robust to draw any policy conclusions.
101
Clearly, other conditions must be satisfied; for example, consumers must value variety to a sufficient degree.
Page 49
contracts. Ex-ante, downstream firms underinvest because they foresee that in times of scarcity
the monopolist will be able to renegotiate contracts. This inefficiency creates an incentive to
integrate vertically. On the other hand, integration can also enhance supply assurance problems
of competing firms, leading to a form of market foreclosure. Therefore, vertical integration may
have both positive and negative effects on welfare.
Under what conditions the efficiency occurs
Such problems are likely to be particularly significant in industries where goods are perishable
or otherwise decline in value. For example, in food industries the failure of the market to clear
could give rise to stale or rotten goods. Markets characterised by a high rate of technological
progress provide another relevant example: a computer manufacturer may fear being left with a
glut of chips which will be outdated within the quarter.
Box 3.2: A classic model of integration as a response to uncertainty
Carlton (1979) considers a standard setup, where both downstream firms and upstream
input suppliers are perfectly competitive. He starts from the following basic assumptions:
(a) prices do not adjust immediately to clear the market, but stay fixed in the short term;
further, there are no spot markets, but buyers have to search for a seller, bearing the
risk of not finding the sufficient amount of the good; such market friction may create
situations where there is both unsatisfied demand and unused supply;
(b) upstream firms decide how much to produce before observing demand, for instance
because production is not instantaneous;
(c) overproduction is costly because of the costs of inventories, while underproduction
leads to dissatisfied customers (which in equilibrium translates into a loss of
customers).
The market friction described in point (a) leads to demand uncertainty for the downstream
firms. These can produce instantly to satisfy demand, but only if they can procure a
sufficient amount of input. As a result, demand in the input market is also affected by
uncertainty. Therefore, in the above business environment, firms expose themselves to the
risk of under- or over-producing, and so price at a mark-up to cover the risk. The author
recalls that “it is a characteristic of markets that operate under uncertainty that the
equilibrium price exceeds the cost of production”. By integrating with the input supplier, a
downstream firm can save on this mark-up, but it has to bear the risk of overproducing,
which was previously incurred by the upstream firm. Overall, downstream firms have a
(private) incentive to integrate upstream, since those firms “can use their own input to satisfy
their ‘high probability’ demand, and use the input market to satisfy their ‘low probability’
demand”.
Page 50
3.3.3
Improved information flow and coordination
Source of efficiency gain
In this section we discuss the efficiency gains that firms can obtain by improving the exchange
of information between them and by coordinating their activities. The efficiencies thus obtained
are to be distinguished from economies of scope or scale, whereby the same production is
realised at a lower cost by pooling the parties’ productive assets (including human capital). The
benefits of increased coordination, instead, usually amount to qualitative improvements of the
productive activity. The outcome may be higher quality products and/or the achievement of a
productive organisation not previously available. The benefits include, for example, a higher
responsiveness to the evolution of consumers’ demand and competitors’ product introduction.
The exchange of information and/or the coordination of decisions with a partner firm can be
particularly critical in the following sectors:
•
R&D activity: In R&D-intensive industries, innovation often requires upstream and
downstream firms to exchange information about technology and design of
products. This may be necessary to ensure the products are compatible, to avoid
extra costs of adjustment or to increase functionality102; the introduction of new
technologies often requires changing the interface between products.103 For
example, car manufacturers and component suppliers usually closely coordinate
their R&D activities. Analogously, the manufacturers of computer microchips devote
considerable efforts to ensure that downstream computer manufacturers and
complementary software firms have the information necessary to develop their
products. Indeed, the importance of such information exchange has been
highlighted during the recent European Microsoft case.104 Vertical integration can
thus facilitate coordination of product design, leading to lower costs, higher product
quality and shorter time-to-market.
•
Distribution: Improved coordination and management of demand/supply needs of
vertically related firms can allow Just-In-Time production and distribution. This form
of organisation may allow firms to achieve a number of competitive advantages.
Section 3.3.2 discusses the direct savings on the cost of inventories and the
reduced chances of shortages. Additional benefits include, for example, the
possibility of customising the product to the specific needs of each client by
manufacturing it “on demand”. This is common, for example, in the car industry. It
also happens in the computer industry, where some assemblers are vertically
integrated into retailing (for example through the internet), and allow customers to
choose most of the components. A similar mechanism can be found in the fashion
industry, whereby manufacturers are able to offer a wider range of products, or to
102
Milliou (2004). We note that whenever coordination allows for avoidance of wasteful duplication the resulting efficiency
gain can also be defined as an economy of scope.
103
Farrell and Weiser (2003).
104
EU Case COMP/37.792 - Microsoft/ W2000.
Page 51
change their collections more frequently by closely monitoring demand and stock
levels in each outlet. From an economic point of view, customisation can be
described as an increase in the number of varieties. This is beneficial to consumers
because it allows a better match of the product with their preferences. The increase
in the number of varieties, in turn, is made possible by the reduction or elimination
of inventory costs.
•
Marketing: Downstream firms may have better information than upstream firms on
consumer demand characteristics such as taste, price sensitivity, etc. They are also
better placed to obtain feedback from consumers on product quality and reliability.
This can help the manufacturer to determine the design of a new product, to
improve the design of an old one, or to calibrate its marketing strategy.105 The
benefit can result in higher value for consumers, and/or faster time to market for
new products.
•
Choice of production technology and pricing decisions: Carlton (1979) notes the
very general principle that “price signals” prevailing when non-integrated industries
trade inputs in a spot market may be insufficient to coordinate the introduction of a
new production technology. This may happen because price interaction only signals
the direction of change in the equilibrium price after a possible marginal
improvement, while a new technology may be profitable only when a completely
different market equilibrium is realised. In the model proposed by the author,
industry profits and consumer welfare increase upon the introduction of a new
technology. However, this only happens if firms coordinate their decisions because
unilateral innovation by one firm would be unprofitable if its vertical partner does not
simultaneously modify the input price.
Effect of vertical integration
Vertical mergers that improve coordination and information flow between the parties generally
increase the merged entity’s competitiveness with respect to its rivals, by making its products
more attractive to customers and/or their production more cost-efficient. Such effects are
generally pro-competitive.
Vertical integration, however, may also have a negative impact on social welfare if it leads to a
reduction of cooperation between the merged firm and other non-integrated firms. This may
happen as the integrated firm may be able to exploit or distort the collaboration to provide an
advantage to its other division. For example, the diffusion of confidential information on R&D
innovations towards competitors is described as a horizontal spillover, and is analysed in
Section 3.4.2. Knowledge spillovers also concern other situations: for example, innovations in
business organisations are information goods similar to R&D, and can be revealed through
close cooperation; moreover, they cannot be patented or licensed. In other cases, firms may
fear that the integrated vertical partner would disclose information about their demand for inputs
105
Page 52
Riordan and Salop (1995).
or their bids.106 Nonetheless, even when the merged firm has the means to engage in such
behaviour against competitors, it does not necessarily have the economic incentive to do so.
Examples of these circumstances are provided by the papers summarised below. Thomas
(1996) considers an industry with duopoly in both upstream and downstream sectors, and
shows that both divisions of the integrated firm are indifferent about the transfer of information
on the rival’s bids (the leakage would not increase their profits). Hughes and Kao (2001)
analyse the possible disclosure of private information on the rival’s demand. The results from
their model imply that the upstream division of the integrated firm has no incentive to pass
information to its downstream division.107
In general, the overall impact of integration on industry efficiency and social welfare is therefore
the result of two opposing effects. Integration improves the cooperation and the information flow
between the merging parties, and may reduce the amount of that activity carried out between
the integrated firm and other non-integrated parties. The net effect may be positive or negative.
For example, in the case of R&D activity, total industry investment in innovation may rise or fall,
as described in Section 3.4.2. Assessing the likely outcome and impact on consumer welfare
can therefore be a difficult task. General guidance is provided by the ICE argument and
exceptions to it, presented in Sections 2.3.2 and 2.3.3 respectively.
Impact of vertical integration on consumer welfare
The benefits of improved coordination and information flow more often result in qualitative
improvements than cost efficiencies. Such improvements include, for example, closeness of the
product to consumers’ needs (thus increasing the product value to consumers), increased
variety, shorter time-to-market of new products (thus increasing the rate of innovation), as well
as more traditional quality improvements such as increased reliability of the product.
These benefits to consumers may be offset by the negative effects on the quality of competing
products, described above. But even in the cases where there is a reasonable expectation that
competitors would be damaged by the merger, this does not necessarily imply a reduction of
consumer welfare. The actual impact on consumers depends primarily on rivals’ ability to
respond with strategies suitable to increase their own competitiveness, including possible
counter-mergers. Unless rivals are marginalised and cannot respond, resulting in a strong
increase in market power of the combined firm, the merger should therefore be considered as
pro-competitive.
Nonetheless, in differentiated product markets, some consumers may be damaged in the short
term if the quality of some rival products is reduced. In that case, consumers with a strong
preference for these products might suffer from a welfare reduction. Still, the overall effect on
consumers may be positive if the gains enjoyed by some groups are larger than the losses
incurred by others.
106
Concerns about these potentially detrimental effects of vertical integration have prompted competition authorities to
require behavioural commitments from the merging firms in a number of recent cases.
107
We also note that, in their framework, firewalls increase industry profit, and decrease both consumer and total welfare.
Page 53
Under what conditions the efficiency occurs
As discussed above, the efficiency gains from increased information flow and coordination are
most often of a qualitative nature. Qualitative improvements are notoriously hard to quantify,
although the benefit to consumers that they provide can be substantial. The assessment of
these efficiencies is therefore likely to be difficult, and open to controversy. In some cases,
useful indications can be derived by comparison to similar firms in the same industry or,
possibly, in different industries. For example, where the parties seek to imitate the
organisational structure introduced by another industry player, and where the latter has been
able to gain significant market shares through that innovation, we can reasonably expect the
benefits to be substantial. Furthermore, the increase in market shares indicates a gain for
consumers.
Can the efficiency be achieved through other means?
Clearly cooperation and information sharing can in principle be achieved through a number of
contractual agreements. Indeed, this often happens. In some cases, information is simply traded
as a good, for example, information on consumer demand, at the retailer level, can be
purchased from specialised companies.
However, mergers may be a preferable solution where transaction costs become excessively
burdensome (perhaps to the point that no transaction would take place). A general discussion of
transaction costs is presented in Section 3.5. In the context of cooperation agreements, we note
two problems that are generally likely to raise issues of contractual incompleteness: (1) any
cooperation is easily susceptible to moral hazard problems (such as, but not only, hold-up), and
(2) knowledge is generally prone to free riding, since it may be freely observed and then
costlessly passed on.108
3.3.4
Managerial efficiency
Mergers may have an important role to play in preventing inefficient management.109 If a
management team is inefficient, then the share value of the company will fall below the “true”
value of its assets; i.e., that which would prevail if the management were efficient. Thus, if an
outside management team believes it can do a better job of managing a firm than the current
board, it may try to acquire the firm. This would create a gain to both consumers and
shareholders.
Although the role of managerial efficiency is generally considered in the context of conglomerate
mergers, and indeed any management team can potentially identify an underperforming firm,
vertically related firms may be particularly well placed to assess the possibility of such gains.
For example, they are likely to have a good knowledge of the market the firm operates in, and
Page 54
108
The problem of free riding on R&D effort is discussed in detail in Section 3.4.2
109
See, for example, Slusky and Caves (1991).
they may have gained insights on the way the firm is run during negotiations, special
collaborations, or long term buyer/seller relationships. Analogously, they may have direct
information on its competitors. Further, as explained in the section on economies of scope
above, the management team is more likely to efficiently run the “target” firm than other
managements, with little knowledge of the industry, would do. Finally, as discussed in Section
2.2, firms have an added incentive to increase sales of vertically related firms, as, due to the
complementary nature of the products involved, their own profitability increases.
Clearly, firms may seek mergers and acquisitions for reasons other than those discussed in this
section. In particular, we note that, although improving management efficiency may be among
the positive effects of mergers, there is no empirical evidence that it is a direct consequence of
all (or even most) mergers.
3.4
Preventing Profit Expropriation
This section discusses situations in which a certain firm is unable to appropriate the full benefits
of its investment because some of the gains are reaped by other firms. This effect is described
in economic terms as a positive externality, or a spillover, and the firms which benefit from the
investment of others are termed “free riders”. These situations tend to lead to under-investment;
in extreme cases, investment may be deterred altogether. Such sub-optimal levels of
investment harm consumers, but may also harm firms at other levels of the supply chain. These
may therefore seek arrangements suitable to prevent such inefficiencies. This section is
structured as follows:
3.4.1
•
Section 3.4.1 discusses the common example of free riding on promotional activity;
•
Section 3.4.2 discusses how horizontal spillovers reduce incentives to invest in
R&D;
•
Section 3.4.3 discusses how vertical spillovers reduce incentives to invest in R&D.
Free riding on pre-sale services
Source of inefficiency
Free riding occurs when one firm benefits from investments made by its competitors. For
example, a distributor may invest in promotional activity, by providing brochures or product
demonstrations. This is also likely to benefit rival distributors of the same product. Where one
firm’s actions affect another firm, other than through the pricing mechanism, the benefit is
termed an externality. Free riding is a form of horizontal externality as the firms affected by the
investment are at the same level of the supply chain. Since the firm is unable to appropriate the
full benefit of its investment, it would tend to underinvest in such activities.
Page 55
In addition, however, the retailer’s investment decision may also affect the manufacturer of the
good: if the retailer’s investment in pre-sales activity was at the optimal level, the manufacturer
would sell more (this is the effect of the vertical externality, already discussed in several other
sections110). Thus the manufacturer has an incentive to try and over-come such free-riding
problems. This suggests a second condition for free riding to happen: it depends on whether the
service is contractible. In the cases where it is possible to write a simple contract between the
manufacturer and the retailer specifying the amount of promotional activity to be carried out, for
instance by the retailer, and if the contract itself is not costly, then free riding is not a problem.
Finally, the firm’s investment decisions have an effect on society as a whole. For example,
investments in promotion are generally valuable to consumers when they provide information
about the product. Part of the surplus generated by the promotional activity is captured by the
firm making the investment through additional sales. However, in most circumstances where the
firm is not a monopolist, consumers can also be expected to benefit. Thus, if the retailer is
deterred from making such investments due to free-riding problems, society as a whole is worse
off.
Free riding can occur at the upstream, as well as the downstream, level. Suppose, for example,
that a manufacturer invests in training the retail staff that sells his product. It may be that this
training enhances the sales of other manufacturers’ products, as well as its own. If the
manufacturer factors this as a negative effect of the training, it may be deterred from investing in
training.
Effect of vertical integration
The free-riding problem can only be overcome by reducing or eliminating competition at the
level where free riding takes place.111 If the problem is at the downstream level, it requires a
reduction of intra-brand competition: for example, the upstream manufacturer can decide to
integrate into retailing, and exclude competing retailers; thus any benefit from promotional
activity accrues directly to the single integrated entity, and there is no free riding. If free riding
takes place at the upstream level (e.g. between manufacturers), inter-brand competition must
be reduced; for example, this can be achieved by excluding competitors’ products from retail
outlets.
Underlying this is the problem of “contractual incompleteness” discussed in Section 3.5. Vertical
integration may reduce such agency problems because under common ownership the
incentives of the retailer are aligned with the incentives of the manufacturer. For example, a
manufacturer could impose a contract specifying that trained retail staff only sell its own brand,
while untrained staff are employed for competing brands, but such arrangements would be very
hard to carry out in practice when the firms are independent. In particular, the retailer would
110
Page 56
In contrast to the horizontal externality, the vertical externality acts between firms at different levels of the supply chain;
this fundamental characteristic of vertical relationships (and, similarly, of complementary relationships) was described in
Section 2.1; the vertical externality is also responsible for double marginalisation effect, and also of the vertical
spillovers of R&D activity (Section 3.4.3).
always have an incentive to cheat by using trained staff for other products, and it would be hard
to monitor this behaviour. However, under common ownership, retailer staff would have no (or
much less) incentive to cheat on the boss’ orders.112
Under vertical integration, investment decisions take into account the benefit to the whole
vertical structure, thereby eliminating the vertical externality. Therefore, we can expect the level
of investment in pre-sale activity to rise with respect to pre-merger level. However, in order to
eliminate free riding entirely, additional measures may be required. For example, where free
riding is downstream, a possible solution involves excluding non-integrated retailers from selling
the good.
Under what conditions the inefficiency occurs
A first necessary condition for free riding to take place is that the party undertaking the
investment only appropriates part of the benefit. Pre-sales information is the most common
example of such investment, but free riding can also take more subtle forms. For example,
investments in brand names are often subject to external effects: when franchisees sell under
the same brand name (e.g. McDonald’s, Benetton), the higher quality of service in one outlet
(such as insufficient personnel, cleanliness, etc.) can improve the reputation of all the others.
Second, the investment subject to free riding must be non-contractible. As explained above, if
such investment could be easily observed and specified in a contract, firms would generally
explicitly contract the optimal investment level, and the problem would be eliminated at the root.
However, a number of services and investments are not contractible or, if they are, such
agreements involve excessively high costs (see Section 3.5.1). For example, free riding often
happens in relation to information goods (e.g. promotions, training, reputation, etc.). These have
well-known “public good” characteristics,113 due to the fact that once information has been
transferred it cannot be returned and, moreover, that it can be costlessly passed on.
Third, the problem of free-riding is likely to be greater the higher the degree of competition at the
level where the free-riding takes place.
Free-riding is particularly likely in certain environments. For example:114
•
Free-riding by retailers is particularly common where the good is new, technically
complex or of high value, in which case consumers may not appreciate the full
value of the good without appropriate information.
111
Tirole (1988), p.183.
112
See Hart and Moore (1990).
113
Public goods and services are usually described as those which show the following two properties: (1) they are nonrival, i.e., once they are provided, the cost of servicing one additional user is zero; (2) they are non-excludable, i.e.,
once they are provided to one person, it is impossible to stop another person from using them. For example national
defence or clean air may be classed as public goods.
114
European Commission (2000).
Page 57
•
Promotional activity may also be very valuable when a firm enters new markets and
investment is needed in order to establish the brand.
•
Free-riding between (upstream) suppliers is more likely where promotion takes
place at the buyer’s premises (downstream) and is generic, i.e., not brand specific.
However, free riding is by no means restricted to these situations. For example, manufacturers
may want retailers to promote even relatively straightforward goods, such as CDs or books, or
to raise the profile of an established brand. Similarly, in the case of supplier free riding,
“promotion” should be interpreted widely to include, for example, the provision of credit or
storage facilities which could be used for other manufacturers’ goods.115 In both cases, these
provisions may also benefit rival manufacturers, and would therefore open to free riding.
Can the efficiency be achieved through other means?
The horizontal externality that causes free riding can be eliminated through a number of vertical
restraints.116 These have to be appropriately designed to solve this problem as well as the
common double marginalisation problem. Further, they should take into account the
characteristics of the business environment. In a simple environment, the following standard
restraints are suitable to address the problem:
•
RPM, or minimum RPM (i.e., a price floor), by preventing retailers from competing
on price, are straightforward ways to induce retailers to compete on service,
promotions, etc. Double marginalisation can then be eliminated by using a two-part
tariff;
•
Closed territory distribution is also a suitable means to reduce intra-brand
competition, thus solving retailer free riding. Under such agreements, the
manufacturer would commit to supply only one retailer in each territory;
•
Exclusive dealing can be used to solve manufacturer free riding.
However, as discussed in previous parts of this report,117 in the presence of uncertainty or
asymmetric information, these vertical restraints may not achieve an efficient outcome, and
alternative solutions may have to be devised. In certain cases, competition between retailers
may turn out to provide the best outcome.118
Page 58
115
For example, Dutch brewers provided financial support for their downstream outlets in order to finance refurbishments,
amongst other things. Brewers are particularly well-placed to provide such services as they can monitor the business
risks associated with the loans; in addition, they also benefit from the increase of sales themselves. However, the
provision of such credit allows free riding as, for example, improvements to premises also increase the sales of rival
beers.
116
See Tirole (1988) and Tan (2001) for a review. Papers on specific issues include Winter (1993), Mathewson and Winter
(1998), Bolton and Bonanno (1988).
117
E.g. in the discussion of contractual solutions for the double marginalisation problem, Section 3.2.1.
118
Rey and Tirole (1986a, 1986b).
Moreover, in some cases, such vertical restraints may raise significant competitive concerns.
For example, Rey and Stiglitz (1995) show how exclusive territories can be used by producers
to reduce inter-brand competition and increase profits, reducing consumer surplus and total
welfare.
3.4.2
Internalisation of R&D benefits: horizontal spillovers
Source of inefficiency
Since investment in R&D is generally a sunk cost, it has to be recovered through appropriate
mark-ups (and profits) once the innovating product reaches the market. However, when an
innovation can be easily and quickly copied by rival firms, the innovating firm is unable to
appropriate the benefits, as competition would drive profits towards zero. Foreseeing such
outcomes, firms would underinvest in innovation or, in extreme cases, refrain from investing at
all.119 Therefore, it is essential for an innovating firm to be able to protect its innovations.
Effect of vertical integration
A merger may increase a firm’s incentive to invest in R&D by guaranteeing better protection for
intellectual property rights. Vertical integration can protect a product innovation by making
reverse-engineering more difficult by embedding its innovation in the end product. Alternatively,
or additionally, this may allow a firm to maintain confidentiality about its innovation until the final
product is ready to be released on the market, thus improving the time-to-market advantage.
However, non-integrated firms may be induced to reduce their cooperation with the integrated
firm for fear that the integrated firm would have a greater incentive to leak information to its
subsidiary. For example, consider an upstream monopolist that integrates with one of several
downstream firms. The remaining downstream competitors may fear that cooperation with the
upstream division of the integrated firm may lead to private information being passed on to their
downstream rival. This amounts to an “expropriation” of part of the reward from the investment
in R&D of the independent firms. As a result, the merged firm would increase its R&D
investment level, while other firms would reduce it. However, recent theoretical papers have
shown that the overall impact of vertical integration on social welfare may still be positive, since
the net effect may be to increase total R&D investment.120
119
In economic terms, the sunk costs to be recovered include not only the investment necessary to realise the successful
innovation, but also a fair share of the cost undertaken in unsuccessful projects, since R&D activity is intrinsically
characterised by uncertainty.
120
Milliou (2004).
Page 59
Impact of vertical integration on consumer welfare
When a firm internalises a larger part of the benefits from an investment, in equilibrium it tends
to invest more and, absent other distortions,121 its investment level comes closer to the socially
optimal level. Because of this effect, social welfare tends to increase. Higher investment levels
in R&D may also increase consumer welfare. In particular, investments in product innovation
are very likely to benefit consumers, since the launch of a new product is often considered in
itself a gain for consumers (at least when the number of product varieties is not socially
excessive). Investments in process innovation may benefit consumers by bringing about lower
production costs. These savings, in turn, are generally passed on to consumers in terms of
lower prices, although this may happen in different degrees, depending on the intensity of
competition.122
However, these benefits of vertical integration (and in particular those relative to product
innovation) can be offset, at least in part, by a possible reduction in the level of investment
undertaken by other firms (see above).
Can the efficiency be achieved through other means?
Patents and licences can also be used to protect intellectual property. However, the protection
provided by patents and licences is not always sufficient, and the innovator may fear that other
firms will be able to copy its design. For example, the innovation can, in certain cases, be
imitated, circumventing the patent protection. Moreover, innovations in business organisation
cannot be patented. Further, where time-to-market advantage is crucial, for example in
industries characterized by a high pace of innovation, even a minor leakage of information such
as the general characteristics and functionalities of a new product can greatly reduce the profits
from the innovation.
In some cases it has been suggested that it may be appropriate to achieve the efficiencies
through a merger, but that special behavioural commitments should be imposed on the merging
parties in order to ensure that no confidential information disclosed by third parties is passed on
to the other division/subsidiary of the merged firm.123 This remedy, also known as a “firewall”, is
intended to alleviate concerns regarding the possible reduction of incentives to invest in
innovation for non-integrated firms.
Page 60
121
In certain circumstances firms may have an incentive to invest in R&D by more than the socially optimal amount (see,
e.g. Delbono and Denicolo 1990).
122
A large literature evolved around this issue building on the seminal papers by D’Aspremont and Jacquemin (1988) and
Kamien, Muller and Zang (1992).
123
This approach was used by the FTC in a number of recent merger cases in different industries: pharmaceuticals
(Merck/Merck-Medco), satellites (Boeing/Hughes), energy (PacifiCorp/ Energy Group), defence (Raytheon/Chrysler,
Martin Marietta / General Dynamics, Lockheed/ Martin Marietta, Lockheed Martin / Loral).
3.4.3
Internalisation of R&D benefits: vertical spillovers
Source of inefficiency
When a firm is part of a vertical supply chain, some of the benefits from its R&D effort may be
reaped by its vertical partners. Consider, for example, the case where a firm has the
opportunity, by investing in R&D activity, to reduce the marginal cost of production of a certain
good. Such process innovation would lower the price of the final good, thus increasing output
and consumer welfare. However, if the firm’s distributor has the power to charge a mark-up over
its marginal cost, it will appropriate part of the increase in surplus, as the decrease in the price
of the goods will increase its profits. Analogously, input suppliers upstream of the innovating firm
will, if competition is imperfect, respond to an increase in demand caused by downstream
innovation by raising input price, thus expropriating from the innovator part of the benefit of the
downstream investment. The source of this inefficiency lies in the vertical externality already
encountered in other sections of this report (see Section 2.2.2).
The same sort of inefficiency, due to an expropriation of R&D returns, is in general also present
between producers of complementary goods. This is due to the fact that the externality
experienced by these firms is very similar to a vertical externality.
Effect of vertical integration
An integrated firm, in general, can be expected to adopt the investment decisions that maximise
overall company profits. Therefore, the vertical externality between the two parties is eliminated.
Since a larger part of the surplus from investment is appropriated, the incentive to innovate
increases and the distortion is reduced.124 A numerical example of this is given in Box 3.3.
The impact of vertical integration on competitors depends on several factors.125 Banerjee and
Lin (2003) point out that since a process innovation126 by one downstream firm generally
increases the price of its inputs (through the increased demand), it also has the effect of raising
rivals’ costs. Thus, when firms are not vertically integrated, the RRC effect may increase a firm’s
incentive to innovate, mitigating the negative impact of the surplus expropriation. Indeed, in the
authors’ model, where inputs are monopolistically supplied, the downstream firm innovates
more if it is subject to oligopolistic (Cournot) competition than if it is a monopolist.
124
A firm is never able to appropriate the full benefit of the investment, because parts of it are generally reaped by nonintegrated input suppliers and customers. Moreover, as long as knowledge of the invention spills over to society, the
benefits are spread more broadly (for example as other firms may be able to build on the innovation). See also Farrell
and Katz (2000) for a discussion of the cases where integration may not increase the investment level, or where the
increase may be socially undesirable.
125
For example, Choi (2003) analyses R&D incentives and the decision to integrate (with a producer of complementary
goods), in the context of duopolistic competition and differentiated goods.
126
A process innovation refers to more efficient production of the same good, as opposed to a new product, or a quality
improvement of an existing product, which is labelled as product innovation.
Page 61
Under what conditions the inefficiency occurs
The inefficiency occurs when firms at other levels of the supply chain are able to charge a markup over their marginal costs. This condition is similar to the one which holds for the double
marginalisation effect. As in that case, the presence and magnitude of the inefficiency does not
depend on whether the mark-up is a pure profit gain, or is needed to recover fixed or sunk
costs. In general terms we expect the magnitude of the inefficiency to be greater, the larger the
gross margins127 of firms in the supply chain independent from the innovator.
Impact of vertical integration on consumer welfare
The merged firm’s increased incentive to invest in (socially valuable) innovation generally
results in increased social welfare, and may also increase consumer welfare. The reasons for
this are discussed below.
For R&D aimed at process innovation, consumers generally obtain part of the surplus if the
marginal cost of production is reduced. (In certain circumstances, analysed in Section 2.2.3,
fixed costs reductions can also result in lower prices). The effect is larger if the demand for the
final good is more elastic. A larger price elasticity of demand implies that final consumers get a
larger share of the efficiency gain, because the suppliers (at all levels of the chain) have a larger
incentive to expand output by reducing prices. Note that what matters here is the residual
demand for the final product in question, not the total market demand. Indeed, even if total
demand is inelastic, so that a price reduction would not expand the market size, it will still, in
general, expand the market share of that supply chain.128 Therefore, a firm may have an
incentive to reduce price even when total market demand is inelastic. The converse argument
also holds: if a firm’s output is unlikely to increase, the price is unlikely to fall.
Page 62
127
I.e., selling price minus variable costs.
128
Here we are assuming that firms at different levels of the supply chain integrate, or otherwise engage in close
cooperation so as to maximise efficiency of the whole chain. In this situation, competition would take place between
different chains of supply. (See, e.g. European Commission, 1997).
Box 3.3: Vertical integration and R&D incentives in a simple model
The following example illustrates how the vertical partners of a firm can appropriate part of
the profits resulting from innovation and, as a consequence, may cause the firm to
underinvest in R&D. Consider an industry composed of an upstream monopolist (e.g. the
manufacturer) and a downstream monopolist (e.g. the distributor). One unit of the
intermediate product is necessary to produce one unit of the final good. The firms are
independent, and sell at a uniform unit price. This industry structure is illustrated in Figure
3-2. The inverse demand function for the final product is
pd = a − q ,
where a > 0 ,
q is output, and p d is the price to final consumers. The time structure of the
model is as discussed in Appendix A.3: The upstream firm sets its price, p u , in the first
stage; subsequently, the distributor sets the price to consumers, taking the input price as
given. Since both firms charge a mark-up over marginal cost of production, the final price
is inefficiently high, and output inefficiently low. However, the upstream firm can exploit its
first-mover advantage to earn a higher profit than the downstream firm. In equilibrium,
output and profits are as follows:
( a − C d − Cu ) 2
Πu =
,
8
(a − Cd − Cu ) 2
Πd =
16
.
It is clear from the above results that a hypothetical reduction of the marginal cost of
production (either C d or C u ) increases both firms’ profits. However, the investment in
R&D necessary to bring about the process innovation is undertaken independently by the
two firms. Since each other’s profits are neglected in the investment decision, the firms
would tend to underinvest in R&D. For example, if C d = C u = 10 , and a = 100 , the
marginal profitability of a reduction of C d or C u to the respective firms is
dΠ u
1
= − (a − C d − C u ) = −20,
dC u
4
dΠ d
1
= − (a − C d − C u ) = −10
dC d
8
This means that the distributor would be willing to invest up to 10 to reduce C d or C u by
1, while the manufacturer would be willing to invest 20 to reduce C d or C u by 1.
However, the two firms together would invest up to 30 (i.e., 20+10) to reduce by 1 either
C d or C u . (Furthermore, it can be shown that if the two firms were to integrate, thus
eliminating double marginalisation and expanding output, the merged entity would invest
up to 40 for the same innovation.)
Page 63
For R&D aimed at product innovation, i.e., a new product, or a quality improvement of an
existing product, one can generally assume that the innovation is a gain for consumers. This is
motivated by two observations: first, generally consumers obtain some part of the surplus from
any product (whose output is greater than zero), because the supplier is unable to price
discriminate perfectly (i.e., charge each consumer exactly according to their willingness to pay);
second, even in the case where one supplier can perfectly price discriminate, it would still be
unable to extract all surplus. This holds because the monopolist is constrained by the
consumers’ ability to switch to a different product from which they gain a surplus. An example
involving substitution between TV sets and holidays is provided in Section 4.2.2.
Finally, innovation generally benefits consumers, and society as a whole. In particular, when
subsequent innovations can build on a previous one, consumers may obtain an indirect benefit
from the first innovation even when they do not have any direct gain.
Can the efficiency be achieved through other means?
Banerjee and Lin (2003) suggest that the firms could reduce this inefficiency through a fixedprice agreement in the case where the innovator is downstream and the input supplier is a
monopolist. This would be signed before the R&D investment is made and would bind the
upstream firm not to raise the unit price when the innovation is introduced. According to the
authors, in many cases such a contract would in fact increase the monopolist’s profits. However,
this kind of agreement might be unfeasible or inefficient in business environments characterised
by either uncertainty on production costs or highly volatile demand.
Agreements whereby the upstream firm accepts to finance part of the R&D investment of the
downstream firm (or vice-versa) are generally difficult to realise due to problems of contractual
incompleteness, notably (1) the hold-up problem, and (2) the difficulty in specifying and
monitoring the effort devoted to R&D. These issues are discussed in Section 3.5. Further, we
note the general problem that R&D activity is inevitably characterised by uncertainty, because
the investment is usually committed before knowing the exact outcome. Therefore, one of the
conditions (i.e., uncertainty) that facilitate the emergence of contractual incompleteness
problems (and the associated costs) is automatically satisfied.
3.5
Incomplete Contracts and Transaction Costs
The earlier parts of this section have discussed how vertical integration can eliminate
inefficiencies and thus generate benefits for the integrated firm, and also for consumers.
However, as discussed, in many cases these benefits can also be achieved by contractual
means. Thus, a firm’s decision regarding vertical integration boils down to a choice between
Page 64
replacing external market transactions with internal exchanges: firms will seek vertical
integration when the cost of market transactions exceeds the cost of internal exchanges.129
Thus, the question of whether a certain efficiency gain can be achieved by contractual means or
requires a merger ultimately lies in the analysis of the extra costs of such external transactions.
Therefore, the issue of transaction costs is central to the analysis of non-horizontal mergers,
both from a conceptual and a practical point of view.
The transaction costs of a simple cash-on-delivery sale are usually small. However, even these
transactions may give rise to significant costs when the value exchanged is small and the
number of deals large, or when considerable search costs are incurred in order to find the
appropriate trading partner.
Alternatively, and perhaps more often, transaction costs are large because the deal involves the
establishment of a complex relationship between the parties. As the complexity of the trading
relationship – and of the business environment – increases, contracts also become more
complex. In uncertain environments, for example, agreements may specify the parties’
undertakings under many (but inevitably a limited number of) possible circumstances. Further,
certain assets, such as know-how, are difficult to trade because they cannot be easily specified
in a contract. The same may hold for product quality, or for the effort delivered by an employee.
Where contracts are not fully specified, firms may fear the other party will exploit contractual
loopholes after entering into the agreement. These circumstances introduce further costs
associated with the transaction: writing up the contract, incentivising the parties, monitoring
compliance and, when necessary, pursuing litigation.
This section discusses how transaction costs arise, and in what situations they are more likely
to be significant. It also discusses the extent to which vertical integration solves the problems,
and analyses some of the common alternatives to a full merger. Section 3.5.1 below presents a
general introduction to contractual incompleteness and the risk for opportunistic behaviour that it
may present; Section 3.5.2 analyses a particularly important instance of the incomplete
contracts problem, the so called hold-up problem; finally, Section 3.5.3 discusses the costs
associated with simple cash-on-delivery transactions.
3.5.1
Incomplete contracts
Source of inefficiency
As explained in the introduction to the section, the establishment of contractual agreements
between independent firms is complicated by a number of problems related to the fact that
contracts, in general, cannot specify the parties’ undertakings under all possible future
circumstances. This is true for two reasons:
129
Coase (1937).
Page 65
•
First, many events are not easily observable or measurable, and therefore cannot
be specified in a contract or enforced by a court. A typical example is the effort
placed in R&D activity by a hypothetical subcontractor. When a research project
fails, it may almost be impossible to tell whether this is due to lack of effort or bad
luck.
•
Second, even in the case where all actions were measurable, the mere number of
potential events may make it impossible to write a complete contract. The more
uncertain the business environment, the more potential outcomes it will be relevant
to consider.
All these problems are described in the economic literature as contractual incompleteness.130
The problem with incomplete contracts is that loopholes may allow the parties to behave
opportunistically once the contract has been signed. The risk of such an outcome is usually
described as moral hazard.
Incomplete contracts usually imply a higher cost to one of the parties. In the simplest form, this
is the cost of monitoring the contracted actions. Alternatively, it is the cost of the incentive given
to the other party in order to behave. To describe this point accurately, some definitions are
required. In the simplest setting the first party, called the principal, contracts the provision of
some service or good to the second party, the agent. However, the principal does not have full
control of the agent’s behaviour, because she cannot observe or write a contract specifying all
the qualities of the service provided.131 Thus, once the contract is signed, the agent may have
an incentive to behave opportunistically, undermining the value of the contract for the principal.
For example, in the case of R&D outsourcing, an external researcher may be able to reduce his
effort and blame the lack of results on bad luck.
In order to motivate the agent to behave, the principal must grant a suitable incentive, or
engage in costly monitoring (when this is possible). In other words, the conflict of interests
experienced by the agent is solved by aligning his incentives with the principal’s incentives.
However, the need for reward or monitoring introduces costs for the principal, and therefore
reduces her motivation to engage in the contract in the first place. These costs are usually
described as agency costs. In extreme cases, the cost of incentivising an agent may offset a
principal’s incentive to engage in the contractual relationship at all. This constitutes a social
loss, as well as a private loss, whenever a value-generating activity is not undertaken.
Effects of vertical integration
Agency problems usually affect the relationship between employer and employee, as well as the
one between shareholders and managers. For this reason, an integrated and larger firm is not
necessarily more efficient (in terms of agency costs) than two smaller independent firms.
Page 66
130
For a general review see e.g. Milgrom and Roberts (1992).
131
Or, even if the principal can observe the agent’s behaviour, a court could not verify compliance to the contract.
Therefore, agency costs, in general, are not eliminated by vertical or lateral integration, but in
certain cases they can be reduced.
From a theoretical point of view, this issue was analysed by Grossman and Hart in a seminal
paper of 1986.132 The article discusses the fundamental question of how the scope for
opportunistic behaviour changes when a contractual relationship between two independent
parties is replaced by an employer-employee relationship under common ownership. The
authors distinguish two types of contractual rights: specific rights, which are specified in the
contract, and residual rights, defined as all the rights of control over the parties’ assets not
specifically given away in the contract. “When it is too costly for one party to specify a long list of
the particular rights it desires over another party’s assets, it may be optimal for this party to
purchase all the rights except those specifically mentioned in the contract”.133 The difference
between contracting and integration amounts to the purchase of these residual rights of control.
The ownership of residual rights affects the parties’ incentives because it determines the value
each party can get by walking away from the relationship. In this way it changes the parties’
bargaining positions and thus the distribution of ex-post surplus. Since the parties’ ex-ante
investments are determined by the ex-post benefit they expect, the authors conclude that
ownership by firm 1 of firms 1 and 2 is optimal when production requires a large relationshipspecific investment by firm 1 and a small investment by firm 2, while a contract is preferable
when a modest investment level by both firms is required.
Along the same lines, there are two further advantages of integration.134 First, a single entity can
selectively fire (or reward) any employee, while if one firm subcontracts to another, it has no
control on the employees of the latter (it can cut the ties with the subcontracting firm, but the
degree of influence is not the same). This changes the employees’ incentives. Secondly,
integration reduces the number of parties that can bargain over how to split the surplus (i.e., the
profits). This increases the expected profits of each of the remaining parties, thus increasing
their ex-ante incentive to invest in the relationship.
The residual property rights approach illustrated above appears more suitable to describe
owner-managed firms rather than large companies, at the very least for the reason that “key
decision makers in such companies (the CEOs, for example) do not have substantial ownership
rights before or after the merger”.135 Moreover, agency costs within a firm depend on the
organisational structure adopted (e.g. multiple-divisions vs. subsidiaries, degree of centralisation
of decision-making, etc.), on the incentive schemes applied to managers and employers and,
ultimately, on the efficiency of corporate governance. They also depend on the efficiency of
internal (and external) auditing. A well managed company can exploit all these means in order
to minimise agency costs.
132
The paper builds on previous works by Coase (1937), Klein, Crawford and Alchian (1978) and Williamson (1979). Along
the same lines, see also Hart and Moore (1990) and Hart (1995).
133
Grossman and Hart (1986).
134
Hart and Moore (1990).
135
Hart and Holmstrom (2002). On this and additional arguments see also Holmstrom and Roberts (1998).
Page 67
Under what conditions the inefficiency occurs
Moral hazard is likely to present a particular problem in the following cases:
•
First, where one of the parties’ undertakings is hard to measure, contracts will tend
to be more complex and monitoring costs will be higher. (However, Grossman and
Hart argue that auditing capabilities do not increase with integration.)
•
Second, in uncertain economic environments, there is a greater number of potential
outcomes which the contract has to cover.136
•
Finally, where there is a complex relationship between two firms, an equally
complex contract is also likely to be required. This increases the risk that the
contract will be incomplete, and therefore also increases the chance of moral
hazard.
These criteria are likely to be met in particular circumstances. For example, in the discussion
above, we highlighted the problem of contractual incompleteness in the case of R&D activity,
while the following section analyses the hold-up problem. However, other examples include:
Page 68
•
Human capital: Market contracts for the transfer of know-how or organisational
knowledge are generally difficult to specify and to enforce. Therefore, such
transfers are more easily accomplished within a firm “due to more highly developed
information networks, auditing capabilities and internal reward schemes”137.
•
Promotions: In vertical relationships, promotional investment by one firm in the
supply chain will benefit another. For example, a retailer can offer its customers
services such as free delivery, free alterations, credit, pre-sales information or
elaborate premises. By increasing sales, this will also benefit the manufacturer.
Therefore, the manufacturer may wish to stipulate contractual terms specifying the
amount of promotional activity provided by the retailer; however, it may be very
costly for the manufacturer to monitor compliance. Left to its own devices, the
retailer would only invest to the extent that it benefits its own sales, and would not
take into account any extra benefits that may accrue to the manufacturer.
•
Quality uncertainty: Moral hazard may also arise in vertical relationships if
downstream manufacturers are unable to easily verify the quality of the input into
their production process. For example, they may be unable to test the durability of a
component in their product, or to test the life expectancy of mechanical equipment.
In this case, the downstream firm may have an incentive to pass off low grade
goods as high quality. In order to overcome such quality uncertainty, the
manufacturer may prefer to produce the input or equipment itself. Hennessy (1996)
136
John and Weitz (1988) also introduce behavioral uncertainty, measured by the typical time between an initial contact
and the ultimate placement of the order, on the basis that large lags in the selling environment make it difficult to infer
which actions were responsible for the observed outcomes.
137
Levy and Haber (1986); see also Teece (1980, 1982). However, Grossman and Hart (1986) argue that auditing
capabilities do not increase with integration.
suggests this may provide a motivation for vertical integration in the food industry,
where it is costly for food processors to accurately asses the quality of raw inputs.
•
Cost-plus agreements: Moral hazard may also arise where contracts are
undertaken on the basis of “cost-plus” agreements which specify a mark-up over
production costs. For example, where the cost of a raw material may be highly
variable, the downstream firm may be in a better position to bear the risk than the
intermediate goods producer. In this case, the downstream firm may purchase the
good from the intermediate firm on the basis of a cost-plus agreement. However, in
this case the input provider has little incentive to reduce its cost by, for example,
choosing the most cost effective raw material.138
Can the efficiency be achieved through other means?
Apart from contracting appropriate incentives, firms can sometimes solve such moral hazard
problems by building a reputation, or by proving a commitment not to behave opportunistically.
For example, if a firm (the agent) engages in similar relationships on a regular basis, possibly
with different firms, it may find it profitable to build a reputation for not exploiting hard-tomeasure promises in contracts. The long term gain from such a strategy may exceed the short
term profit from cheating on one relationship. Alternatively, as described in the section on the
hold-up problem, the firm may take actions which commit it not to cheat. Where there are other,
less costly means of achieving the efficiency benefits, the benefits of vertical integration in terms
of reducing agency costs, are clearly more limited.
3.5.2
The hold-up problem
Source of inefficiency
The hold-up problem arises when an investment which is specific to the relationship between
the two firms is required.139 Such an investment could be undertaken at either the upstream or
the downstream level. For example, a component manufacturer may need to develop a version
of the component which is particular to a client’s product. Alternatively, a distributor may
produce brochures and other marketing materials which relate to a manufacturer’s products in
order to sell them more effectively.
Such an investment makes the firm very dependant upon the vertical relationship, and therefore
puts it in a weak position with respect to the other party, which may exploit this advantage once
the investment has been undertaken. For example, a manufacturer may agree to design and
produce specific components for a client’s product. However, once the manufacturer has
produced the design, and invested in the machinery for the component, the purchaser may
threaten to switch to another supplier, rendering the investment useless, unless the
138
Williamson (1975).
139
E.g. Williamson (1975, 1979, 1980 should this be 1989), and Klein, Crawford and Alchian (1978).
Page 69
manufacturer lowers its price. The parties could try and draw up contracts in advance which
specify all details of the relationship. But, in practice, such contracts will be highly costly, and
will rarely cover all possible means by which the purchaser may undermine the relationship.
Foreseeing its weak bargaining position following the investment, the manufacturer may be
deterred from investing.
Under what conditions the inefficiency occurs
There are three necessary conditions that must be met for an investment to cause a hold-up
problem.140 First, the investment must be relationship-specific. That is, an investment
undertaken by a supplier cannot be used to supply any other customers. Equally, any
investment undertaken by a buyer cannot be used for purchases from any other supplier. If the
investment can be profitably applied to other relationships, the firm will be able to recoup the
investment through another vertical relationship and will not be dependent on the other party.
Second, the investment must represent a sunk cost; i.e., if the investment takes the form of a
purchase of specific machinery or equipment, the hold-up problem will only occur if the
purchase cannot be sold without a significant loss. Third, it must be impossible or excessively
expensive to prevent opportunistic behaviour through an enforceable contract.
In addition, the hold-up problem may be more likely to occur in certain economic environments.
For example, the problem is often associated with long-term investments. This is because,
where an investment is only short-term, or can be quickly recouped, the firm will not be
dependent on the vertical relationship for any great length of time. This lessens the need for
long-term protection in the form of a vertical agreement or vertical integration.
Knowledge based transactions are also particularly open to exploitation; this is because once
such knowledge has been passed on it cannot be retracted. For example, an accountant may
put forward a particular solution to a client. If, having received this advice, the client does not
pay the bill there is little the accountant can do.
Finally, the hold-up problem may also be more common where the contracting parties undertake
relationship-specific investments of very different amounts, or where those investments are not
simultaneous (asymmetric investments). If investments of similar importance are undertaken on
both sides, the parties are equally dependant on each other, and thus there is a reduced
chance of exploitation.
Such investments may fall into one of four different categories:141
•
Page 70
Site-specific assets: where the parties locate “cheek-by-jowl” in order to minimise
transportation or inventory expenses. For example, where the process of converting
a raw material involves adding substantial bulk, transportation may be costly. It will
140
These conditions are recognised by the European Commission (2000), §115 (4).
141
Williamson (1989).
therefore be easier to transport the good before it is processed, and locate near the
client, in order to minimise transport costs.
•
Physical asset specificity: i.e., assets designed specifically for the other party.
Examples of this include specialised dies used for producing a component specific
to a manufacturer’s requirements.
•
Human-capital specificity or “know-how”: Staff may have skills which are specific to
a particular relationship.142 Such skills may be acquired by formal training or by
“learning-by-doing”. For example, a distributor may employ staff to sell a particular
piece of machinery. This member of staff may have detailed knowledge and training
about that make, and would be significantly less equipped to sell other brands.
•
Dedicated assets: These are assets which are potentially general purpose, but
which would not have been bought without the business of a particular client. The
loss of that client would leave the firm with significant excess capacity. For
example, a manufacturer may require an extra production line to serve a particularly
large client. If there are no other sizeable buyers in the market, it may have difficulty
filling this capacity should it lose the business.
Can the efficiency be achieved through other means?
It is possible that hold-up problems can be overcome with vertical restraints. For example, when
the investment is on the upstream side, the manufacturer could impose a quantity forcing
obligation on the distributor. In this case, the distributor has to sell a minimum quantity of the
manufacturer’s goods, and will therefore be less inclined to disrupt the relationship. When the
investment is made by the retailer, obligations by the manufacturer not to sell through other
retailers (such as exclusive distribution, exclusive customer allocation or exclusive supply
obligations) may alleviate the problem. However, such contracts will never align the interests of
the two parties as fully as in the case of full integration, and they may also introduce further
transaction costs.
In addition, Rey and Tirole (2003) provide a notable example of a case where the hold-up
problem is solved by a commitment device. A vertically integrated firm sells high-tech
intermediate products to firms in the downstream market, where it also competes itself. The
downstream partners have to choose a technology for their production, but once adopted it
cannot be changed easily. A hold-up problem may arise, as the downstream firms may fear that
if they choose the platform of the integrated firm, the upstream firm may later exploit the
relationship to reduce their competitiveness to the benefit of its own downstream division.143
Therefore, the integrated firm may rationally decide to divest its downstream division, in order to
show its commitment not to foreclose. According to the authors, this explains the divestiture by
142
These may include knowledge about the abilities and efforts of the personnel, information about the matching of
employees and tasks, and knowledge specific to the production methods or organisational methods of a firm (Prescott
and Visscher, 1980).
143
This could be done, for example, by providing competitors’ inferior products, raising prices, etc.
Page 71
AT&T of its equipment division Lucent Technologies. However, divestiture and other such
commitment devices may involve the loss of other benefits of vertical integration and will also
introduce additional costs. Thus the acquisition of the downstream firms may provide a more
efficient solution.
3.5.3
General transaction costs
As described above, when there is moral hazard, transaction costs can be incurred in
monitoring or incentivising other firms. However, transaction costs arise to a greater or lesser
extent in almost all contractual relationships, even when moral hazard does not occur.
For example, there are costs involved in drawing up contracts, enforcing them, and pursuing
litigation in the event of a breach of promise. In addition, “search costs” may be incurred in
looking for a party offering the suitable good, or a party that offers a sufficiently large stock of
the good. Search costs may arise where a good is supplied by a large number of competing
firms and trade does not happen simultaneously in a centralized marketplace. Hennessy (1996)
gives the example of the food industry, where food-processing firms need to procure a large,
steady supply of raw inputs.
Simple cash transactions generally involve low costs. However, costs incurred by the seller can
increase if the buyer is allowed to pay with a delay (e.g. 30 days) instead of “cash-on-delivery”.
In this case the seller is implicitly providing credit, and has to bear the costs of funds and the
risk of insolvency.
As transaction costs are an inefficiency their reduction is beneficial to society and may result in
customer benefits through lower marginal costs and/or lower fixed costs (See Section 2.2.3).
Page 72
4 Pro-competitive Motivations for Conglomerate Mergers
4.1
Introduction
Efficiencies in conglomerate mergers
This section considers the sources of efficiency that can arise from conglomerate
concentrations. Following the same approach used in Section 3, we analyse efficiencies giving
rise to price and cost reductions, but also to improved quality and variety of goods and services,
as well as increased innovation leading to new products.
Figure 4-1 below summarises the potential sources of efficiency gains, dividing them in the four
categories outlined in Section 3.1, i.e., increasing pricing efficiency, increasing productive
efficiency, preventing profit expropriation, and reducing problems associated with contractual
exchanges (namely incomplete contracts and transaction costs). The figure also specifies which
type of conglomerate merger is most likely to give rise to each efficiency: a merger of
complementary goods, unrelated goods, or both types. This classification is, however, intended
to be a presentational tool, rather than a sharp definition of ranges of applicability.
Figure 4-1: Classification of conglomerate efficiencies
Efficiency
Merger Type
Increasing pricing efficiency
Avoiding the Cournot effect
Complementary
Price discrimination (metering)
Complementary
Reducing customer heterogeneity
Unrelated
Increasing productive efficiency
Economies of scope and scale
Complementary and unrelated
Financial efficiency
Unrelated
Improved information flow and coordination
Complementary and unrelated
Countervailing buying power
Complementary and unrelated
Managerial efficiency
Complementary and unrelated
Preventing profit expropriation
Quality standardisation
Complementary
Internalisation of R&D benefits: horizontal spillovers
Complementary
Internalisation of R&D benefits: complementary spillovers
Complementary
Incomplete contracts and transaction costs
Incomplete contracts and moral hazard
Complementary and unrelated
One-stop shopping
Complementary and unrelated
Page 73
Tying and bundling
Conglomerate firms often have the choice of selling their products (whether complementary or
unrelated) either independently, as stand-alone components, or tied together, as a “package”.
There are several ways of tying or bundling products, which are described below, and several
reasons for doing so. This report does not focus on the anti-competitive motivations for tying
and bundling.144 Instead, we note that some of the efficiency gains listed in Figure 4-1 can only
be reaped when the products of the integrated firm are tied together (i.e., metering, illustrated in
Section 4.2.2, reducing customer heterogeneity, discussed in Section 4.2.3, and protecting
innovation in Section 4.4.2); in other cases, selling a “package” allows further efficiency gains
(e.g. the Cournot effect in Section 4.2.1) or is the most certain and effective way of achieving
the efficiency (e.g. quality standardisation discussed in Section 4.4.1). Therefore, there are
strong pro-competitive motivations for bundling and tying, as demonstrated in the following
sections.
In the economic literature there are considerable divergences in the precise definitions of
bundling and tying.145 In this report, we often refer to both practices simply as “bundling”.
However, where a precise distinction must be drawn between the different strategies, we adopt
the following definitions. We distinguish between bundling and tying according to whether the
packages offered fix a precise proportion between the products. For example, in a packaged
holiday the offer includes one airplane ticket for each hotel “bed” purchased; while a customer
purchasing a copier machine can successively buy any number of toner cartridges, provided
they are of the same brand as the copier. We define the former as “bundling”, and the latter as
“tying”. The precise definitions of these practices are as follows.
Page 74
•
Pure bundling. This refers to the practice where two products have to be purchased
together, as they are not sold or are useless on a stand-alone basis. Further, the
products have to be purchased in fixed proportions, e.g. a new car sold with the
four tyres, without the possibility of choosing the brand of the tyres.146
•
Mixed bundling. This practice implies that two goods are offered to customers as a
package in addition to being sold individually. The package, or bundle, is sold at a
discount with respect to the sum of the prices of individual components. Further, the
proportion of the products within the bundle is fixed.
•
Tying. Tying occurs when purchase of product B requires that customers also
purchase product A. Once A has been purchased, the customer can buy any
quantity of the product B. Thus, tied products can be purchased in variable
144
For a recent review of the anti-competitive motivations for bundling see Nalebuff (2003) and Nalebuff and Majerus
(2003); see also OECD (2002).
145
See for example Nalebuff (2003) and Church (2004).
146
Upstream firms may sometime bundle products that are substitutes for the final consumer but are perceived as
complements by downstream firms (e.g. retailers).
proportions. Further, in tying the purchase of A and B need not be simultaneous. A
typical example of tying is copiers and toner cartridges described above.147
•
Conditional quantity rebates. This commercial practice can be distinguished from
tying and bundling as customers can choose whether to purchase products A, B or
both, from the same brand, and the goods can be purchased in variable
proportions. However, the customer is entitled to a discount on the unit price of B if
it also purchases product A, i.e., the rebate is conditional on buying A.
Note that bundling and tying can be achieved either as a contractual obligation or through the
technical design of the products. The former case is generally referred to as “commercial”
bundling or tying, while the latter can be described as “technical bundling” or “technical tying”.
Thus, technical tying or bundling denotes the practice of making two products technically
incompatible with third parties’ products. Here, the notion of “incompatibility” implies that the
system good obtained by mixing incompatible components does not function, or works less
effectively than a system made of compatible products. Technically tied goods are also
described as forming a closed system, as opposed to open systems where both components
can be freely chosen on the market between competing brands, and then matched together.
Clearly, technical tying or bundling can only be employed with complementary products, and not
with unrelated products.
As a final remark, we notice that bundling and tying are not always viable strategies, even when
they are economically advantageous. The nature of the products and/or consumers’ demand
can make some of these commercial practices impractical or impossible. Which one is more
realistic (if any), is to be assessed on a case by case basis. For example, most car
manufacturers offer credit to their customers. The car and the credit are complementary
products, and pricing efficiency dictates that the bundle should be offered at a discounted price
with respect to the sum of individual components. In this case the only viable strategy is mixed
bundling, since pure bundling would impede sales to those customers who do not need credit.
Degree of complementarity
For several efficiency sources described in this section, the magnitude of the gains depends on
the degree of complementarity between the products of the merging firms. In the following we
provide qualitative and quantitative criteria to assess the degree of complementarity between
two goods.
147
There may be some ambiguity between the definitions of mixed bundling and tying when customers are offered to
choose from a menu of bundles which differ only in the proportion of the two goods (e.g. A+B. A+2B, and A+3B). In this
case, we would tend to denote the practice as bundling if the customer only has a limited set of choices. While we
Page 75
Qualitative criteria:
•
Two goods are stronger complements when the value of using them jointly is higher
than the sum of the values of using each of them independently; this implies that
secondary uses of the two goods are less important.
•
Two goods are stronger complements when they must be used in fixed or nearly
fixed proportions.
For example, cars and petrol verify the first criterion but not the second: each good is valueless
without the other, but different people use different amounts of petrol for the same car. When
the above criteria hold perfectly, i.e., the goods are valueless unless used together and must be
used in fixed proportions, the two goods are often described as “perfect complements”. For
example, nuts and bolts, or car chassis and car engines can be regarded as perfect
complements. In this case, we generally assume that customers are not sensitive to the price of
a single component, but only to the total price of the system. Therefore, when the price of one
good increases, demand for both goods decreases by the same amount.
Quantitative criteria. When suitable data are available, the degree of complementarity can be
measured by the cross price elasticity of demand. This provides an estimate of the percentage
of sales that would be lost by product A upon a one percent increase in the price of product B.
All other things equal, the higher the elasticity, the stronger the complementarity. Note that the
cross price elasticity is in most cases not greater than the own-price elasticity of demand. This is
intuitive, since demand for a good is generally more sensitive to the price of that good than to
the price of a complementary good.
Organisation of this section
In the rest of this section, we consider each class of efficiency in turn. Section 4.2 deals with
pricing inefficiencies; Section 4.3 considers improvements of productive efficiency; Section 4.4
considers how benefits can arise through overcoming horizontal externalities; and, finally,
Section 4.5 considers how the inefficiencies arising from incomplete contracts and transaction
costs can be resolved through conglomerate mergers.
Several of the efficiencies created by conglomerate mergers have at their source the same
economic mechanism that gives rise to analogous “vertical” efficiencies. In general, the reason
for this is that vertically related firms indeed produce complementary goods or services: e.g.
retailing is complementary to manufacturing (see Section 2.1).
Efficiency sources that are common to both vertical and conglomerate mergers are mainly
described in Section 3 of this report. These include the Cournot effect (which is the
conglomerate analogue of double marginalisation), managerial efficiency, internalisation of R&D
would denote it as tying if the customer can choose any amount of product B; indeed, with tying the customer may even
ignore its future requirement of product B.
Page 76
benefits, moral hazard and the hold-up problem. In these cases, the present sections focus on
the differences to the vertical analogue, or the different interpretations of the theory, referring to
the section on vertical efficiencies for a detailed description of the economic models.
The remaining efficiencies, instead, are specific to conglomerate mergers, and are discussed in
full. For each efficiency, we consider how a non-integrated environment can produce suboptimal
results, and show how conglomerate mergers can improve the outcome. We also describe the
situations in which the efficiency gain occurs, highlighting the factors which affect the size of the
benefit, and considering whether the results hold under more general scenarios. In addition, we
consider whether the benefits could be achieved without resorting to full integration: the
motivation for integration will be particularly strong where the efficiencies cannot be easily
achieved by other means.
4.2
Increasing Pricing Efficiency
This section is concerned with pricing inefficiencies that can be overcome by conglomerate
integration. The three sources of efficiency gains discussed in this section are the following.
•
Avoiding the Cournot effect is described in Section 4.2.1, and is relevant to mergers
of complementary products. It is also relevant to neighbouring unrelated products
when these are regarded as complements at a certain level of the supply chain.148
The economic externality between complementary products that gives rise to the
Cournot effect mirrors the vertical externality leading to double marginalisation
problems. Therefore, this section relies heavily on the discussion of vertical
industries presented in Section 3.2.1.
•
Price discrimination (metering) strategies can sometimes be implemented following
complementary mergers, as explained in Section 4.2.2.
•
Reducing customer heterogeneity, which is sometimes described as a form of price
discrimination, is discussed in Section 4.2.3. It mainly concerns unrelated products.
A conglomerate merger can bring about efficiency gains related to metering or to reducing
customer heterogeneity to the extent that it facilitates bundling and tying. In the case of the
Cournot effect bundling or tying are not necessary to realise efficiency gains. However, these
practices may increase the magnitude of the gains.
148
For example, gin and whisky are mainly substitutes for the final consumer, but can be considered as complements by
retailers (see the Guinness/Grand Metropolitan case study in Section 5).
Page 77
Can the efficiency be achieved through other means?
In principle, it is not necessary for two firms to merge in order to offer customers a bundle.
Various kinds of contractual arrangements can achieve the same result. However, where
contracts induce significant transaction costs, monitoring costs and/or present the risk of
opportunistic behaviour by one of the parties, a merger may be the most efficient mean to
realise the gain. Further, a merger may bring about additional efficiencies, such as production
efficiencies, that could not be realised otherwise.
4.2.1
Avoiding the Cournot effect
Source of inefficiency
Section 3.2.1 discusses the “double marginalisation” effect in the context of vertical integration.
A very similar source of inefficiency arises in the case of complementary goods. This is due to
the economic similarity between vertical and complementary industry structures (see Section
2.1). In particular, as explained in the next paragraph, the same sort of externality that leads to
the double marginalisation problem also exists between suppliers of complementary goods.
This source of inefficiency in the pricing of complements was identified by Augustin Cournot in
1838. Cournot considered two complementary products that are valuable only when used
together (perfect complements). He gave the example of copper and zinc which are combined
to make brass. In this case two independent producers with some degree of market power
would set inefficiently high prices. In its pricing decision, the zinc producer does not internalise
(i.e., does not take into account) the profit increase of the copper supplier that would result from
a reduction in the price of zinc; similarly for the copper producer. This outcome can be seen as
a double mark-up charged to final consumers for the combined good. Instead of a “vertical”
double mark-up, as discussed in Section 3.2.1, it is a “lateral” double mark-up.
Page 78
Figure 4-2: Complementary monopolies
Cost = CA
Cost = CB
Monopoly A
Monopoly B
PA = CA + MA
PB = CB + MB
Consumers
PA+B = CA + CB + MA + MB
This is illustrated in Figure 4-2, which shows two monopolies, A and B, with marginal production
costs of CA and CB respectively. To this cost, they add a mark-up of MA and MB respectively,
giving the price of the goods, PA and PB. Thus, assuming consumers buy one unit of each good,
the price of the system, PA+B, is the sum of both marginal costs and mark-ups. Analogously to
the double marginalisation effect, the inefficiency stems from the fact that each firm, in its pricing
decision, only takes into account the effect of its price on its own sales, overlooking the negative
impact that a price rise has on sales of the complementary good.149 As a result of this pricing
externality, firms charge (overall) excessive mark-ups.
If the two producers coordinated their decisions, instead, prices would be lower and demand
higher, making both suppliers and consumers better off (see Figure 4-3 and Figure 3-3). This is
shown more formally in Box 4.1 which describes a simple model illustrating the Cournot
effect.150
149
The Cournot effect and double marginalisation are so similar that in many cases the same theoretical models can be
used to describe both.
150
We note that the models employed to describe double marginalisation in Appendix A and Box 3.1 are slightly different
from the one presented in Box 4.1. The disparity relates to different assumptions on the time structure of the model. In
our conglomerate model prices are set simultaneously and rents are split equally, while in the vertical setting one firm
sets prices before the other, giving one firm an advantage which results in an unequal division of the surplus. These
differences, however, are irrelevant to the general economic mechanism that is illustrated.
Page 79
Figure 4-3: Complementary integration and pure bundling
Cost = CA
Cost = CB
Monopoly A
Monopoly B
PA+B = CA + CB + M
Consumers
Clearly, this kind of inefficiency can only arise if both products are sold under imperfect
competition. Similarly to the case of double marginalisation, the smaller the mark-up charged by
the firms, the smaller is the inefficiency (see Figure 3-4). In the limiting case when at least one
product is competitively supplied, and thus is sold at marginal cost, the Cournot effect
disappears. In such situation, a merger would not increase pricing efficiency through the
elimination of the Cournot effect.
This consideration is at the basis of the “one monopoly profit” principle for complementary
markets. A classic “Chicago School” argument, the principle states that the monopolist of good
A cannot earn any additional profits by also monopolising good B, where A and B are perfect
complements (see Section 2.2.3).
Page 80
Box 4.1:
A simple model of the Cournot effect.
In the following we describe a simple theoretical model of the Cournot effect as a source
of inefficiency. We also discuss how a conglomerate merger eliminates the inefficiency
and why alternative contractual agreements with the same purpose may be ineffective.
We consider a hypothetical industry where two monopolists, A and B, produce perfectly
complementary goods, i.e., goods A and B are valuable only when used together in 1:1
proportions (see Figure 4-2). Therefore, we can assume that the demand function for each
good is has the form:
D( p A , p B ) = u − v( p A + p B ) ,
where p A and p B are the prices and u , v > 0 are constant parameters. Assuming
marginal
costs
of
production
are
denoted
by
cA
and
c B , firm A’s profit
Π A = ( p A − c A ) D( p A , p B ) is maximised when the price p A is given by
p A=
1 u
2 v
− 12 ( p B − c A ) .
Note that since goods A and B are only valuable together, consumers’ demand depends
only on the total price of the system; as a result, the optimal price for firm A decreases if
the price of good B is higher. Similarly for firm B. When the two firms set their prices
independently there is only one Nash equilibrium, given by
pA =
1 u
3 v
+ 23 c A − 13 c B ,
pB =
1 u
3 v
+ 23 c B − 13 c A .
This means that if any of the two firms charged a lower price for its good, the other firm
would have an incentive to increase its price. The Nash equilibrium above is the only
configuration of prices where none of the firms is willing to change its behaviour. This
equilibrium, however, is inefficient. The total price of the system good is thus
p = p A + p B = 2u 3v + (c A + c B ) 3 . If the two firms, instead, cooperate to maximise
joint profits, as would happen for a merged firm, the total system price would be
p∗ =
It can be shown that
1 u
2 v
+ 12 (c A + c B ) .
p is always greater than p ∗ (because demand must be non-
negative). Thus, consumer welfare is higher under common ownership than under
separate ownership. Moreover, total profit is higher.
In principle, this pricing inefficiency can also be eliminated through an agreement between
the two firms, whereby they consent to charge prices such that the system price is not
greater than
p ∗ . However, such agreement would incur in two difficulties: (1) deciding
how to split the rent, and (2) monitoring compliance. The first problem arises because
there is an infinity of different ways to set prices
∗
∗
∗
∗
p A and p B such that p A + p B = p ∗ .
The second problem is a consequence of the fact that any such price configuration is not a
Nash equilibrium, so that each firm has an economic incentive to cheat on the agreement
and raise its own price. These two facts make it considerably more difficult to reach and
sustain such kind of pricing agreements.
Page 81
Complementary mergers and bundling
One important difference between vertical and complementary mergers is that, in the case of
complements, mere integration generally brings about only a part of the potential gains when
the complementary components are sold separately: some form of bundling is necessary in
order to reap the full benefits. The need for bundling (or tying) occurs in two cases:
•
When there is some competition in the markets, and customers are free to “mix and
match” a component from the merged firm with another from a competitor (see
Figure 4-4). For example, in the GE/Honeywell merger case, aeroplane
manufacturers were able to purchase avionics from Honeywell and engines from
Rolls Royce. In this case, a hypothetical price reduction of good A would increase
sales of all suppliers of good B. Thus, while part of the benefit would be reaped by
the “B division” of the merged firm, other suppliers of good B would also gain. The
pricing externality is internalised only in part. Clearly, the efficiency gain will be
more important the larger the market shares of the merging parties. In the extreme
case of a merger between two monopolists (and perfectly complementary products)
bundling is not necessary.
•
Bundling is also necessary when individual components are valuable on their own.
If some customers purchase good A but not good B, for example, a generalised
price reduction of good A would benefit these customers without bringing about any
gain to the firm.
In these situations, therefore, the merged firm would choose to sell the individual components at
a higher price, while the price of the package, or bundle, can be discounted with respect to the
sum of pre-merger prices of the two goods. The economic rationale of post-merger prices is
analysed in the following paragraphs.
Effect of conglomerate integration
The case of the integration between monopolists discussed by Cournot is straightforward, and
constitutes a good example of a transaction bringing benefits to all parties (i.e., to firms as well
as consumers).151 However, where there is some competition in both markets, and products are
differentiated, the outcome of the transaction depends on a number of factors. The effects of the
merger on consumer welfare may be ambiguous and depend on the impact on competitors.
These issues are discussed in the following. (See also Section 3.2.1 for a discussion of these
issues in the vertical context.)
To fix ideas, consider the following setup, discussed by Choi (2003).152 There are two
complementary goods, A and B, which are valuable only when used together in 1:1 proportions,
and two differentiated brands of each component, A1, A2, and B1, B2. The structure of the
Page 82
151
This is an example of what economists call a Pareto improvement.
152
Similar set ups are considered by Nalebuff (1999 and 2000), Nalebuff and Lu (2001) and Choi (2001).
industry is illustrated in Figure 4-4 (left hand side). Each consumer has a preference for one of
the four varieties of composite goods (A1B1, A2B2, A1B2, and A2B1). Consumers are
indifferent between the remaining three non-preferred varieties. Initially, all firms are
independent and enjoy the same market share of 50%. This situation is compared to the
equilibrium after the merger of firms A1 and B1. Moreover, this model of complementary
industry provides insights also on vertical mergers and the double marginalisation effect,
discussed in Section 3.2.1. Figure 4-4 illustrates, besides the complementary industry analysed
by Choi, the corresponding “vertical” analogue. The interpretation of the vertical set up is as
follows. Good A, supplied by upstream firms, is used as an input into the production of
downstream firms. Downstream firms produce internally good B, and combine it with good A to
produce the final good. Each downstream firm uses both differentiated inputs, and offers
consumers two varieties of the final goods (so B1 offers A1B1 and A2B1, and B2 offers A1B2
and A2B2).
Figure 4-4: Complementary integration in duopolistic differentiated markets and analogy
with vertical integration
Complementary Structure
Vertical Structure
Pre-Merger
A2
B1
B2
Consumers
A1B2
A2B2
B2
A2B1
B2
A1B1
B1
B1
A2
A2
A1
Products
A1
A1
Product Market B
Products
Product Market A
Consumers
Post Merger
Consumers
A2
B1
B2
A2B2
A1B2
A2B1
A1B1
Products
B2
B2
A2
A2
B1
B1
A1
A1B1
Products
A1
A1
Consumers
Note: Thick/dashed lines indicate post-merger reductions/increases in prices to final consumers
with respect to pre-merger prices (see Choi 2003).
Page 83
The results of the model are sensitive to the details of the demand system and the production
functions. In the rather general case illustrated by Choi (2003) the following outcomes result:
•
Exploiting the efficiency gain due to the elimination of the Cournot effect, the
merged firm sells the bundle at a discount with respect to the sum of pre-merger
prices (i.e., it would engage in mixed bundling). However, single components are
sold at a higher price than before the merger. This is because some of the
customers that the firm would lose as a result of the price increase now buy the
bundle, while pre merger they would have chosen competitors’ products. This
increases the incentive of the merged firm to raise prices of the individual products.
•
Rivals, driven by competitive pressure, reduce their prices, but less than the
merged firm, because they are affected by the double mark-up problem and thus
price inefficiently.
•
As a result of the new prices, the merged firm increases its market shares with
respect to the pre-merger level, while competitors’ market shares decrease.
•
In most cases, the profits of the merged firm increase, while the competitors’ profits
decrease (although this result is less general than the others).
This is intuitive: the model describes a situation where the merging parties become more
efficient than their rivals and as a result make larger profits, while competitors lose. Crucially,
this model is static, i.e., it assumes that rivals cannot react to the merged-firm's mixed bundling
strategy. In the real world, rivals can be expected to respond to the merger with counterstrategies aiming at recovering the lost shares. These include R&D, product repositioning, and
counter-mergers.
The immediate consequence of the merger is therefore to increase the efficiency of the merging
parties, and increase the competitive pressure on rival firms. In Choi’s model, social welfare
decreases if the composite products are close substitutes, while it increases if they compete
weakly with each other (i.e., are highly differentiated). Further social loss may result in the
longer term if competitors are unable to reproduce the efficiency gain obtained by the merging
parties and are marginalised, so that the market power of the merged firm increases
significantly.
Another interesting result arising from Choi's and Nalebuff's papers is that pure bundling or tying
are less profitable strategies than mixed bundling. However, the competitors’ suffer more under
pure bundling. Thus, pure bundling or tying may be more effective in markets where foreclosure
or marginalisation of rivals is possible, while mixed bundling is more likely to be undertaken to
achieve an efficiency gain.
Finally, note that in many situations rival firms have the possibility and the incentive to respond
with a counter-merger. When this occurs, the final outcome is a single oligopolistic market
where all players are integrated. In such markets, all firms internalise the price externality (at
least for the goods that are entirely internally-produced), but the benefits of the increased
Page 84
efficiency are “competed away” and passed on to consumers. All firms end up making smaller
profits than before the “merger wave”. Several theoretical models illustrate this situation as a
typical prisoner’s dilemma.153 In other situations, in contrast, rivals of the merging parties may
find it more profitable to remain non-integrated (although they still make less profit than when all
firms were non-integrated).154 In these cases there is a first-mover advantage, and the firm that
integrates first “wins”, while consumer do not enjoy any additional benefit from rivals’ response.
Nonetheless, Nalebuff notes that even in these situations rivals may still choose to integrate
because of “dynamic considerations”. For example, in markets where competition in R&D is
important, rivals of the merging parties may choose to follow this strategy because they prefer a
“lower but level playing field”. Indeed, since profits also finance R&D investment, rival firms may
“prefer earning 9 in a (9, 10) outcome to earning 10 in a (10, 20) outcome”.155
Impact of conglomerate integration on consumer welfare
The direct effect of the merger on consumer welfare, without taking into account the reaction of
rivals, is generally ambiguous. However, additional benefits to consumers may result from
possible strategic reactions of rival firms, as described above.
Under what conditions the inefficiency occurs
The presence of mark-ups is a necessary precondition to the Cournot effect, and the magnitude
of the inefficiency tends to diminish when gross margins are smaller. However other important
factors can affect the magnitude of the inefficiency, and thus the size of the possible efficiency
gain generated by a merger. These are listed below:
Price discrimination
The models described above assumed that the same unit price is set to all final
customers. For many goods, this is only approximately true. Even consumables such as
toothpaste and toothbrushes, at the retail level, can be offered with volume discounts,
coupons, and other kinds of price discrimination devices. Products such as airplane
engines and avionics are purchased through tenders, implying prices may differ between
consumers, particularly where the suppliers might have some information on customers’
willingness to pay.
Nalebuff and other authors156 remark that when perfect price discrimination is possible,
the actual transaction price only reflects customer’s reservation value, and therefore is
not affected by any pricing externality. Assume, for instance, that the buyer prefers to buy
153
See for example Economides (1993) and Matutes and Regibau (1988); see also Gaudet and Long (1996) in the context
of vertical mergers. However, note that Gaudet and Long find that non-integration can also be an equilibrium when
there is a sufficiently large number of firms in the market.
154
Nalebuff (2000)
155
Nalbuff (2003).
156
Nalebuff and Lu (2001), Nalebuff and Majerus (2003).
Page 85
the product produced solely by the integrated firm (i.e., A1B1 in Figure 4-4), and is willing
to pay a higher price for this than for products involving components of other firms. Then,
in a bidding competition where preferences are observable, the merged firm would set its
price equal to competitors’ marginal costs plus a “preference premium”, and would always
win without the need to bundle its products.157 On the other hand, the buyer may, instead,
like only one of the components produced by the integrated firm, preferring the rivals’
version for the second component. In this case, if the integrated firm were to deny its
customers access to the single components this would reduce its profits. The relevance
of the pricing externality under imperfect information, where perfect price discrimination is
not possible, is still an open question.
Degree of complementarity
Many theoretical models assume that the goods produced by the merging firms are
valueless if not combined together, and are used in fixed proportions. When the goods
are only imperfect complements, the magnitude of the inefficiency tends to be smaller.
More precisely, the smaller the degree of complementarity, the smaller is the
inefficiency.158
Complements have similar values
As in the case of a vertically structured industry discussed in Section 3.2.1, these results
further depend on the relative values that customers attribute to the components; the
more similar the value of the complements is, the larger the efficiency gain tends to be.
Bundle size
A bundle of many components allows larger efficiencies than a bundle of few
components. If the final good is made of n components, each component’s price has an
impact on the sales of the other n-1. Thus, the size of the possible efficiency gain per
component is amplified. This implies that the magnitude of the total gain increases more
than linearly with the number of components.159 Large bundles can be particularly
common in the case of information goods, since the low marginal cost of production
reduces the inefficiency of providing a customer with a good it does not require.160 Again,
the same effect arises in vertically structured industries.
Page 86
157
Unless the merged firm’s total marginal cost is inferior to the premium.
158
Criteria to assess the degree of complementarity are provided in Section 4.1. Note that the cross price elasticity of
demand between two products may also be reduced as a result of competition in one of the two (or both) product
markets. This has also the effect of reducing the mark-up that firms can charge, thus reducing the size of the
inefficiency. In this case, the lower cross price elasticity does not necessarily reduce the inefficiency further: i.e., the two
effects should not be cumulated.
159
See, e.g. Nalebuff (2003).
160
On the rationale for large bundles of information goods see also Bakos and Brynjolfsson (1999), and the discussion of
this paper in Section 4.2.3.
Figure 4-5:
Criteria to assess the relative magnitude of the firm’s efficiency gain
Larger efficiency gain
Smaller efficiency gain
Pre merger: Large mark-ups on both goods
Pre merger: Small mark-ups on both goods
Post merger: The goods are sold as a bundle
Post merger: The goods are not bundled
One unit price for all customers
Sufficient degree of price discrimination
Strong complementarity
Weak complementarity
Complements have the similar values
Complements have very different values
Larger number of complementary inputs
Smaller number of complementary inputs
Finally, we note that even if there are efficiency gains from eliminating the Cournot effect, firms
may choose not to bundle their goods. In particular, if there is imperfect competition and the
price elasticity of total market demand for the system (the final good) is small, rivals may
respond aggressively to price cuts, such that bundling is not a profitable strategy for the
integrated firm. The reason is that the bundle discount has two effects: market size expands,
and the firm’s share increases. If the market is inelastic rivals are induced to price more
aggressively to try and retain market shares, leading to lower equilibrium prices and lower
profits for all players. Although in this case bundling would still be beneficial to customers (at
least to some of them), it is not a likely outcome of the merger. 161
Can the efficiency be achieved through other means?
Eliminating the Cournot effect inefficiency through a contractual agreement between two
independent firms is more complicated than eliminating double marginalisation in a vertical
relationship. As we have seen, firms operating at different levels of a supply chain can choose
between a number of contractual arrangements suitable for eliminating pricing externalities,
although these may not always be the most efficient way to do so. Contractual solutions in a
vertical industry are facilitated by the fact that the firms are already trading partners, with an
established buyer-seller relationship. Therefore the only change required is replacing a constant
unit price with a more sophisticated pricing mechanism (e.g. a franchise fee, RPM, etc), perhaps
with the addition of some exclusivity clause.
Firms producing complementary products, instead, usually do not have any relationship, and
certainly do not have a trading relationship in place: if they did, they would be vertically related
through the creation of a joint distributor or a distribution contract. Instead, they both sell directly
to customers, who are finally responsible for matching the complementary products. This fact
makes contractual alternatives to a merger more difficult to realise. Depending on the
characteristics of the industry, we can generally distinguish two types of solutions.
•
In the cases where a merged firm would choose to bundle the products (as
discussed above), the contractual arrangement would also require bundling. This
161
Nalebuff (2000) obtains this result in a setting similar to Choi (2003).
Page 87
would involve either (a) transforming the complementary relationship into a vertical
one, whereby one of the two firms buys the complementary product and distributes
the bundle; or (b) reaching a common distribution agreement with a third party
(perhaps a joint-venture company), which is in charge of distributing the bundle. In
both cases, the agreement involves the creation of a vertical relationship, and the
contract would in fact be similar to a vertical restraint of the kind described in
Section 3.2.1 on pricing in a vertically structured industry.
•
In the (rare) cases where bundling is not strictly necessary (i.e., there is no
competition and no secondary use), a simple agreement to jointly cut prices can
potentially raise both firms’ profits. This effect is illustrated in Box 4.1. However, as
discussed in the box, these lower prices would not be stable, in the sense that each
firm would have an incentive to cheat on the agreement and raise its price,
effectively appropriating some of its partner’s profits. This implies that for such
agreement to be sustainable in the long term, some mechanism has to be devised
to monitor compliance and punish deviations. In practice, the difficulties are similar
to those of sustaining a cartel agreement, except that the goal is to reduce prices
instead of raising them.
Finally, in both cases, the firms would have to agree on how to split the revenues (this difficulty
is also illustrated in Box 4.1). This may be particularly difficult if there is uncertainty regarding
production or distribution costs and/or uncertainty regarding demand.
4.2.2
Price discrimination (metering)
Source of inefficiency
When a firm with some degree of market power charges one price to all its customers, it faces
the basic problem that consumers have different valuations for the good. Any possible price
inevitably loses some customers (those with a lower valuation) and also sheds some profits –
since high valuation customers are willing to pay more. This is a source of inefficiency, and
leads to a reduction of social welfare: if each customer could be charged exactly his valuation
for the good, output would be expanded by sales to low valuation customers.162 Clearly, the
problem is more severe the wider the range of valuations. 163
Effect of conglomerate integration
Firms try and minimise this inefficiency through various forms of price discrimination. Bundling
or tying complementary goods can be a very effective tool for discriminating between
Page 88
162
However, high value customers would pay more so that the overall impact on consumer welfare is, in general,
ambiguous.
163
For example, the welfare loss will be larger if customers’ valuations are uniformly distributed between €40 and €160
rather than between €90 and €110.
customers, and since bundling strategies are more easily implemented if all goods are sold by
the same firm, conglomerate integration can be undertaken in order to achieve price
discrimination. We note that bundles of unrelated (or weakly complementary) goods can also be
sold with the same goal of reducing the pricing inefficiency described above. However, in the
case of unrelated goods, the economic mechanism is different form the one described in this
section, and it is therefore examined separately in Section 4.2.3.
In very general terms, tying complements facilitates price discrimination because it increases
the firm’s information on customers’ valuation. The strategy consists in using one good to
“meter” the use of the complementary good, and make consumers effectively pay a “per use”
fee. Accordingly, this practice is also known as “metering”. This is the case, for example, of
copier machines and toner cartridges. In order to maximise profits, copier manufacturers would
like to charge a high price to intensive users, who tend to value the machine more, but a lower
price to infrequent users, who tend to value the machine less.164 This can be achieved by using
the toner cartridges to meter use; by increasing the price of toner cartridges and decreasing the
price of the copier, the total price to intensive users goes up, while the price to infrequent users
goes down.
The necessary conditions for implementing metering are the following: (1) the goods are
valueless if not used together (i.e., no secondary use); (2) the goods are not used in fixed
proportions (and consumers’ valuation for the system is proportional to the ratio); and (3) the
products are tied together, technically or through other means, in such a way that customers
cannot “mix-and-match” components from different manufacturers.
Impact of conglomerate integration on consumer welfare
Assessing the impact of price discrimination on consumer welfare is a highly non-trivial
exercise, and presenting a complete account of the relevant issues would go beyond the scope
of this report. Nonetheless, we summarise below some of aspects of the current debate. The
theoretical benchmark is perfect (i.e., first degree) price discrimination in a monopolistic market.
In such conditions, social welfare increases, the firm’s profits rise, while the effect on consumer
welfare may be ambiguous.
Generally, low valuation customers gain, while high valuation customers lose. A necessary (but
not sufficient) condition for consumer welfare to increase, is that output must increase.165 Social
welfare, instead, may increase even if output decreases.166
In addition to the “redistributive” effect, whose total impact on consumers may be ambiguous,
there is also a net effect, as the firm’s profits increase.167 This outcome can be desirable if
164
The different valuations may not be fully accounted by the fact that an intensive user will have to purchase a new
machine sooner than the infrequent user.
165
See, for example, Varian (1989), Stole (2003).
166
For example, this kind of result holds in the model of Chen (1999).
167
Note that total output will also normally increase.
Page 89
market power is indeed necessary in order to recoup high fixed or sunk costs, such as R&D.
Economists recognise that price discrimination allows the maximisation of profits at the lowest
possible cost to consumers as a whole and, as such, price discrimination can efficiently promote
innovation and other kinds of investments, bringing further benefits to consumers in the longer
term.168 This is also recognised by antitrust authorities, where many regulated monopolies are
allowed some kind of price discrimination.169
In concrete cases, price discrimination is never perfect, making the assessment even less
straightforward. Further, competition from imperfect substitutes, or even from inferior goods, can
significantly change the overall impact on social and consumer welfare. For example, it can be
shown that, in differentiated products industries, as a result of competition consumers may
obtain considerable residual surplus. In certain circumstances, industry profits may even be
lower when all firms price discriminate. 170
We also note that, contrary to a common prejudice, consumers are likely to enjoy some surplus
even under perfect price discrimination in a monopolistic market, owing to general equilibrium
considerations. This is shown by the following example. Consider a TV sets monopolist
engaging in perfect price discrimination. Assume further that consumers can also buy other,
unrelated products which are sold at a uniform unit price, e.g. holidays. Most consumers enjoy
some surplus on the purchase of these other products, which contributes to their welfare.
Therefore, utility-maximising consumers would choose to buy holidays instead of TV sets, up to
the point where the surplus they enjoy from the two products is the same.
On the negative side, instead, several authors point out that the reward to firms from achieving
price discrimination can be so large to induce them to undertake socially wasteful activities.171
Similarly, a loss of social welfare may result from the behaviour of consumers, and in particular
of high-valuation consumers, seeking to avoid price discrimination. Nalebuff (2003) analyses
these issues, and presents a detailed account of the (wasteful) stratagems employed by
manufacturers of copier machines to protect their “metering” strategy.
Can the efficiency be achieved through other means?
Alternative means suitable to realise this efficiency gain may be available in certain
circumstances. These include the following:
•
Directly metering use can sometimes be possible. For example, a copier machine
may be leased with a “per copy” fee; some car insurances offer “per kilometre”
contracts.
168
Page 90
Hausman and MacKie-Mason (1988), Farrell and Weiser (2003). Albeit in a different context, Rey and Tirole (2003)
argue that antitrust authorities may consider refraining from prosecuting certain firms’ behaviours because the extra
profits thus obtained compensate the firm for its investment (Rey and Tirole, 2003, Section 5, “Forbearance as a reward
to investment or innovation”).
169
E.g. telephone companies, railways, etc.
170
Stole (2003), and references therein.
171
See for example Farrell and Weiser (2003), also discussed in Section 2.2.3.
•
4.2.3
Often, price discrimination devices other than metering are available, and can
employed in selling individual components. In that case tying is not necessary.
However, in these cases, the question arises of whether the use of metering, either
as an additional device, or as an alternative, can further increase efficiency.
Reducing customer heterogeneity
As described in the previous section, in the presence of market power, flat pricing (i.e., applying
the same mark-up to all customers) is generally inefficient. Firms may try to avoid this
inefficiency by metering use, therefore setting different prices to different groups of customers
(see Section 4.2.2). Alternatively, firms may bundle unrelated goods in order to reduce customer
heterogeneity, thus increasing the efficiency of flat pricing. The latter method is the topic of this
section.
Effect of conglomerate integration
Bundling different products may reduce the pricing inefficiency by reducing the heterogeneity of
customers’ valuation. In general, the valuation for the bundle will vary less from customer to
customer than the valuation for the individual goods (in relative terms).172 This is most easily
demonstrated through an example, such as that provided by Brandenburger and Krishna
(1990). Consider the problem of optimal pricing of tickets for a hypothetical concert hall, and
assume there are four types of concert goers: Romantic, Neo-classical, Tchaikovsky lover and
Sophisticate. There are two concerts, Berlioz/Tchaikovsky and Bartok/Stravinsky. Each fan has
different valuations for concerts, as shown in Figure 4-6.
Figure 4-6: Valuations of concerts
Customer type
Berlioz/Tchaikovsky
Bartok/Stravinsky
Both concerts
Romantic
40
20
60
Neo-Classical
20
40
60
Tchaikovsky lover
45
5
50
Sophisticate
5
45
50
Further, assume that marginal cost is zero. If concerts are sold separately, the optimal price is
40; at this price, the Romantic and Tchaikovsky lover buy a ticket for the first concert, and the
Neo-Classical and Sophisticate buy a ticket for the second, and total profits are 160.173 If the
concerts are bundled together, and sold at a price of 50, all consumers buy the combined ticket
(as their combined valuation of both concerts is higher than or equal to 50), and the concert
172
This effect was first noticed by Stigler (1963), and further developed by Adams and Yellen (1976), Salinger (1995), and
other authors.
173
A rise in price to 45 for example increases the surplus gained from the Tchaikovsky lover and Sophisticate, but profits
are decreased as this is more than offset in by the loss in sales to the Romantic and Neo-Classical; a decrease in price
to say 20, increases ticket sales, but profits decrease as this is not compensated for by the decrease in margin.
Page 91
provider makes a profit of 200. Since there is less difference between consumers’ valuation of
the total package, the pricing inefficiency is reduced. In this example, total consumer surplus
(i.e., the difference between the consumers’ valuation and what is paid for the ticket) is also
increased, although some consumers gain at the expense of others.174 More generally, the
effect on consumer welfare is ambiguous but social welfare (i.e., the sum of consumer welfare
and the firm’s profits) will always be higher. However, even when the (static) consumer surplus
is decreased, the resulting increase in profits may promote dynamic efficiency (e.g. innovation),
thus bringing benefits to consumers in the longer term. This latter argument is very similar to the
one made for price discrimination (see Section 3.2.4).
Under what conditions the efficiency occurs
Efficiency gains increase in the difference between consumers’ valuation of the independent
goods. In the example above, customers’ valuations for the two goods are negatively correlated:
customers who have a high valuation for one concert tend to have a low valuation for the other.
In such cases the benefits of a packaged offer are most apparent, as the differences in the
valuations “balance each other out” when the goods are combined, reducing the extent of
heterogeneity. However, bundling is also profitable for a monopoly when customers have
independent valuations for the two goods, or even weakly positively correlated.175 Thus, the
result is very general: for any two goods whose valuations are independent, e.g. pens and beer
or clothes and bananas, there will be less variation in the valuation of the bundled good, and a
monopolist can therefore increase its profits by offering the bundle at a discounted price with
respect to the optimal monopoly price of the two goods.176 Further, a bundle of three or more
goods is even more profitable. However, when consumers’ valuations are aligned (i.e., strongly
positively correlated) – as is the case for goods that are strong complements – bundling is less
effective. This is because if consumers’ valuations of the two goods are strongly correlated, the
valuation of the bundles will still vary strongly from one consumer to the other.177
Another factor affecting the size of the potential gain in efficiency and in profits is the marginal
cost of production. If marginal cost is large relative to customers’ average valuation, then
bundling is less profitable, and less socially efficient. For example, when marginal cost for a
good is larger than the average valuation, the supplier wants to increase rather than decrease
the dispersion of customers’ valuations. Indeed, if all customers have the same valuation (equal
to the average valuation), the producer would make losses, while if valuations are dispersed,
some customers may be willing to pay more than the marginal cost.
Therefore, we are more likely to see bundles where goods have low marginal cost. Information
goods, for example, usually have marginal cost close to zero. These include application
Page 92
174
With a single price of 40, the Tchaikovsky lover and Sophisticate both have a surplus of 5, while the Romantic and NeoClassical consumers have no surplus; with the bundled price, the Romantic and Neo-Classical consumers have a
surplus of 10, while the remaining two consumers have no surplus.
175
Schmalensee (1984); McAfee, McMillan and Whinston (1989).
176
See Nalebuff (2003), p.36.
177
For example, suppose concert goers have exactly the same valuation for Berlioz/Tchaikovsky as for Bartok/Stravinsky,
then the optimal price for the bundle is just twice the optimal price for the individual components, and there is no
efficiency gain.
software, news, music files, etc, in particular when purchased and delivered on the Internet. As
pointed out by Bakos and Brynjolfsson (2000) and Nalebuff (2003), such goods are often sold in
bundles. MS Windows operating system now includes a collection of applications that used to
be sold as independent products. Websites providing news usually charge a fixed fee for the
access to the entire set of hundreds or thousands of articles. On cable television, pay-per-view
is a relatively minor phenomenon: most common are “bundles” of programs, i.e., packages of
one or more channels.
Summarising, the factors affecting the size of the gain in profits and in social welfare that can be
obtained through bundling are summarised below:
4.3
4.3.1
•
The supplier has market power. All models presented in the literature assume that
the supplier is a monopolist. However, bundling is likely to be profitable also when
competing suppliers exist, but the products are sufficiently differentiated (i.e., there
is monopolistic competition).
•
Bundling is more profitable the larger the dispersion of customers’ valuations
(except in the case when marginal costs are large in relation to average customers
valuations, as explained above).
•
Bundling is more profitable when customers’ valuations of the two goods are more
negatively correlated. Gains tend to disappear as soon as the correlation is more
than weakly positive. Therefore, bundling tends to be unprofitable for
complementary goods.178
•
Gains increase when the number of bundled goods increases.
•
Bundling is less profitable the larger the marginal cost of production relative to the
average valuation.
Increasing Productive Efficiency
Economies of scope and scale
Source of efficiency
A merger leads to “economies of scope” if it leads to a reduction of the average unit cost of
production by combining the production of different goods. This is opposed to “economies of
178
However, bundles of complements may be profitable (and socially efficient) for other reasons, e.g. elimination of double
marginalisation (see Section 4.2.1).
Page 93
scale”, whereby a reduction in average unit cost is a consequence of increasing output of the
same good.179
In principle, only economies of scope are relevant to conglomerate mergers; economies of scale
result from increasing the amount of the same activity, and are therefore more closely linked
with horizontal mergers. Nevertheless, to the extent that all firms are vertically integrated to
some extent, a conglomerate merger between firms which are primarily active in different
markets may also include some horizontal overlap in secondary upstream or downstream
activities, allowing the merged firm to capture economies of scale.
Figure 4-7: Horizontal overlaps in a conglomerate merger
Firm A
Firm B
Production A
Production B
Admin.
Admin.
Distribution
Distribution
Product A
Product B
Consider, for example, cost reductions arising from shared overheads. Two manufacturers of
different chemical products may produce different products, but require similar administration
and distribution processes, as shown in Figure 4-7. A merger between two manufacturers of
different chemical products may allow them to use the same sales force to market both products
and to share these costs, as depicted in Figure 4-8. Strictly speaking, these are economies of
scale arising from the fact that both companies are vertically integrated into sales and
administration; the companies could, in principle, obtain these services from specialised
accountancy, sales and management companies. However, since most firms undertake these
activities themselves, in practice, such economies often arise from non-horizontal mergers.
Similarly, where the parties also distribute their products, a merger may give rise to economies
of scale in distribution.180
Page 94
179
The output increase can be the result of integrating the production of two previously separate firms.
180
Note that economies of scale in administration and distribution usually result in different types of efficiency gains:
distribution is generally part of the marginal cost, so that integration may lead to lower prices; on the contrary,
administration overheads are normally fixed costs that do not have a direct impact on prices.
Figure 4-8: Economies of scale/scope in a conglomerate merger
Merged Firm
Production A
Production B
Admin.
Distribution
Product A
Product B
Under what conditions the efficiency occurs
Such benefits are often cited by merging parties as a motivation for the deal. Indeed, economies
of scope and scale have the potential to result in lower marginal cost of production, and in that
case they can generally be expected to lead to lower prices in the short term. Further, the
potential gains are usually verifiable, and the merging parties may have good estimates of these
types of cost savings.
However, such efficiency gains are not always merger-specific: often they could be realised
through contractual means. Therefore, economies of scale and scope should be regarded as
gains from the merger only in the cases where such alternatives are not viable. These include
the following situations:
•
Specialised indivisible physical assets. Assets that are indivisible are very likely to
be incompletely utilised in the production of a single good. The spare capacity of
the asset can then be used for a different product. However, the specialised nature
of the asset implies that contracts designed for sharing the resource would be
prone to opportunistic behaviour (see Section 3.5.2). This might motivate a
merger.181 For example, a conglomerate merger of a car and a motorcycle
manufacturer may allow the firms to share a wind tunnel for aerodynamics testing,
increasing the usage of that asset.
•
Specialised human capital. The typical example of economies of scope resulting
from specialised human capital is R&D activity: personnel engaged in such work
are usually able to exploit their knowledge and experience in different (but related)
markets. For example, an electronic engineer that designs music amplifiers might
181
Teece (1980, 1982); see also the hold-up problem in Section 3.5.2.
Page 95
be a good designer of CD players; pharmaceutical researchers can and do engage
in research on drugs for different pathologies. However, economies of scope in the
use of human capital are not limited to R&D: marketing and management activities
may also involve specialised knowledge that could be used for different products.182
•
Immaterial assets. These include brand names, reputation, etc. For example, it
could be argued that one of the major advantages of the Virgin group’s
conglomerate expansion is that it can transfer an established brand name into each
new market that it enters.
Impact of conglomerate integration on consumer welfare
The efficiency gains resulting from economies of scope can take the form of lower marginal
costs and/or of lower fixed costs of production. Both kinds of cost reductions may be passed on
to consumers in the form of lower prices, although this is more likely to happen in the case of
marginal cost savings (see Section 2.2.3). Alternatively, fixed cost savings can increase the
profitability of investments, thus leading the firm to increase investment level. This may happen,
for example, if the parties can reduce the cost of innovation by pooling together their R&D
departments. The resulting benefits to consumers may take the form of higher product quality
and more frequent innovation (see also Section 3.3.1).
Can the efficiency be achieved through other means?
In all the cases described above, the nature of the asset hinders an effective exchange of the
asset through market contracts. In principle, specialised assets can be leased; human capital
shared through secondment; and brand names franchised. But such contracts are usually
difficult to specify and to enforce, opening the possibility for opportunistic behaviour, leakage of
know-how, etc. In such cases, therefore, economists recognise that a merger can create true
efficiency gains.183
One possible critique, however, is that in many cases a firm could achieve the same gain
unilaterally, i.e., without a merger or other contractual means. This argument has been raised by
Farrell and Shapiro (1990, 2001) in the context of horizontal mergers, and in particular for
economies of scale. These authors demonstrated that in an industry characterised by
economies of scale and competition in quantities (à la Cournot), prices rise post merger, unless
integration between the parties generates efficiency gains beyond mere economies of scale.
Further, as these authors remark, the larger the potential gain from economies of scale, the
stronger the incentive for the firm to expand unilaterally, even without a merger. This is likely to
Page 96
182
Teece (1980, 1982).
183
Teece (1980, 1982), Farrell and Shapiro (2001).
result in excess capacity, thus increasing the intensity of competition and generally bringing
larger benefits to the consumers.184
This critique is generally less applicable to conglomerate mergers than to horizontal mergers.
Indeed, unilateral expansion to a complementary or unrelated market where the firm was not
previously involved is likely to require the acquisition of specific skills and assets. This can make
the expansion to the new market costly, lengthy, and uncertain. Therefore, unilateral expansion
should be generally considered as a less realistic alternative to a full merger in the case of nonhorizontal integration.
4.3.2
Financial efficiency
Conglomerate mergers may enable firms to diversify risk, making them less financially
vulnerable to possible downturns in one of their lines of business. This is particularly important
for firms operating in markets characterised by high volatility of prices or by long term demand
cycles which are hard to predict (e.g. commercial aviation). A good example of such
diversification is a manufacturer of heating equipment which enters the market for air
conditioning equipment in order to reduce the risk of exposure to unusually hot or cold years.
Even where it is not possible for firms to offset profits precisely, a firm may diversify into an
unrelated activity. An example of this might be Gillette’s diversification into men’s toiletries,
which made it less reliant on its position in the razor market where it was particularly vulnerable
to advances by its main rival, Wilkinson Sword.185
By stabilising its cash flow in this way, a conglomerate firm reduces the risk of bankruptcy. As a
result, it may be able to borrow at a lower interest rate.
In economic terms, the reasoning for this goes as follows. The cost to a firm of external
financing (e.g. debt) is in practice always higher than the cost of internal financing.186 This extra
price, or premium, is due to the agency costs: the lender has imperfect information on the actual
revenues accrued to the borrower, and bears the risk of opportunistic behaviour. This moralhazard problem affects any borrower-lender relationship. However, when a firm’s balance sheet
is distressed, and/or when the firm’s future cash flow is highly uncertain, such extra cost may
become excessively high. As a result, the firm may refrain from investing in an otherwise
valuable project; or, in extreme cases, the lender may simply refuse to provide the funds. A
diversifying merger can allow a firm to reduce the cost of external finance by strengthening its
balance sheet and stabilising its cash flow.187
184
This argument was used by the Court in FTC vs. Cardinal Health Inc., 12 F. Supp. 2d 34 (D.D.C. 1998).
185
Lipczynski and Wilson (2001).
186
This efficiency stems directly from well known imperfections of the capital market, due to asymmetric information and
moral hazard problems between borrower and lender. If lenders faced no costs associated with monitoring the
borrowers, and if complete contracts were possible, the risk of bankruptcy could be perfectly insured by financial
markets. In fact, for a bank, it would be equivalent to lend capital to two independent firms instead of a single, merged
firm, since the risk would be the same.
187
See, e.g. Diamond (1984), Bernake and Gertler (1986, 1987), Pautler (2001), Hubbard (1998), and Hubbard and Palia
(1999).
Page 97
A recent paper by Inderst and Müller (2003) shows that conglomerate integration, by allowing
the firm to pool the cash flows generated by different activities, can reduce the agency costs of
borrowing, thus increasing efficiency and allowing the continuation of valuable projects.
However, the authors also show that such centralised borrowing can be inefficient if it allows
firms to self-finance some of their projects. A lender is better protected against moral hazard if
the borrower needs a constant flow of external funds. When, instead, a firm is financially
independent, at least in part, the lender would adopt a structure of repayments that is less
efficient, thereby raising the total cost of funds. Nevertheless, we note that a merged firm may
be able to choose the most efficient organisation, while independent firms can not. Indeed, a
conglomerate firm that has reached the size where centralised borrowing is inefficient, can
usually organise itself in financially independent divisions and negotiate debt at the division level
(decentralised borrowing).
Impact of conglomerate integration on consumer welfare
Since agency costs are socially wasteful, a merger achieving greater financial efficiency
increases social welfare. Further, the reduced cost of funds can increase consumer welfare
through one of the following mechanisms:
•
if the efficiency gain allows an investment that would otherwise be deterred, social
and consumer welfare generally increase;
•
if fixed costs are reduced, and output expands, prices may fall (see Section 2.2.3).
•
if the marginal cost is reduced (i.e., where capital is an input in production), prices
are likely to fall (see Section 2.2.3).
Under what conditions the efficiency occurs
To summarise, the inefficiency is likely to be larger in the following cases:
•
when cash flow is highly uncertain (e.g. because of price, cost, and/or demand
volatility);
•
when cash flow is hard to measure, in the sense of “incomplete contracts” theory
described in Section 3.5.1 (i.e., it cannot be verified by a court);
•
when the liquidation value of one of the firms is low relatively to the value of its
assets in “full use”; this may happen because the productive assets are highly
specialised and non-redeployable (i.e., few prospective buyers in case of
liquidation);188
188
Page 98
For example, a steel plant is non-redeployable because it can only be used to manufacture steel. In case of insolvency,
it can only be sold to another steel manufacturer, and the small number of prospective buyers may lower the liquidation
value of the asset. On the contrary, a clay tennis court is a specialised asset (it can only be used for playing tennis), but
•
when the firms’ activities require significant and sustained investment.
The efficiency gain is likely to be larger when the cash flows from the merging parties’ activities
are negatively correlated,189 or at least uncorrelated, and expected cash flows are of similar
magnitude.
4.3.3
Countervailing buying power
Source of efficiency
Conglomerate mergers can give rise to increased buyer power where the merging parties
employ the same input. For example, the merger can lead to larger volume-discounts, and
thereby reduce costs. Gains may be particularly substantial where purchases are made through
tenders: bidding firms tend to compete more intensely when tenders are infrequent and highly
valuable.
Such increases in buyer power always benefit the firm. Where downstream competition ensures
that costs reductions are passed on to consumers in the form of lower end prices, increased
buyer power may be considered as a benefit of the merger by antitrust authorities concerned
with consumer welfare. However, where the buyer achieves monopsony power, consumers can
be harmed. This is because the monopsonist may have the incentive to reduce its input
purchases in order to drive the input price down. By doing this, it may have to reduce output too,
thus harming consumers. However, where the buyer power counters market power upstream,
buyer power may be beneficial. In general, the impact of integration on consumer welfare is
therefore ambiguous, and depends on the details of the industry.190
Can the efficiency be achieved by other means?
In general, countervailing buyer power can be achieved without a merger through common
procurement. For example, this is common practice for US private hospitals. By unifying their
purchases in a single tender for a long term exclusivity contract, these firms can obtain large
price reductions from pharmaceutical companies and other suppliers. The introduction of such
practices is one of the reasons for the decrease of average prices of certain pharmaceuticals in
the US during the second half of the 1990’s. However, such joint purchasing practices can also
facilitate anti-competitive practices in the downstream market. In particular, by increasing the
exchange of information, joint purchasing agreements can help companies to establish and
maintain collusive agreements.191
it is re-deployable, since the largest part of the asset value is constituted by the land, which can be used for urban
development or agriculture.
189
I.e., one activity’s profits are large when the other’s are small, and vice-versa.
190
For a detailed discussion see Dobson, Waterson and Chu (1998).
191
For example, information on future output and prices can be inferred from the knowledge of input purchases; see, e.g.
Kühn (2001) and European Commission (2001), §§115-138.
Page 99
4.3.4
Managerial efficiency
Conglomerate mergers can result in increased managerial efficiency. The economic rationale
for fostering the market for corporate control is exactly the same as for vertical mergers, and is
discussed in Section 3.3.4 of this report.
4.4
Preventing Profit Expropriation
4.4.1
Quality standardisation
When a system good is composed of two individually sold components, the demand for each
component also depends on the quality of the complementary product. This may constitute a
problem if consumers only observe the quality of the system good, so that in case of bad
performance of the system they are unable to assess which component is responsible. In such
cases, the manufacturer of the good-quality component is harmed by the behaviour of the other.
Therefore, in order to prevent these problems, it may seek to gain control over the supply of the
complementary component.
This pro-competitive argument was presented by the defence in the HILTI AG case.192 HILTI
was a major manufacturer of nail guns and cartridges for nail guns. In 1982, it was accused of
illegally tying cartridges and nails for nail guns, thus harming competition. According to the
defence, the reason for tying was to ensure quality and safety to the final user. It feared that low
cost, low quality nails produced by other firms could undermine the value of his nail cartridges
and guns and, in case of accidents, even lead to claims for damages.193
A firm can use several means to ensure that consumers only employ quality complements. First,
the firm may engage in pure bundling, whereby the two products are only sold together, as
HILTI did. Secondly, it may employ technical tying, if the technical interface between the
complements can be patented. In this case, the firm does not need to be a supplier of the
product, since in principle it may earn the same profits by licensing the patent (see, however,
the commitment problem discussed in Section 2.3.3). This alternative solution, however, may
require monitoring the quality of the product supplied by the licensees. Where such monitoring
is costly, a merger can be more efficient (see “moral hazard” problems in Section 3.5.1). Finally,
the firm may simply publicly disclaim any responsibility for malfunctioning systems that used
unauthorised components. Such reputation-building strategy, however, may not always be
viable, or effective.194
Page 100
192
European Commission Decision of 22 December 1987 relating to a proceeding under Article 86 of the EEC Treaty
(IV/30.787 and 31.488 — Eurofix-Bauco v. HILTI), and Judgment of the Court of First Instance (Second Chamber) of 12
December 1991 (HILTI AG v. Commission of the European Communities).
193
See also the case study in Nalebuff and Majerus (2003), pp.17-21.
194
For example this strategy would be less effective for established products than for new products; it may involve
significant advertisement costs; finally, despite the protection it would provide against legal claims, it may be insufficient
to protect the brand’s reputation.
Thus a merger is not strictly necessary in order to ensure quality of a complementary product,
but it may often be the fastest or most effective way to achieve the goal. Further, where a
merger would raise significant anti-competitive concerns, the alternative strategies described
above may not be sufficient to remove them.
4.4.2
Internalisation of R&D benefits: horizontal spillovers
As discussed in Section 3.4.2 in the context of vertical mergers, R&D activity can easily result in
knowledge spillovers towards competing firms. In that case competitors can effectively reap part
of the benefits from another firms’ investment in innovation. When there is a risk of such
outcome, R&D investment may be deterred, to the detriment of social welfare.
A complementary merger can reduce the risk of horizontal spillovers in two ways. First,
horizontal spillovers can be the consequence of an information exchange between two firms
that are cooperating in R&D or other activity, if one of the parties leaks the information to third
party competitors. For example, assume that products A and B are complements. Firms Alpha
and Gamma manufacture product A, while firm Beta is a supplier of good B. Being high tech
products, the design of A and B requires a significant amount of collaboration between the
respective manufacturers in order to improve compatibility and enhance performance of the
overall system. However, Alpha fears that its collaboration with Beta may lead to confidential
information being leaked to Gamma, which also cooperates with Beta. Since there is a risk that
Alpha’s investment in R&D might partly benefit its competitor, Alpha might reduce its investment
in valuable innovation. A possible solution to this problem is for Alpha to merge with Beta.
However, while this might reduce the risk of spillovers from Alpha to Gamma, in could also
increase Gamma’s fears of spillovers towards Alpha. This mechanism is essentially identical to
the one described in the context of vertical mergers (see Section 3.4.2). Note that, in general,
information exchange is often required in conglomerate relationships. In particular, cooperation
can be essential in rapidly innovating markets, where time-to-market advantage is crucial:
innovating computer hardware can be valueless if complementary components are not available
soon enough, allowing competitors the time to catch up. (The analysis of the economic
mechanisms underlying the efficiency gains from improved coordination of R&D activities is
fundamentally similar to the vertical case, which is examined in Section 3.3.3.)
Secondly, a merger may prevent horizontal spillovers by allowing a firm to embed its innovation
into a larger system, thereby making reverse engineering more difficult. In this case, the
merging parties may not be engaged in any cooperation prior to the merger.
The impact of a complementary merger of the incentives of the merging parties to invest in R&D
as well as on the incentives of non-integrated firms are essentially the same as in the case of
vertical mergers. A detailed analysis of these effects is presented in Sections 3.4.2 and 3.3.3.
Page 101
4.4.3
Internalisation of R&D benefits: complementary spillovers
Producers of complementary goods are linked by a common interest, even where they are
completely independent entities (see Section 2.2.2). For example, a faster computer makes all
the relative application software run faster, therefore increasing the value of the software to the
consumer. Analogously, in the medical diagnostic imaging industry, improved contrast media
have increased the value of imaging equipment, by enhancing the image quality and range of
applicability of any given machine.195 When the supplier of the complementary good is
independent and has sufficient market power to price at a mark-up, the innovator fails to
appropriate parts of the benefits from its investment. Foreseeing this result, a firm would
rationally invest in innovation less than the socially optimal amount.
This “link” existing between suppliers of complements is dubbed, in economic terms, a “positive
externality”. In this case, it is fundamentally similar to the vertical positive externality between
firms at different levels of a supply chain, which is considered in Section 3.4.3. As discussed in
that section, the inefficiency is likely to be more important when:
•
the products are stronger complements;196
•
suppliers of the complement have larger gross margins.
In addition to the above conditions, the efficiency gain is also likely to be larger when:
•
post merger, the market share of the independent producers of complementary
goods is smaller;
•
post merger, independent producers of complementary goods have smaller gross
margins.
Note that the inefficiencies related to horizontal and complementary spillovers, discussed in this
and in the previous section, are fundamentally different, even if they can lead to similar
outcomes. First, horizontal spillovers benefit competitors, while complementary spillovers
benefit producers of complementary goods. Secondly, the economic mechanism is different.
Horizontal spillovers arise when rival firms obtain confidential information through collaboration
with the same supplier of the complementary product. Complementary spillovers, instead, are a
direct consequence of the externality that characterise all complementary goods (as described
in Section 2.2.2), and does not require any exchange of information or other interaction between
the firms. Indeed, since the economic mechanism underlying these two inefficiencies is
different, these effects may add up.
Page 102
195
See case No. M.3304 - GE/Amersham.
196
The criteria to assess the degree of complementarity between two products are described in Section 4.1.
4.5
4.5.1
Incomplete Contracts and Transaction Costs
Incomplete contracts and moral hazard
The issue of contractual incompleteness and asymmetric information are described in detail in
Section 3.5. As discussed in that section, the implications of such contractual problems for
merger analysis are twofold:
4.5.2
•
In principle, all the benefits of conglomerate integration could be achieved by some
contractual arrangement, however complex. In practice, this is not always the case.
The theory of incomplete contracts explains why certain efficiency gains are indeed
merger-specific.
•
Even in the case where the efficiency can be achieved contractually, such
agreements may involve significant costs for one of the merging parties. These
include the costs of monitoring and enforcing the contract, but also the economic
incentive that has to be provided to the other party in order to prevent opportunistic
behaviour (agency costs). Where such costs are substantial, a merger may bring
about an additional efficiency gain by allowing a reduction of such contractual costs.
One-stop-shopping
Source of efficiency
Conglomerate mergers between related goods may give rise to “one-stop shopping” advantages
for customers. When several different goods or services can be traded with a single transaction,
the possibility arises for saving the costs associated with multiple transactions. These may
include the costs of negotiating, drawing up contracts or search costs. Such costs are usually
either (a) incurred directly by customers, or (b) incurred by the seller and passed on to
customers, at least in part.
A simple example is provided by petrol stations selling newspapers,197 or supermarkets selling
petrol. A retailer may prefer dealing with single wholesaler in order to minimise the number of
monthly deliveries.198 These examples involved unrelated products. However, such advantages
may also commonly arise in the sale of complements; for example consumers may want to take
out insurance at the same time as they buy an electrical good; car manufacturers integrated
with financial companies can offer purchase financing, saving on contractual costs.199
197
Regulation preventing petrol stations from selling non-petrol products was dropped in recent years in some member
states.
198
According to market studies performed by the British Department of Trade and Industry, this is one of the motivations
for pubs to have a single distributor for beer. See Office of Fair Trading (2000).
199
In the latter example there is also a second, more important, pro-competitive motivation related to economies of scope:
see Section 4.3.1.
Page 103
When transaction costs are significant for either of the trading partners, one-stop shopping
constitutes a genuine source of efficiency gains, and increases social welfare.
Impact of conglomerate integration on consumer welfare
The extent to which the gains are passed on to consumers depends on the intensity of
competition in both markets. Where intense competition prevails, this benefit will generally be
passed on to consumers. On the contrary, if the merging partners have some degree of market
power, the gain to consumers can be appropriated by the firm, at least in part, through an
increase in prices. However, a price increase could still leave consumers better-off, as the
added convenience may more than compensate for the higher price. 200
Can the merger be achieved with other means?
One-stop shopping advantages are often obtained through contractual arrangements; for
example coffee shops have located inside banks and bookshops, and internet companies
selling products such as music and books have linked with those selling holidays and travel.
This suggests that, in these cases, such gains may not require a full merger. However, in other
cases, bargaining difficulties or contractual incompleteness may prevent such arrangements,
and thus motivate a merger. For example, a brewery may be able to provide one-stop shopping
advantages to retailers by distributing non-competing products supplied by other companies.
However, the third party may fear that the brewery lacked sufficient incentive to market its
products but would find it hard to specify such marketing contractually.201
Page 104
200
Price increases could be justified if one stop-shopping is more expensive to provide, e.g. renting space in petrol stations
might be more expensive than corner shops.
201
See, for example, the case study of Guinness/Grand Met in Section 5.
5 Policy Recommendations
5.1
Introduction
Drawing on the analysis presented in Sections 2 to 4, this section provides practical policy
recommendations for the competitive assessment of non-horizontal mergers, and in particular
considers how efficiencies generated by such mergers should be taken into account. We have
two main policy recommendations.
The typical approach to assessing efficiencies as set out in the EC horizontal merger guidelines
divides the competitive assessment into two broad steps: first the potential anti-competitive
effects of the merger are examined and then, where a merger is held to give rise to anticompetitive outcomes, the extent to which efficiencies generated by the merger might offset
such concerns is examined.202 Our first policy recommendation is that this two-step approach is
both impractical and inappropriate for assessing many of the efficiencies generated by nonhorizontal mergers.
The focus of the EC horizontal merger guidelines’ assessment of merger efficiencies is on
marginal cost savings. The impact of such efficiencies on pricing incentives can readily be
assessed separately from any lessening of competition arising from the horizontal merger.203
But, in practice, many of the most important efficiencies generated by non-horizontal mergers
(and some would argue also horizontal mergers) are the result of synergies arising from the
combination of complementary assets rather than direct cost saving as such.204 Incorporating
these types of efficiencies, to which non-horizontal mergers are most likely to give rise, into the
overall competitive assessment cannot easily be done as a separate stand-alone exercise.
Indeed, the two-step approach to assessing efficiencies is appropriate if and only if the following
two conditions hold.
•
The anti-competitive issues raised by the merger can be assessed separately from
the likely effects of any efficiency.
•
A direct trade-off comparison can be made (i.e., the efficiency can be readily
translated into an impact on pricing incentives).
Not only is it difficult to make a direct trade-off comparison between the pro-competitive
efficiency effects and the adverse effects on pricing and/or quality,205 but in many cases it is
202
In practice, it should be stressed that this approach to assessing the competitive effects of a horizontal merger is only
appropriate for (a) assessing unilateral effects concerns and (b) only applies in practice to marginal cost efficiencies.
203
Indeed, it can be argued that the current approach to assessing efficiencies generated by horizontal mergers is itself
too restrictive with its emphasis on marginal cost reductions.
204
This point has also been made by Kolasky (2004). See also Teece (1980, 1982), and Jorde and Teece (1990).
205
For example, how can one translate the efficiency generated by combining complementary assets into an impact on
pricing incentives?
Page 105
simply not possible to assess the potential anti-competitive concerns separately from the procompetitive effects for the simple reason that both have the same source. Indeed, in many
instances, the theory of competitive harm posited in non-horizontal mergers arises because the
merger generates an efficiency. For example, in assessing the competitive effects of a merger
that is expected to give rise to bundling, the anti-competitive concern (foreclosure of rivals who
are not able to bundle) has the same source as the pro-competitive benefits (the elimination of
pricing inefficiencies).206 This implies that for many types of efficiency generated by nonhorizontal mergers, the assessment of efficiencies necessarily forms part of an integrated
unified competitive assessment.207
Second, there should be an economic presumption that non-horizontal mergers are procompetitive.208 This conclusion derives from a fundamental difference between horizontal and
non-horizontal mergers. By definition, a horizontal merger removes a direct competitive
constraint and this raises the possibility that post-merger prices will increase to the detriment of
consumers.209 In marked contrast, non-horizontal mergers do not, also by definition, remove a
direct competitive constraint. Moreover, the general impact of a non-horizontal merger on prices
is to reduce them as a result of eliminating externalities and other inefficiencies that might have
existed pre merger. While this is not to say that no non-horizontal merger is anti-competitive, it
does indicate an economic presumption (albeit rebuttable) that such mergers are procompetitive.210
This view is echoed by the European Court of Justice in its Tetra Laval judgement:
“The analysis of a conglomerate-type concentration is a prospective analysis in
which, first, the consideration of a lengthy period of time in the future and,
secondly, the leveraging necessary to give rise to a significant impediment to
effective competition mean that the chains of cause and effect are dimly
discernible, uncertain and difficult to establish. That being so, the quality of the
evidence produced by the Commission in order to establish that it is necessary
to adopt a decision concerning the concentration incompatible with the
common market is particularly important, since that evidence must support the
Commission’s conclusion that, if such a decision were not adopted, the
economic development envisaged by it would be plausible”. 211
This economic presumption has important implications for the stance adopted in the integrated
assessment of the likely impact of a non-horizontal merger where the pro-competitive efficiency
Page 106
206
See Church (2004) for a recent review of the potential anti-competitive effects of non-horizontal mergers.
207
Paradoxically, in such cases, the more modest the efficiency the less likely it is that the required marginalisation of
competitors will take place.
208
See Section 2. In particular, most economists recognise that non-horizontal mergers are generally much more likely to
generate efficiencies than horizontal mergers. See for example Reiffen and Vita (1995), and Scheffman and Higgins
(2004). A similar view is expressed by Cooper et al. (2005) in the closely related context of vertical restraints.
209
We disagree with those commentators who argue that every horizontal merger gives rise to a price increase. Not only
are such claims not well-founded in theory, they are unsupported by empirical evidence.
210
This economic presumption does not necessarily imply a legal presumption. Nonetheless, it entails that compelling
evidence will be required to establish that the merger leads to a significant impediment of effective competition.
211
Judgment of the Court in case C-12/03 P, 15 February 2005, paragraph 44.
effects of the merger cannot be disentangled from the potential anti-competitive effects. Namely,
in the competitive assessment of those commercial strategies that might give rise to both
efficiencies and potential anti-competitive concerns, it should be assumed that the procompetitive effects are predominant, i.e., any efficiency generated by a non-horizontal merger is
deemed to give rise to pro-competitive outcomes, unless it can be demonstrated otherwise. This
is in marked contrast to the approach, often labelled the efficiency offence, whereby efficiencies
tend to be viewed as inherently problematic.
The remainder of this section is organised as follows. Section 5.2 summarises the efficiencies to
which vertical and conglomerate mergers respectively give rise, by reviewing the necessary
conditions for the efficiencies to arise and highlighting their impact on consumers. We also
consider the scope for obtaining such efficiencies via non-merger alternatives. This summary
demonstrates that there are numerous efficiency rationales that motivate vertical or
conglomerate mergers.
Section 5.3 discusses efficiencies generated by non-horizontal mergers and how these should
be integrated into the competitive assessment by expanding on the above discussion. Section
1.4 then presents an analytical framework for assessing the competitive effects of those nonhorizontal mergers where it is necessary to assess efficiencies as part of a unified competitive
assessment.
5.2
Efficiencies Arising From Non-Horizontal Mergers
This section provides a summary of the discussion of efficiencies that can arise from vertical
and conglomerate mergers that were discussed in Sections 3 and 4 respectively.
5.2.1
Vertical mergers
This section summarises the main conclusions that can be derived from the analysis of vertical
mergers presented in Section 3. The summary provided below is intended as a cross-reference
tool that can be used to identify quickly the efficiencies that may be relevant to a specific case
as well as the relative sections of this report. It does not constitute a checklist for the
assessment of the efficiencies. The efficiencies are grouped into four categories used in Section
3. These are:
•
increased pricing efficiency;
•
improved productive efficiency;
•
the prevention of profit expropriation; and
•
overcoming problems arising from incomplete contracts and/or transaction costs.
Page 107
5.2.1.1
Increased pricing efficiency
This section summarises factors relating to the assessment of pricing efficiencies arising in
vertical mergers. Amongst these, the elimination of double marginalisation has the broadest
applicability. This inefficiency occurs when there are multiple mark-ups in a supply chain, that is,
price exceeds the marginal cost of production at more than one stage in the supply chain. Due
to the ubiquitous presence of fixed costs and market power in the economy, prices often exceed
short run marginal cost so that mark ups are likely to be an extremely widespread phenomenon.
However, vertical integration can also give rise to consumer benefits where goods are
differentiated and imperfect competition exists at both levels of the supply chain, although in this
case the overall impact of vertical integration on consumer welfare can be positive or negative.
The double marginalisation problem becomes greater the higher the mark-up, the greater the
market elasticity, the larger the number of inputs, and where the value of the components in the
integrated good are more similar. On the contrary, it is less important when it is possible to price
discriminate, where intermediate goods have a secondary use or where inputs are used in fixed
proportion. In some circumstances, the double marginalisation problem can be eliminated or
reduced through specific contractual arrangements between firms. Whether this is possible
depends on the complexity of the business environment and the complexity of the relationship
between the two firms (the more complex, the less likely). In particular, when firms have
additional reasons to cooperate, the relationship can grow excessively complex for a contract,
and a merger may represent the most efficient solution or indeed the only solution. Finally, in
practical cases vertical integration may not solve the problem if the internal transfer price
between the divisions of the merged firm is set above marginal cost.
Mark-ups can also produce pricing inefficiencies due to input substitution. This occurs when a
firm substitutes away from an input that is subject to a mark-up, towards an alternative product
that is sold at a lower mark-up. The use of inputs is then no longer efficient and, by integrating,
the firm can eliminate the mark-up, and increase efficiency. Moreover, vertical integration can
lower the cost of inputs to rival firms. This may be shown with the help of the same theoretical
models usually employed by the literature investigating the effect of a vertical merger on raising
rivals’ costs. Depending on the value of the parameters used in the models, the price of inputs
and of final products may rise or fall. Finally, we discussed how vertical integration can increase
profits through price discrimination, although the welfare effects in this case are often
ambiguous, as some consumers lose while others gain.
Page 108
Figure 5-1: Pricing efficiencies in vertical mergers
Necessary conditions for the
efficiency gain
Short term impact on consumers
and rival firms
Potential non-merger alternatives
- Constant unit price (at least for
marginal purchases);
- Homogeneous products: consumer
welfare rises;
- RPM;
- Positive mark-ups both upstream
and downstream;
- Differentiated products: some
consumers certainly gain, others
may lose; the number of varieties
may decrease;
Double Marginalisation (Section 3.2.1)
- Post merger, internal transfer price
is closer to marginal cost.
- Franchise fee/two-part tariff;
- Quantity forcing.
- Rivals feel increased competitive
pressure.
Input Substitution (Section 3.2.2)
- Several inputs can be used in
variable proportions;
- Market power of the integrated firm
increases;
- Inputs are sold at constant unit
price (at least for marginal
purchases);
- Price to consumers may rise or fall,
depending on the size of the
efficiency gain relative to the
market power increase;
- Inputs are charged different markups.
- RPM;
- Franchise fee/two-part tariff;
- Quantity forcing;
- Bundling.
- Rivals feel increased competitive
pressure.
(Upstream) Price Discrimination (Section 3.2.4)
- Downstream firms vary in their
degree of sensitivity to input price;
- Downstream firms would engage in
arbitrage against different input
prices.
- Price-sensitive downstream firms
pay lower prices and expand
output;
- Means suitable to prevent arbitrage
of input prices between
downstream firms.
- Price-insensitive downstream firms
pay higher prices;
- Lower input prices are passed
through to consumers;
- Higher prices may be passed
through to consumers.
5.2.1.2
Improved productive efficiency
Section 3.3 provides guidance on the assessment of productive efficiencies arising from vertical
mergers. Economies of scope resulting from technological synergies and shared indivisible or
hard-to-trade assets are often cited by merging parties as a rationale for mergers. These
efficiencies are particularly likely to require a merger because, in case of contractual
agreements, the specialised nature of the productive asset exposes the firms to the risk of
opportunistic behaviour (see Section 3.5.2). In addition, vertical integration might improve
productive efficiency by increasing coordination of demand and supply between upstream and
downstream firms. Although economic models for such supply assurance are complex, this
efficiency has intuitive appeal, and is often cited by business people as a motivation for vertical
Page 109
integration. Firms may also benefit from increased coordination and information flow in the fields
of R&D, distribution, marketing, and for the adoption of new production technologies. Increasing
managerial efficiency is also an important motivation for mergers.
Figure 5-2: Productive efficiencies in vertical mergers
Necessary conditions for the
efficiency gain
Short term impact on consumers
and rival firms
Potential non-merger alternatives
- If variable costs fall: consumers
enjoy lower prices and larger
output;
- Unilateral expansion (generally
unlikely).
Economies of Scope (Section 3.3.1)
- The merger realises technological
synergies;
- OR the merger allows the parties to
share indivisible, hard-to-trade
assets;
- If fixed costs fall: prices may fall.
- As a result of the above, production
costs of one or both products falls.
Supply Assurance (Section 3.3.2)
- Input market characterised by
uncertainty of prices and/or
availability;
- Significant risks of input rationing
or costly over-production.
- Possibly lower prices to final
consumers;
- Increased coordination of the
supply chain.
- Possibly qualitative improvements
such as increased product variety;
- May enhance competitors’ supply
assurance problems
Improved Information Flow and Coordination (Section 3.3.3)
- Improved coordination can
effectively increase the efficiency
of: R&D, distribution, marketing or
pricing.
- Possibly qualitative improvements
such as increased product variety;
- Contractual agreements.
- Possible reduction of fixed costs
may result in lower average prices
in the long term, or increased
investment;
- May harm cooperation with nonintegrated firms.
5.2.1.3
Preventing profit expropriation
Section 3.4 considers efficiencies arising from externalities which prevent firms from
appropriating all of the gains of their investment. For example, a firm may fear that competitors
may be able to “free ride” on its investment. This reduces the firm’s incentive to invest to the
detriment of itself and its vertical partners. For example, a retailer may be unwilling to promote a
manufacturer’s product if this benefits its competitors. In consequence, the service is therefore
not provided or at a lower than optimal level, and output falls. If the upstream firm integrates
with the downstream suppliers and excludes independent retailers, all the gains are
appropriated by the single, integrated entity, and this increases the incentive to invest.
Page 110
Partial expropriation of profits by other firms is also common in R&D activity. Firms may be
concerned that its competitors will find out about its innovation and know-how, and then copy
(“free ride” on) the results. Vertical integration can mitigate such horizontal spillovers, by
reducing the chance that “know-how” will be leaked to competitors via other firms in the supply
chain, or by providing the opportunity to embed the innovation in the upstream product. Finally,
innovating firms generally increase the profitability of their vertical partners, e.g. upstream
components manufacturers and downstream distributors. Vertical integration is in this case a
means to appropriate a larger part of the benefits from the investment in R&D. This reduction in
vertical spillovers can increase the firm’s incentive to innovate, to the consumers’ advantage.
Figure 5-3: Prevention of profit expropriation in vertical mergers
Necessary conditions for the
efficiency gain
Short term impact on consumers
and rival firms
Potential non-merger alternatives
Free riding on pre-sale services (Section 3.4.1)
- Competition is effective at the level
where free riding takes place;
- Investment in pre-sales services is
prone to expropriation by
competitors;
- The investment level is difficult to
specify in a contract.
- Consumers enjoy better pre-sales
services and pay slightly higher
prices;
- Consumer welfare generally does
not decrease and may increase,
provided inter-brand competition is
effective;
- RPM;
- Exclusive territories distribution
(against retailer free riding);
- Exclusive dealing agreements
(against manufacturer free riding).
- Some retailers (manufacturers)
may be foreclosed from the input
(outlet).
Internalisation of R&D benefits: horizontal spillovers (Section 3.4.2)
- Innovation can be imitated or
copied by rivals;
- Competitors’ incentive to invest in
R&D may be reduced;
- The imitation would be sufficiently
timely and effective to expropriate
a part of the returns from the
innovation;
- In the medium term, increased
quality and/or variety of the
products supplied by the integrated
firm (because of the increased
incentive to invest);
- The prospective loss would be
sufficiently significant to deter the
R&D investment;
- Integration would materially reduce
the above factors so to restore the
incentive to invest.
- In the long term, knowledge
spillovers from the increased R&D
activity benefit competitors and
society.
Internalisation of R&D benefits: vertical spillovers (Section 3.4.3)
- Pre merger, the vertical partner of
the innovating firm price at a markup.
- For process innovation, consumer
welfare rises if output increases;
- Co-participation to the R&D
investment;
- For product innovation, consumers
gain from increased variety;
- Fixed-price agreements.
- In the long term, knowledge
spillovers benefit competitors and
society.
Page 111
5.2.1.4
Overcoming problems arising from incomplete contracts and/or transaction
costs
Section 3.5 considers efficiencies arsing from the existence of incomplete contracts and
transaction costs. This is a central issue in the analysis of pro-competitive motivations for
vertical mergers, since virtually all efficiencies could be realised with a contractual agreement if
it was possible to specify such contracts completely at no cost, and there was no information
asymmetry. Transaction costs include the agency costs incurred in providing appropriate
incentives or in monitoring firms where “moral hazard” can arise due to incomplete contracts. A
specific example of this is the “hold-up” problem, where a relationship-specific investment
makes one party very dependent on the other. This gives rise to the opportunity for exploitative
behaviour once the investment has been undertaken. Where agency costs imply that protection
from such behaviour is impossible or at least prohibitively expensive, the investment may be
deterred. At a general level, even where there are no specific incompleteness problems,
transaction costs are always incurred to a greater or lesser degree in searching for parties and
drawing up contracts.
Figure 5-4: Incomplete contracts and transaction costs in vertical mergers
Necessary conditions for the
efficiency gain
Short term impact on consumers
and rival firms
Potential non-merger alternatives
- Incentives to invest generally
increase;
- Reputation;
Moral Hazard (Section 3.5.1)
- The parties’ undertakings cannot
be fully specified ex-ante in a
contract;
- Ex-post, the informed party has the
economic incentive to exploit the
contractual loopholes, to the
detriment of the uninformed party:
- Integration allows a better
alignment of incentives of all the
parties involved (including
employees).
- Social welfare increases;
- Consumers generally benefit from
the increased investment level;
- Commitment devices: i.e.,
eliminate the incentive to behave
opportunistically.
- Over time, prices may fall, to the
extent that lower agency costs are
accounted as lower average costs
of production.
Hold-up Problem (Section 3.5.2)
- Moral Hazard conditions hold;
- Same as for moral hazard.
- The investment is relationshipspecific;
- For upstream investments: quantity
forcing;
- For downstream investments:
exclusive dealing;
- The investment is a sunk cost.
- Reputation;
- Commitment.
General Transaction Costs (Section 3.5.3)
- Small and frequent transactions;
- OR significant searching costs;
- OR delayed payment for sales.
Page 112
- If variable production costs fall and
output increases: consumers enjoy
lower prices;
- If fixed costs fall: prices may fall.
- Long term contracts.
5.2.2
Conglomerate mergers
This section summarises the main conclusions that can be derived from the analysis of
conglomerate mergers presented in Section 4. The summary given in this section is intended to
provide a cross-reference tool for the quick identification the efficiencies that may be relevant to
a specific case as well as the corresponding sections of this report. It is not meant to be a
checklist for the assessment of efficiencies.
The efficiencies are grouped into four categories used in Section 4; namely:
5.2.1.5
•
increased pricing efficiency;
•
improved productive efficiency;
•
the prevention of profit expropriation; and
•
overcoming problems arising from incomplete contracts and/or transaction costs.
Increased pricing efficiency
Figure 5-5 below summarises the efficiency gains arising from eliminating pricing inefficiencies.
These are particularly relevant because they can result in immediate price reductions. The
Cournot effect is the conglomerate equivalent of the double marginalisation inefficiency arising
in vertical industries. The importance of the Cournot effect is due to its very broad range of
applicability. In general, any two firms which supply complementary products and price at a
mark-up are affected by this source of inefficiency. A merger reduces the inefficiency by
eliminating the externality between the merging firms. Further, unlike double marginalisation,
contractual alternatives to a full merger are not simple and rarely used in practice. Note that, in
order to fully reap the benefits of the internalisation of this pricing externality, complementary
products need to be sold jointly through tying or bundling. The reason for this is that, if products
are sold individually, customers are free to “mix and match” components of different brands, so
that rival firms may appropriate part of the gains from a price reduction. In those cases, the
magnitude of the efficiency gain will depend on the portion of customers who buy the products
together even without a tie.
Complementary mergers may also increase pricing efficiency when they allow the merged entity
to price discriminate by “metering”. With metering, the welfare of customers with a low valuation
for the goods generally increases, while the welfare of high valuation customers generally
decreases. However, welfare of all consumers may increase if post merger all competitors price
discriminate.
Finally, bundling or tying may increase pricing efficiency by reducing the heterogeneity of
customers’ valuations. The efficiency gain is larger if the bundled products are unrelated, and
disappears if the goods are strongly complementary. This mechanism may be particularly
relevant for information goods.
Page 113
Figure 5-5: Pricing Efficiencies in Conglomerate Mergers
Necessary conditions for the
efficiency gain
Short term impact on consumers
and rival firms
Potential non-merger alternatives
- Differentiated products: some
consumers gain and others lose;
overall impact is ambiguous;
- Agreements allowing to bundle
products and split revenues.
Avoiding the Cournot effect (Section 4.2.1)
- Products are complementary;
- Constant unit price (at least for
marginal purchases);
- Positive mark-ups in both
complementary markets;
- The gain is larger if, post merger,
complements are bundled or tied.
- Homogeneous products: consumer
welfare increases;
- Rivals feel increased competitive
pressure.
Price discrimination: metering (Section 4.2.2)
- Products are complementary;
- Products have no secondary use;
- Products are used in variable
proportions;
- Post merger, the complements are
tied.
- Low valuation customers generally
gain;
- High valuation customers generally
lose;
- Other price discrimination devices
not requiring tying may be
available;
- Directly metering use.
- If rivals also implement price
discrimination, consumer welfare
may increase further;
- Welfare may be reduced by
wasteful activities undertaken by
firms and consumers.
Reducing Customer Heterogeneity (Section 4.2.3)
- Customers valuation of both goods
is dispersed;
- Customers valuations of the goods
are not strongly positively
correlated;
- Output generally increases;
- Impact on consumer welfare is
ambiguous: some consumers gain
and others lose.
- Price discrimination devices when
available.
- Products are sold at a mark-up;
- Post merger, products are bundled
or tied.
5.2.1.6
Improved productive efficiency
Conglomerate mergers can achieve higher productive efficiency in a number of ways.
Economies of scope and scale essentially arise as a consequence of some sort of horizontal
overlap between the merging parties (for example in distribution and marketing, administration,
or R&D). However, for the efficiency gain to be merger specific, a synergy must be generated by
the common use of some indivisible, hard-to–trade, productive asset. This may include the
sharing of human capital or immaterial assets such as brand names. These efficiency gains
may result in lower marginal cost or lower fixed costs. Where marginal production costs are
reduced, the merged firm would have an immediate incentive to reduce prices.
Page 114
Conglomerate mergers may also increase productive efficiency by reducing the cost of capital,
thereby increasing the merged entity’s financial efficiency. A firm with a distressed balance
sheet, or which faces significant uncertainty over its future cash flow, may reduce the cost of
borrowing by merging with a firm whose cash flow is uncorrelated with its own, thus reducing
the risk of default. The efficiency gain is greater if the cash flows of the two firms are of similar
size (or if the more volatile cash flow has the smaller magnitude). Since capital can be one of
the inputs for production, a lower cost of capital may result in lower marginal costs. Otherwise, it
would lead to lower average cost and/or increased incentive to invest, which may also benefit
consumers.
Figure 5-6: Productive efficiencies in conglomerate mergers
Necessary conditions for the
efficiency gain
Short term impact on consumers
and rival firms
Potential non-merger alternatives
- If variable costs fall: consumers
enjoy lower prices and larger
output;
- Contractual agreements;
Economies of Scope and Scale (Section 4.3.1)
- Integration allows the parties to
share indivisible, hard-to-trade
assets, resulting in a reduction of
production costs.
- Unilateral expansion (unlikely).
- If fixed costs fall: prices may fall.
Financial Efficiency (Section 4.3.2)
- At least one party faces significant
uncertainty over its future cash
flow;
- Cash flows from the two activities
are uncorrelated (or negatively
correlated);
- Expected cash flows from the two
activities are of similar magnitude;
- OR one company is financially
distressed and the other has
sufficient resources for both (failing
firm defence).
- If efficiency gain allows an
investment that would otherwise be
deterred, social and consumer
welfare generally increase;
- If fixed costs are reduced, prices
may fall;
- If variable costs are reduced (i.e.,
where capital is an input into
production), prices fall and output
expands.
Increased Buyer Power (Section 4.3.3)
- The merging parties employ the
same input, or inputs that are
supplied by the same firm;
- Supply of the input(s) is imperfectly
competitive;
- Prices to final consumers generally
decrease as a result of the lower
marginal cost of the downstream
firms.
- Common procurement
agreements.
- Common procurement would lead
to better prices;
- Downstream competition is
effective.
Page 115
Productive efficiency can also be improved by increasing buyer power. Clearly, this can only
lead to a reduction of input prices when the upstream supplier has some degree of market
power. If competition in the downstream market is effective, the price reduction is likely to be
passed on to consumers, at least in part. However, the same result can be obtained by
independent firms through common procurement agreements, when this is feasible. Finally,
conglomerate mergers can increase managerial efficiency (Section 4.3.4) if the acquirer
replaces inefficient administrators in the target firm.
5.2.1.7
Prevention profit expropriation
Complementary mergers can promote efficiency by helping to prevent profit expropriation. First,
a merger can facilitate quality standardisation. Controlling the quality of both complements may
be important when two goods are only used together (as a system good), and if consumers can
observe the quality of the system, but not the quality of the single components (i.e., there is
asymmetric information between consumers and firms). Consumer benefits generally take the
form of higher quality products. However, in certain cases the problem may be solved without
integration through contractual agreements and/or building brand reputation and consumer
awareness.
Secondly, a merger can increase the incentive to invest in innovation by allowing a firm to
internalise R&D benefits. Expropriation of the returns from R&D investments is a common
problem, mainly for two reasons:
Page 116
•
Profit expropriation from horizontal spillovers arises when innovations can be
copied or imitated by competitors and patents do not provide sufficient protection.
Although this is a common phenomenon, consumer harm arises only if the profit
expropriation is sufficiently strong to reduce the firm’s investment in valuable
innovation. Also, the impact of the merger on competitors’ incentives to innovate
must be estimated.
•
Profit expropriation from complementary spillovers generally arises whenever two
products are complementary, and is more significant the stronger the
complementarity. A further necessary condition for profit expropriation to occur is
that the supplier of the complementary product prices at a mark-up over marginal
cost.
Figure 5-7: Prevention of profit expropriation in conglomerate mergers
Necessary conditions for the
efficiency gain
Short term impact on consumers
and rival firms
Potential non-merger alternatives
- Consumers enjoy higher quality
products.
- Licensing agreements with quality
monitoring (may be costly);
Quality Standardisation (Section 4.4.1)
- Products are complementary and
mainly used together;
- Consumers can observe the quality
of the system, but not the quality of
the single components.
- Unilateral expansion (i.e.,
production and branding of the
complement).
Internalisation of R&D benefits: horizontal spillovers (Section 4.4.2)
- Innovation can be imitated or
copied by rivals;
- Competitors’ incentive to invest in
R&D may be reduced;
- The imitation would be sufficiently
timely and effective to expropriate
a part of the returns from the
innovation;
- In the medium term, quality and/or
variety of the products supplied by
the integrated firm would increase;
- The prospective loss would be
sufficiently significant to deter the
R&D investment;
- In the long term, knowledge
spillovers benefit competitors and
society.
- Integration would materially reduce
the above factors so to restore the
incentive to invest.
Internalisation of R&D benefits: complementary spillovers (Section 4.4.3)
- Products are complements, and
are mainly used together;
- For process innovation, consumer
welfare rises if output increases;
- Pre merger, the products are
priced at a mark-up.
- For product innovation, consumers
gain from increased variety;
- Sharing of the R&D investment;
- Fixed-price agreements.
- In the long term, knowledge
spillovers benefit competitors and
society.
5.2.1.8
Overcoming problems arising from incomplete contracts and/or transaction
costs
Incomplete contracts and transaction costs play a central role also in the analysis of procompetitive motivations for conglomerate mergers, as well as vertical mergers. The reason for
this is that virtually all efficiencies could be realised with a contractual agreement if there were
no informational asymmetries and it were possible to specify such contracts completely at no
cost. The general analysis of incomplete contract issues, however, does not depend on whether
the parties are in a vertical relationship or operate in complementary or unrelated markets
instead. Therefore, the conditions for realisation of efficiency gains from the elimination of moral
hazard or the hold-up problem are the same as described for vertical mergers. In addition,
Page 117
conglomerate mergers may reduce transaction costs by providing one-stop-shopping
opportunity for customers. This leads to an efficiency gain whenever transaction costs are nonnegligible, either for the seller or for the buyer. However, this benefits consumers only to the
extent that effective competition ensures that the merged firm cannot appropriate all the
efficiency gain.
Figure 5-8: Incomplete contracts and transaction costs in vertical mergers
Necessary conditions for the
efficiency gain
Short term impact on consumers
and rival firms
Potential non-merger alternatives
- Incentives to invest generally
increase;
- Reputation;
Moral Hazard (Sections 4.5.1 and 3.5.1)
- The parties’ undertakings cannot
be fully specified ex-ante in a
contract;
- Ex-post, one of the parties has the
economic incentive to exploit the
contractual loopholes, to the
detriment of the other party;
- Integration allows a better
alignment of incentives of all the
parties involved (including
employees).
- Social welfare increases;
- Consumers generally benefit from
the increased investment level;
- Commitment devices: i.e.,
elimination of the incentive to
behave opportunistically.
- Over time, prices may fall, to the
extent that lower agency costs
reduce the average costs of
production.
Hold-up Problem (Section 3.5.2)
- Moral Hazard conditions hold;
- Same as for moral hazard.
- The investment is relationshipspecific;
- For upstream investments: quantity
forcing;
- For downstream investments:
exclusive dealing;
- The investment is a sunk cost.
- Reputation;
- Commitment.
One-Stop Shopping (Section 4.5.2)
- The merging parties’ products are
purchased by the same customers;
- Transaction costs are nonnegligible, either for the seller or for
the buyer, or both;
- Competition is effective.
Page 118
- Consumer welfare increases;
- Distribution agreements;
- Even if competition is not
sufficiently effective and prices
increase, social welfare may still be
higher.
- Unilateral expansion into
distribution of the other party’s
product.
5.3
5.3.1
Assessment of Non-Horizontal Merger Efficiencies
Introduction
This section discusses the role of efficiencies in non-horizontal mergers. Although the goal of
the competitive assessment of all mergers is the same, (namely, to consider whether, overall,
consumers stand to gain more from the efficiency benefits created by the merger than they lose
from any reduction in competition), it is important to understand the implications of the
fundamental differences between horizontal and non-horizontal mergers for the manner in which
efficiencies and anti-competitive effects should be assessed. As discussed in Section 2, the
economic characteristics of horizontal and non-horizontal mergers are fundamentally distinct.
The main differences can be summarised as follows.
•
First, a horizontal merger results in the direct elimination of a competitive constraint
and this increases the likeliness of a post-merger price increase. Such anticompetitive price increases might however be offset if the merger also gives rise to
marginal cost efficiencies.212 Horizontal mergers may also, give rise to other types
of efficiencies including dynamic efficiencies that cannot readily be measured in
terms of their impact on pricing incentives. In contrast, not only does a nonhorizontal merger not directly remove a competitive constraint, but instead it gives
rise to a general incentive to reduce price by directly eliminating externalities and
inefficiencies. This implies that competition can only be adversely affected
indirectly.
•
Second, the anti-competitive effects of a horizontal merger can usually be analysed
separately from any efficiencies generated by the merger. In contrast, for many
non-horizontal mergers, the source of the potential competitive harm is the same as
the source of the efficiency generated by the merger. Hence, the competitive
assessment cannot sensibly be conducted in two separate steps: i.e., assess the
likelihood of anti-competitive outcomes first and then consider whether efficiencies
more than offset such concerns. Rather, in such cases, efficiencies must form part
of a unified competitive assessment.
Since non-horizontal mergers can only adversely affect competition indirectly, there is (or ought
to be) an economic presumption that such non-horizontal mergers are pro-competitive.213 This
implies that it is assumed that the efficiency benefits arising from the merger predominate
unless it can be demonstrated otherwise. This is in marked contrast to the approach, often
labelled the efficiency offence, whereby efficiencies are viewed as inherently problematic unless
it can be demonstrated otherwise.
212
This statement applies primarily to unilateral effects for the reasons discussed in Section 2.
213
This does not amount to a legal presumption; see also footnote 210. A similar approach is adopted by Cooper et al.
(2005) in the context of vertical restraints.
Page 119
The remainder of this section provides a stylised example of a typical non-horizontal merger.
The implications of the differences between horizontal and non-horizontal mergers mentioned
above are also discussed in more detail and illustrated by reference to this example.
5.3.2
An illustrative example214
This section provides a hypothetical example of a non-horizontal merger and describes some
typical claims for efficiency gains and anti-competitive harm. It also illustrates the potential
tension between the efficiencies generated by a merger and their potential anti-competitive
effects.
The merger involves Firm M, a producer of widget machinery and Firm C, which produces a
vital component of that machinery. Firm C is held to be dominant in the component market. The
production of both the machinery and the components requires considerable investments, to
cover R&D (fixed) costs. Firms therefore charge a mark-up over the marginal production cost.
Moreover, the components are used on a one-to-one basis with the machinery; that is every
machine must have exactly one component in it. The unified entity would therefore have an
incentive to reduce the price of components (eliminating double marginalisation). The merger
may also lead to benefits in the form of improved interoperability of the two products, as well as
more efficient R&D arising from the pooling of human capital.
However, there might also be potential competition concerns arising from the fear that other
machinery suppliers may be marginalised. Even if the merged firm continued to supply
components to other machinery providers, without the benefit of the reduced double
marginalisation and other benefits the merged firm obtains post-merger, the rival firms would not
be as efficient as the merged party. Moreover, if the merged party fears that its new R&D efforts
could be leaked to other parties because of spillover effects, it may not make its components
available to other machinery producers, putting them at a further disadvantage. If rivals are
unable to replicate the gains of the merged party by offering their own competing “bundle”
through a merger with other machinery producers or more loosely organised selling consortia,
they might be marginalised, and become unable to finance and sustain investment in new
product development. Over time, they may exit the market or at least become less efficient
competitors.
If this is the case, the merger would ultimately lead to an increase in Firm M’s market power. If
firms are unable to re-enter the market, this may allow the merged entity to raise price or reduce
R&D and harm customer interests. However, the component manufacturer already has market
power upstream. Therefore, following the “one-monopoly-profit” theory, it already has control
over the pricing of the system. But Firm C may fear that a machinery manufacturer may, in the
long term, develop its own components, threatening its existing market power. The merged
party might be able to guard against this threat if its foreclosure can reduce the potential
entrant’s market share and weaken its financial situation to the extent that it would not have
sufficient funds to enter the component market.
Page 120
5.3.3
For many efficiencies it is not possible to separately assess the anti-competitive
and pro-competitive effects
In those horizontal merger cases which raise the prospect of anti-competitive unilateral effects,
it is possible (at least, in principle) to conduct the anti-competitive assessment independently of
the effects of any marginal cost efficiencies arising from the merger. While this is the case for
some non-horizontal mergers (for example, where the competitive concern relates to raising
rivals’ costs or refusal to supply), for many non-horizontal mergers this is not the case, since the
potential anti-competitive effects and efficiencies are intrinsically interrelated. For such nonhorizontal mergers, it will very often be the case that the source of efficiencies realised by the
merging parties is also the source of the competition concern. Many theories of competitive
harm arising from non-horizontal mergers are based on concerns that the efficiencies generated
by the merger will enable the merged entity to marginalise its competitors by giving the merged
firm a distinctive competitive advantage over its rivals.
Using our hypothetical example to illustrate this point, rivals are only marginalised if the
competitive advantage gained by the merged firm is so overwhelming that – without countermergers or other such strategies – its rivals are unable to compete. Paradoxically, the greater is
the efficiency, the greater the competitive concern. Indeed, in many cases efficiencies
generated by non-horizontal mergers are essential for the success of the alleged foreclosure
strategy! Where this is the case (as it will be in many instances), the efficiencies generated by
non-horizontal mergers cannot be assessed separately from the potential anti-competitive
concerns and must therefore be assessed as part of a unified integrated analysis. 215
Table 1 provides a general categorisation of the main theories of competitive harm. These
theories are divided into two categories: those where efficiency benefits can be assessed
separately so that the two-step approach is likely to be appropriate and those where the alleged
competitive harm has the same source as the efficiency benefit, in which case the two-step
approach is not appropriate.216 It should be stressed that this provides a general categorisation
and that it is possible that in particular circumstances a given theory of competitive harm might
justify the alternative approach.
5.3.4
An economic presumption of pro-competitiveness
As described in Section 2, a horizontal merger can give rise to unilateral effects by directly
eliminating the competitive constraint provided by the other merging party.217 In contrast, nonhorizontal mergers do not directly eliminate competition. They can only reduce competition by
214
A discussion of the theoretical foundations of anti-competitive theories of harm is provided in RBB Economics (2002).
215
Some foreclosure models where efficiencies play a crucial role are the following. In the context of vertical mergers, see
for example Salinger (1988), Gaudet and Long (1996), Higgins (1999), Avnel and Barlet (2000), Choi and Yi (2000),
and Chen (2001). For conglomerate mergers, see Choi (2004), Nalebuff (2000), Denicolo (2000), and Ma (1997). See
Church (2004) for a review.
216
A basis for this categorisation can be found in Sections 2 to 4 of this report and in Church (2004).
217
The authors do not agree with those commentators that assert that all horizontal mergers absent efficiencies give rise
to a post-merger price increase. This assertion is flawed both on theoretical and empirical grounds.
Page 121
bringing about the indirect marginalisation of competitors.218 For example, in the hypothetical
case above, the machinery manufacturer and component manufacturer were not direct
competitors. The merger therefore did not eliminate a direct competitive constraint and
competition can only be adversely affected through some indirect route, for example if the
merged firm denies rival firms access to vital components and in consequence leads to their
marginalisation.
Figure 5-9:
General categorisation of theories of competitive harm
Theory of competitive harm
Efficiency benefits can be assessed separately?
Tying
No
Bundling
No
Raising rivals’ costs
Yes
Refusal to supply
Yes
Portfolio effects
No
However, it is not always possible to reduce competition in this way. For instance, our
hypothetical example relies on Firm C being dominant in the market for components, and on
rival firms being unable to respond in any way or to re-enter if prices were to rise. There is also
no general anti-competitive incentive for the merged firm to enhance its power in the machinery
market. In our example, the incentive only arises because the merged firm fears that a rival
machinery manufacturer may, in the long term, threaten its position in the component market.
Thus, for the ability and incentive to eliminate competition to arise, careful attention must be
given to the specific characteristics of the industry and the products under consideration.
In summary, where it is necessary to conduct a unified assessment of the efficiencies and
potential anti-competitive effects arising from a non-horizontal merger one ought to adopt the
stance that observed efficiencies give rise to pro-competitive outcomes unless it can be
demonstrated otherwise. This is in marked contrast to the approach, often labelled the efficiency
offence, where efficiencies are viewed as problematic unless it can be demonstrated
otherwise.219
Page 122
218
Some authors suggest that in certain particular circumstances non-horizontal mergers may also soften competition,
leading to higher prices in the post-merger equilibrium (see for example Carbajo and Seidmann, 1990, and Chen,
1997). However, unlike unilateral effects in horizontal mergers, these effects tend to rely on rivals’ responses, i.e.,
absent an appropriate reaction from the merged firm’s competitors, the merger would be either pro-competitive or
competitively neutral. In contrast, a horizontal merger can give rise to unilateral effects regardless of rivals’ reactions
(even though these reactions can increase the harm to consumers).
219
A corollary of the above is the inappropriateness of the same strict criteria set out in the horizontal merger guidelines for
the assessment of these types of efficiency.
5.4
A Unified Approach to Assessing Competitive Effects
This section presents an analytical framework for assessing the competitive effects of those
non-horizontal mergers where it is necessary to assess efficiencies as part of a unified
competitive assessment. This analytical framework applies to the assessment of potential
exclusionary strategies, which are generally the main concern in non-horizontal merger cases.
Further, it illustrates a general approach that might be applied also to other types of concerns.
This framework is based on the analysis undertaken by RBB Economics on behalf of General
Electric in its acquisition of Amersham. This analysis was incorporated directly into the
Commission’s Phase I clearance decision.
Our analytical framework has four conditions which need to be present in order for nonhorizontal mergers to lead to anti-competitive effects through exclusionary effects.220 As
discussed above, given the economic presumption of pro-competitiveness, in considering
whether these conditions apply, it is appropriate to adopt an economic presumption that the procompetitive effects of commercial strategies that give rise to both efficiencies and potential anticompetitive concerns are predominant, unless the opposite can be shown. The conditions
required for non-horizontal mergers to harm competition through foreclosure are as follows:221
•
Condition A: Existence of significant market power. Pre merger, at least one of
the merging parties has significant market power or even holds a dominant position
in at least one relevant market. It is commonly accepted that where firms do not
possess significant market power, their actions cannot give rise to anti-competitive
outcomes.
•
Condition B: Rival firms are unable to respond. Rival suppliers to the merging
parties will find it unattractive or impractical to respond by adopting a similar
strategy.222 To the extent that rivals are able to adopt similar strategies either by
merging with other firms or by their own teaming arrangements, the competitive
advantages of the hypothesized strategy (e.g. bundling) will be to a large extent
mitigated.
•
Condition C: Competitors are marginalised. As a result of the merger, competing
suppliers will lose volumes to the merged party and as a result are marginalized. It
is important to be clear as to what is meant by the marginalization of competitors.
All price reductions will adversely affect competitors in the sense that they will find it
harder to make sales at the margins that prevailed prior to the price reduction. But,
as noted above, such price decreases do not generally harm competition. A price
reduction can be said to marginalize competitors only if, at any given price level, the
220
It is also possible that non-horizontal mergers can give rise to coordinated effects (see Nocke and White, 2003, and
Church, 2004). This assessment is equivalent to the assessment of coordinated effects under horizontal mergers.
Further, some theoretical papers essentially based on static models suggest that non-horizontal mergers may result in
softened competition (see footnote 218 and also Seidman, 1991).
221
These conditions focus on the foreclosure of existing competitors. Analogous conditions can be derived for strategies
aimed at excluding potential entrants.
222
I.e., a commercial strategy that gives rise to competitive concerns, but that may also generate significant efficiencies
(see above).
Page 123
competitive constraint currently provided by rivals at that price level were to be
reduced following a temporary price reduction.
•
5.5
Condition D: Competition is adversely affected in the long run. As a result of
the above chain of events, competition is adversely affected in the long run. In
consequence, prices will increase and customer interests will be harmed. This can
only occur if competitors are marginalised to such an extent so as to be forced to
withdraw permanently from the market.223 Moreover, it is important to take into
account how long it takes to get to the “long run”. The longer it takes for any long
run adverse effects to arise, the less likely it is that the merger will be anticompetitive, for two reasons. First, it becomes harder to meet the standard set out
by the ECJ in Tetra Laval. Second, since consumers benefit until the long run
adverse effects occur, it is likely that overall the “short run” benefits will outweigh
the “long run” detriments.
Summary and Conclusions
This section has presented two main policy conclusions.
First, that naively applying the two-step framework for assessing efficiencies as presented in the
EC and US horizontal merger guidelines is inappropriate for several reasons. Most notably, in
many instances it is simply not possible to separate the pro-competitive effects arising from a
non-horizontal merger from its potential anti-competitive effects. In such circumstances, it is
difficult to see how a two-step approach would be implemented, even in theory. Rather, in many
instances, efficiencies generated by non-horizontal mergers must be assessed as part of an
integrated competitive analysis.224
Second, non-horizontal mergers can only adversely affect competition indirectly.225 This implies
that where an integrated competitive assessment is required, the economic presumption must
be that non-horizontal mergers are pro-competitive. Indeed, if an integrated assessment is
required, it should be assumed that the efficiency benefits generated by a non-horizontal
merger predominate any potential anti-competitive concerns unless it can be demonstrated
otherwise.
Page 124
223
Competition could also be harmed if it could be shown that marginalisation had the effect of permanently reducing
investment in new products. This is likely to be extremely difficult to prove in practice.
224
Table 1 provides a general categorisation of which theories of competitive harm it is appropriate to apply the two-step
approach.
225
See Section 2 of this report, as well as Section 5.3.4.
6 Case Studies
This section of the report considers case studies of non-horizontal mergers. We look at several
conglomerate and vertical mergers which were considered by the European Commission and
other antitrust bodies. The cases examined are:
•
Guinness/Grand Met;
•
Boeing/Hughes;
•
AOL/Time Warner;
•
Tetra-Laval/Sidel.
For each case, we first outline the background to the merger considered. We then review the
authorities’ treatment of the case. Lastly, we present our own analysis of the case, considering
the efficiencies that could arise as a result of the merger, and how the authorities’ treatment of
these efficiencies compares to our own recommended policy approach. We also consider
whether any remedies imposed on the cases were well suited to the issues they sought to
address.
6.1
Guinness/Grand Met
This section considers the merger of Guinness and Grand Metropolitan (Grand Met). This case
was notified to the European Commission in 1997226 and was also considered by the US
FTC227. It is particularly notable for being the first case in which the Commission acted on its
“portfolio power” theories regarding conglomerate mergers.
6.1.1
The merger
Both Guinness and Grand Met are involved in the production of alcoholic beverages, including,
in particular, spirits.228 The spirits are sold by retailers worldwide. However, in Europe, the two
companies had largely complementary portfolios in that they did not generally have a strong
position for the same type of spirit within the same country. In the US there was a greater
degree of overlap between the two companies.
The spirits are sold to retailers through intermediate distributors. The distributors are also
responsible for promotion and marketing of the brands. In common with many of the major spirit
226
Guinness/Grand Metropolitan Case No. M.938 OJ l288 (1998).
227
Docket No. C-3801
228
In addition, Guinness was involved in the brewing of beer, and had interests in hotels and publishing; Grand Met had
wine, food and retailing businesses.
Page 125
producers, Guinness and Grand Met are vertically integrated and distribute their spirits
exclusively through their own subsidiaries. (As the Commission acknowledged, this allows
better control of the producers’ brand names 229 – see Section 3.4.1). In addition, both
companies also distributed spirits on behalf of other producers, including many of the smaller
producers, although this third-party distribution was not a significant part of their business. The
merger of Guinness and Grand Met thus brought together two vertically integrated spirit firms,
which also provided limited distribution services to third parties. This is illustrated pictorially in
Figure 6-1.
Figure 6-1: Simplified representation of relevant Guinness and Grand/Met businesses
Spirit
Production
Guinness
Grand Met
Spirit
Distribution
Guinness
Grand Met
Spirit
Retailers
Retailers
Retailers
Third
Parties
The companies claimed a number of benefits from the merger:230
• Complementary and broad product and brand range: i.e., the ability to offer consumers the
most attractive brands and an expanded product range across most major spirits categories.
• Greater geographic breadth: The complementary geographic strengths of the companies
would allow greater scale leading to cost efficiencies in the more developed markets and the
necessary critical mass for growth in emerging markets.
• Enhanced marketing capabilities: The merged company would have the benefit of a high
calibre and experienced management team, translating into superior brand management,
innovative new product development and world class marketing.
• Greater cost efficiencies: The companies estimated that the merger would generate
operating cost savings of some £175 million per annum. This was to come from integrating
national sales forces and back offices of the two spirits businesses in countries where both
groups are active, rationalising head offices, and production and purchasing savings.
Page 126
229
The European Commission states in §20 of the decision: “Integration into distribution allows the supplier to maintain
control over the marketing and distribution of the brand which it owns, thereby safeguarding the all-important image of
those brands in the market-place.” Similarly §36 states: ”Vertical integration into distribution is advantageous in product
markets where branding is important, because it allows the brand-owner to retain full control of product development,
promotion and marketing.”
230
Diageo News Release, 12 May 1997 “Proposed merger of GrandMet and Guinness and £2.4 billion capital repayment”
• Financial capacity to develop the business: The greater financial strength would allow the
company to invest in global brands, expand internationally and launch innovative new brands.
6.1.2
Treatment by antitrust authorities
Market Definition
When considering the market definition, the Commission reached the view that each type of
spirit (such as gin, vodka, or whisky) formed a separate market; thus retailers did not consider
vodka a substitute for gin, for example.231 Each country was also deemed to be a separate
market.
In the US, concerns were mainly horizontal. However, within Europe, Guinness and Grand Met
did not generally have a strong position for the same type of spirit within the same country. For
example, Guinness sold very little vodka, and Grand Met sold very little gin; both companies
sold whisky, but did not generally have strong positions in the same countries. Thus, given this
definition of the market, there was a relatively small amount of horizontal overlap between the
two parties and the companies were not deemed to compete to a great extent.
Despite the lack of horizontal overlap, the Commission considered that competition could be
harmed because of the “portfolio power” arising from the combination of a number of strong
products in neighbouring markets. The Commission had touched on portfolio power theories in
previous cases, but Guinness/Grand Met is notable as being the first case in which such
theories had been acted upon. It is these theories on which this case study focuses.
Portfolio Power
The Commission seemed to accept the companies’ claims that merger would lead to efficiency
gains. It suggested that these advantages would arise in a number of ways, in particular:
•
Economies of scale and scope: The Commission acknowledged that cost savings
might arise through the combination of common sales and marketing activities,
reflecting the companies’ claims regarding cost efficiencies and enhanced
marketing capabilities.
•
One-stop shopping: The Commission also suggested that the merged party might
be able to provide one-stop shopping advantages to its customers.232
231
In the US, the product markets were more narrowly defined as, for example, “premium Scotch” and “premium gin”,
although these definitions were contested by some members of the FTC.
232
§40 and §101 of the Commission’s decision.
Page 127
•
Increased ability to tie and bundle: The Commission phrased this as “greater
flexibility to structure his prices promotions and discounts” and “greater potential for
tying”.
•
Increased financial efficiency: The Commission also accepted that the companies’
claim that the merged entity would, by virtue of its size, have better access to
finance.233
The Commission referred to such benefits from providing a wider range of products as “portfolio
power”. However, rather than interpreting the advantages as pro-competitive efficiency gains,
the Commission feared that they could lead to competitive concerns even where the parties’
product ranges did not compete to any great extent. There seem to be several arguments
underlying these concerns.
First, the Commission appears to have been concerned that the merged firm would be able to
offer a more attractive package, and thus further increase its market power by virtue of its
increased efficiency. 234
In addition, the Commission believed the advantages of the merged firm would increase
potential barriers to entry. This is because competing firms would have to enter on a larger
scale in order to provide attractive alternatives with a similar portfolio of products. The
Commission feared that entry on this scale would not be feasible in many cases. 235
The concerns regarding entry were exacerbated by the Commission’s fear that the merged
distributors would refuse to provide distribution services to third party spirit manufacturers. Even
if entry were not possible on a large scale, small producers may still be viable if they could
distribute their products through the subsidiaries of larger companies. However, the
Commission believed that, where the new brand competed with an existing brand offered by the
vertically integrated company, the distributor was unlikely to promote it heavily. That is, the
companies would effectively foreclose potential upstream competitors. 236
Remedies
The Commission claimed that portfolio power posed a particular problem in Greece. In this
market, Guinness has a strong position in whisky, gin and rum, whereas Grand Met was strong
in brandy, ouzo, tequila and liqueurs. Thus the Commission claimed that the merger “fills the
gaps in the respective portfolios of each party. The resulting combined portfolio will be by far
wider and deeper than that of competitors”.237 It believed that these portfolio effects would
reinforce the existing dominant positions in Greece. To allay these concerns, it required the
Page 128
233
European Commission’s decision, §40.
234
European Commission’s decision, §119.
235
European Commission’s decision, §53.
236
European Commission’s decision, §53.
237
European Commission’s decision, §86.
companies to end a distribution agreement for rum, despite the merger creating no overlap in
the supply of rum in Greece.
6.1.3
Analysis
Assessment of Efficiencies
Both the parties and the Commission seemed to be in broad agreement that the merger could
give the parties a number of advantages over its competitors. The claimed benefits to
consumers do seem justified in the case of economies of scale and scope and one stop
shopping, although the case for tying and bundling and increased financial efficiency seems
weaker:
•
Economies of scope and scale: The companies aimed to rationalise their
distribution systems, to exert buyer power to reduce purchasing costs and reduce
production costs. These reflect marginal cost savings which should be passed
through to consumers. Where distribution agreements are very straight forward,
contractual arrangements may provide an alternative means of achieving these
gains. However, if, as in this case, distribution arrangements are more complex
including, in particular, marketing provisions, such contracts are likely to be beset
by incompleteness problems and moral hazard. Indeed, as the Commission
acknowledged, it is for precisely these reasons that spirit distributors are vertically
integrated.
In addition, the companies aimed to rationalise common overheads such as head
office, back office and marketing activities. These represent fixed cost savings,
rather than reductions in marginal costs, and are therefore less likely to benefit
consumers in the short term.
•
One-stop-shopping: As retailers generally buy products from both parties, the
merger could result in one-stop-shopping advantages (although, the Commission’s
analysis provides no discussion of the likely extent of these gains). If competition in
the downstream market is sufficient, these benefits should be passed on to
consumers, at least in part, rather than being fully appropriated by the merging
parties. Again, as with economies of scope and scale, these benefits could in
principle be obtained through an agreement for common distribution. However, the
problems associated with contracting marketing provisions, as described above, are
likely to make such solutions costly and imperfect.
•
Price discrimination: Tying and bundling may be employed by a firm with market
power in order to exercise price discrimination. This may increase social welfare if it
leads to higher output. However, the impact on consumer welfare is generally
ambiguous: some consumers gain and others lose, so that the overall impact may
be negative.
Page 129
•
Increased financial efficiency: In general, increased financial efficiency is most likely
to generate significant gains where cash flows are highly fluctuating and uncertain
and/or when the companies’ balance sheets are distressed. As there is no
indication that such conditions are met, this seems like a weaker justification for
enhanced efficiency.
The claimed economies in scale and scope do appear to have materialised. 238 In addition, the
companies’ shares outperformed the FTSE 100 following the merger suggesting it successfully
added value to the businesses.239 It is interesting to note, however, that rather than using the
supposed increased financial strength to expand its brands, the company actually divested a
number of brands and chose to concentrate on its core products, 240 although a number of new
products such as Smirnoff Ice™ were also launched.
Overall assessment of the transaction
The Commission’s notion that “a portfolio of brands exceeds the sum of its parts” is in direct
juxtaposition to the Chicago School, which claims that generally, there is only one monopoly
rent; i.e., by extending into other markets, there is no possibility of increasing total market
power. Nevertheless, as discussed in Section 2.3.3, there are exceptions to this rule, in which
case bundling or vertical integration can give rise to anti-competitive fears.
In particular, such fears may be valid where the advantage of a merged firm is so great that it
drives its competitors from the market. If they are unable to re-enter, the merged party may then
be able to raise prices and more than offset the effect of the efficiency gain. In this case,
however, there was little suggestion that existing competitors would exit the market. Indeed, the
Commission believed that competitors supplying important brands would face few problems in
finding access to the market post merger. Rival firms, far from exiting the market, may actually
respond with further strategies designed to increase their competitiveness. Its first concern - that
the more efficient firm would be able to offer a more attractive package and so increase its
market power - is therefore not valid. Harm to competitors seems to have been mistaken for
harm to consumers.
Nevertheless, there is a possibility that the merger may help to preserve market power, if not
increase it, by inhibiting further entry and thus reducing consumer welfare. The Commission
identified two mechanisms for this: requiring entry on a larger scale or denying access to
distribution services. However, as the Commission acknowledges, there are many other of
reasons why entry would not take place, even in the absence of “portfolio power” effects;241 it is
therefore unlikely that the merger was motivated by such considerations. Forcing a divestiture of
some brands would therefore preclude efficiency gains with very little chance of increasing
entry.
238
Page 130
Diageo 1998 Annual Review.
239
Financial Times November 23, 2002 “Diageo considers a fire sale in its craving for focus”
240
Diageo Press Release September 29, 1999.
241
European Commission’s decision, §47
Thus, as we suggest in Section 5 of this report, the efficiency benefits and competitive concerns
were not separable in this case. At the route of the Commission’s concern was an efficiency
benefit, and the greater the likely benefits of the merged company, the greater the possible
harm. However, the merger seems unlikely to have reduced the number of players in the market
or to have a significant impact on further entry. That is, even if the company had engaged in
tying and bundling as envisaged by the Commission, it seems unlikely to have reduced
competition. This violates Condition C of our proposed approach to assessing efficiencies –
which requires that competitors are indeed marginalised by the proposed anti-competitive
behaviour - suggesting the merger should have been cleared.
Remedies
It has been suggested that, were the exclusionary fears valid, a more appropriate remedy might
have been to require that the distribution of the merged firm not be exclusive.242 This would
allow the merged party to benefit from the efficiency gains but nevertheless allow a means of
distribution for new entrants. However, this is unlikely to present a viable alternative as it would
be very difficult to ensure that access would be on equal terms, in particular with regard to the
promotion of the new entrant’s brand names. Indeed, as the Commission recognised, it is
precisely such problems that necessitated the vertical integration in the first place.
Conclusion
In conclusion, the merger seems likely to have had a pro-competitive effect. At least some of the
claimed efficiencies (notably economies of scale and scope) appear to have materialised and
should have resulted in lower prices to consumers. The Commission’s concern that the merger
would marginalise competitors seems unwarranted as rivals were unlikely to exit the market.
Rather they could have been expected to respond by lowering prices or engaging in efficiency
enhancing strategies of their own. The concerns regarding the deterrence of new entrants are
doubtful as entry barriers were high and entry unlikely pre merger. Forcing divestiture therefore
forsakes the efficiency benefits for very little gain.
6.2
Boeing/Hughes
This section considers the merger between the Boeing Company (Boeing) and Hughes
Electronics Corporation (Hughes), which was announced in 2000. The merger was investigated
and cleared on both sides of the Atlantic.243 The EC’s treatment of the merger provides an
interesting example of a case where a non-horizontal concern was considered but then
dismissed because the particular conditions required for anti-competitive effects to emerge were
not present in the market.
242
Nalebuff and Majerus (2003), part 2, §85.
Page 131
6.2.1
The merger
Boeing is active in the commercial aircraft, defence and space industries. Amongst other
activities, it supplies navigation satellites and satellite launch services. Hughes is also active in
satellite-based services. In particular, Hughes Space and Communications Company (HSC) is
the world’s leading manufacturer of commercial GEO satellites. Hughes also produces certain
satellite equipment and is active in the downstream markets for satellite-based communication
services and pay-TV. Boeing identified substantial synergies from the merger in terms of both
production and intellectual capital claiming they would bring “unmatched capability to integrate
space, air and terrestrial information communications systems to provide seamless services”.244
6.2.2
Treatment by antitrust authority
Treatment by the European Commission
The EC considered that satellites did not form a single product market, but could be broken
down into different sectors. Although Boeing and Hughes both manufactured satellites, they
were active in different markets and the merger did not lead to any direct overlaps. Thus, there
were no material horizontal concerns. In addition neither Hughes nor Boeing were deemed to
have a dominant position, despite Hughes having shares of 35-40% in the commercial GEO
satellite market.245 (This lack of market power was based on countervailing buyer power, as well
as on the availability of alternative satellite products.)
The main concerns in the case focused on the combination of the parties’ satellite
manufacturing activities and launching activities. The EC referred to this as vertical integration,
but, since customers usually purchase launch services separately from the satellite concerned
(Delivery on the Ground or “DOG”), it would appear more accurate to describe the case as a
conglomerate integration between complementary products. (Some smaller customers buy a
complete package, known as Delivery in Orbit or DIO, and from these customers’ perspective
the merger could be viewed as a vertical merger.) However, as noted throughout this report,
integration between complementary products and vertical integration bear strong similarities and
the distinction is therefore not important. A simplified graphical depiction of the relationship
which gave rise to the concerns is given in Figure 6-2.
Page 132
243
European Commission (2000), Case No. Comp/m.1879 – Boeing/Hughes, Federal Trade Commission, (2000), File No.
001-0092
244
Phil Condit, Chairman & CEO, Space.com news release, January 20, 2000, Boeing News release, Oct 6, 2000, “Boeing
Concludes Acquisition of Hughes’ Space and Communications Businesses” .
245
European Commission’s decision, §93.
Figure 6-2: Simplified Representation of Relevant Boeing/Hughes Businesses
GEO Satellite Manufacturers
Launch Service Providers
HSC
Boeing
Third Parties
Third Parties
Parties
DIO
Customers
DOG
Customers
The EC and the FTC highlighted several potential concerns with the case, relating to Hughes’
ability to leverage potential market power from the GEO satellite market into the market for
launchers. Several mechanisms were suggested for this:
•
Technical Bundling/Raising Rivals Costs: The merged company could design
current and future generation satellites, or withhold information with regard to their
designs, so that they are less compatible and more costly to integrate with third
party launchers.
•
Vertical Foreclosure: Where DIO customers allow some flexibility regarding the
choice of launch vehicle, Hughes might have all satellites launched on Boeing’s
products;
•
Information Spillovers: Hughes could use sensitive information regarding third party
launchers to Boeing’s advantage. (Similarly, Boeing could use information
regarding other satellites to increase sales of Hughes’ satellites.)
However, the European Commission considered that Hughes was not likely to undertake such
behaviour. This was essentially because Hughes lacked the power in the satellite market for any
such anti-competitive behaviour to be profitable. Thus, the Commission stated that:
“the merged entity will not be in a position to lead a substantial number of customers to
switch to Boeing or Sea Launch vehicles if that were not their initial intention.”246
246
European Commission’s decision, §82
Page 133
It therefore concluded that:
“making integration between HSC satellites and non-Boeing launchers more difficult
would either have no impact on the customer choice or would make launcher
combinations with HSC relatively more expensive than with other satellites, thereby
weakening HSC’s competitive position in satellites”.247
It also concluded that DIO customers buying vertically integrated products would be in a position
to independently select their launch service operator, so that Hughes would not be able to
increase sales of Boeing in integrated contracts.
The exclusionary fears were further allayed by the parties’ undertakings to ensure that Boeing
would not obtain an advantage over other launch suppliers through better information or
integration with Hughes satellites, but these were not a condition of the EC’s approval of the
merger.
Despite the lack of market power in the satellite market, the merger could still have led to anticompetitive effects because of fears regarding spillovers (see Section 4.4.2). If competitors
feared that the merger between Boeing and Hughes would result in leakages of their
innovations, their incentives to engage in R&D could be reduced. However, the merging parties
allied these fears by further undertakings to establish a firewall between the satellite and
launcher divisions to guard against this, although in the EC these undertakings were again
voluntary and not a condition of the merger.
Third parties also claimed that the current over-capacity and the substantial fixed costs involved
in satellite production meant that the loss of even a few contracts could have a material effect
on competition. They suggested that losing contracts would result in them being unable to cover
fixed costs and so becoming even less competitive in future. Such “snowball effects” were
dismissed because economic theory suggests that the ability to make competitive bids should
depend largely on marginal costs rather than fixed costs and marginal costs should be
unaffected by the loss of other contracts. Indeed, a high fixed cost/low marginal cost
environment increased competitive pressure from Boeing/Hughes will tend to make rivals bid
more aggressively for future contracts so as to maximise its recovery of fixed costs.
Treatment by the Federal Trade Commission
The decision in the US contrasted with that of the EC in that, despite the absence of any explicit
comment regarding Hughes’ ability or incentive to foreclose the market, undertakings regarding
compatibility and information of Hughes’ satellites with third party launchers were a condition of
the clearance, rather than a voluntary undertaking.
247
Page 134
European Commission’s decision, §87.
There was also a greater focus on spillovers effects than in the EC. The FTC feared that leaking
of confidential information could have immediate effects because Boeing may bid less
aggressively in upcoming satellite procurements if it had information regarding other satellite
suppliers’ proposals. Moreover, in the long term, if competing satellite producers feared that
Boeing would leak information regarding their innovations to Hughes, their incentives to
innovate would be reduced. However, again, an agreement enforcing firewalls was considered
to remedy these problems although, as with the other undertakings, these were a condition of
the clearance rather than a voluntary undertaking as in Europe.
6.2.3
Analysis
Analysis of Efficiencies
Although the companies do not provide many details of the proposed efficiencies, their claims
that the merger would allow economies of scope production and intellectual capital to improve
integration of the systems seem plausible. In order to launch a satellite, suppliers of launchers
and satellites must work closely together and are therefore likely to own assets which could be
employed in both industries, such as specialised human capital relating to satellite design.
Contractual agreements are unlikely to provide an alternative means of sharing these assets
because of complexity and substantial sums involved suggest a high risk of disagreements
regarding the sharing of the rents or residual rights or other moral hazard problems.
The merger could also conceivably allow better coordination of R&D, for example by aligning
the objectives of R&D projects, ensuring better compatibility between the products and reducing
extra costs of adjustment. It may also increase R&D spending by internalising the benefits. This
is because in a vertically separated environment, an innovation in satellites will be reaped partly
by manufacturers of launch vehicles, and vice-versa. Thus, as discussed in Section 4.4.3,
internalising such benefits by bundling the goods can lead to efficiency gains. These gains are
likely to be larger to the extent that the products are strong complements. However, in as much
as the merged party undertook to sell only mixed, and not pure, bundling, part of the benefits of
innovation will still have been reaped by competitors. Again, moral hazard and contractual
incompleteness are likely to make contractual agreements an unsuitable alternative.
In the event, however, there were questions over the success of the merger. In part, this may
have reflected the timing of the merger; by 2001 there was substantial overcapacity in the
market. In addition, however, cultural differences between the two companies may have
prevented the envisaged sharing of intellectual capacity from being achieved.248
Assessment of the Transaction
The Commission’s assessment of the merger correctly recognises the weaknesses in the theory
of harm, which would not have satisfied the four conditions for assessing efficiencies proposed
248
Financial Times August 5, 2003 “Space and defence prove a difficult orbit for Boeing”
Page 135
in Section 5 of this report. Condition A - which requires the existence of significant market power
- was clearly violated as Hughes was not deemed dominant.
In addition, Condition C which requires that competitors are indeed marginalised by the
proposed anti-competitive behaviour would also have been violated. The exclusionary strategy
could not have led to the monopolisation of the market as the Commission believed that the
“snowball” effects put forward by third parties would not materialise. Moreover, in order to result
in competitive harm, the merged firm must have been able to precommit to an exclusionary
strategy. Without such precommitment, an increase in price of launchers will mean that
customers could insist on buying a Hughes’ satellites but using a third party launcher. The
Commission decided that it would not be feasible for the merged party to credible threaten to tie
its products as customers require the flexibility to change launchers.
In the EC, therefore, the merger was cleared, although voluntary undertakings were offered
regarding the prevention of exclusionary conduct and leakages of confidential third party
information. In the US these undertaking were compulsory. These firewalls may have
encouraged R&D by third parties. However, they may also have prevented the realisation of
further efficiency gains, as the merging parties may have been more inclined to innovate if the
were better able to prevent competitors free riding on this benefit.249 Although this area is still
open to debate, some authors have suggests that such firewalls could have reduced social
welfare, by decreasing total R&D investment.250
6.3
AOL/Time Warner
This section considers the merger between AOL and Time Warner which was notified to the
European Commission in 2000.251 The case was interesting for several reasons. First, at the
time of notification the proposed merger was the largest in corporate history. Second, the theory
of harm involved “network effects” which the Commission feared would amplify upstream market
power because a greater number of upstream firms providing content to AOL’s services
attracted more downstream consumers to AOL which in turn attracted more content providers
and so on. Finally, the high hopes the companies had for the merger failed ever to materialise.
6.3.1
The merger
The primary business of AOL is internet on-line services. As an Internet Service Provider
(hereafter ISP), it offers dial-up internet connections which allow users to access content, in
some cases proprietary to the ISP. This content is provided by media publishers, including Time
Page 136
249
The commitments regarding its continuing relationship with third parties also implied that a number of other efficiency
gains were not realisable. These include eliminating quality uncertainty, discussed in section 4.4.1. Similarly, given the
undertakings regarding supply of information, integration to protect innovation, discussed in 4.4.2, cannot be a concern.
250
See discussion of Milliou (2004) in Section 3.4.2.
251
European Commission (2000), Case no Comp/M.1845 – AOL/Time Warner. The case was also considered in the US
by the FTC and the FCC (Federal Communications Commission), but raised mainly horizontal issues.
Warner. The media group Bertelsmann also provides content which can be accessed over the
internet and at the time of the merger notification it had a joint venture with AOL and a further
agreement relating to joint promotion, distribution and sales. The merger would therefore
contain a horizontal element, as it would bring together Time Warner and - through AOL’s joint
venture - Bertelsmann. However, it would allow also Time Warner to integrated forwards into
AOL’s ISP services. This is illustrated in Figure 6-3.
Figure 6-3: Simplified Representation of Dial-up Internet Market
Content
Providers
Narrow Band
ISPs
Bertelsmann
Time Warner
AOL
Third Parties
Third Parties
Consumers
AOL also offers software, including, in particular, the Winamp music player. Music players are
software devices incorporated into the PC that can be used to play music distributed by ISPs. In
order for the ISP to supply music on-line, it must first obtain the appropriate rights. In most
cases these rights are assigned to the music publisher. The music recording is then converted
into a suitable electronic form and can then be supplied through narrow band ISPs. ISP
customers can access the recordings with using a music player. A graphical depiction of the online music market is shown in Figure 6-4.
At the time of the investigation, music players were usually given free over the internet. There
were various formats for music files and music players supported a number of different
technologies. AOL and Bertelsmann had agreed to work together to ensure that Bertelsmann’s
music was formatted to be playable by AOL’s Winamp.
Page 137
Figure 6-4 Simplified Representation of AOL/Time Warner On-Line Music Businesses
On-line
Rights
Holders
BMG
(Bertelsmann)
Warner Music
(Time Warner)
ISPs
AOL
Third Parties
Music Player
Producers
Winamp
(AOL)
Third Parties
Third Parties
Consumers
6.3.2
Treatment by antitrust authorities
Introduction
The European Commission had concerns regarding the general market for dial-up or narrow
band internet access as well as the market for on-line music which was delivered through these
connections. The concerns in both cases centred on the presence of network effects.
Market Definition
The Commission believed that the market for narrow band internet access was national (and
separate from the broadband market which was still emerging at the time). Although AOL was
active across Europe and was the largest ISP in the world at that time, it had less than 10% of
the customer base in each country (with the exception of the UK), and was not the largest
internet provider in any country. In the UK, AOL had around 15-25% of active subscribers (but
was still only the second largest ISP in the market).
The Commission decided that the market for “on-line music delivery” was least EEA wide.
Bertelsmann and Time Warner would together have had 30-40% of music publishing rights in
the EEA. The market for music players was considered global; AOL’s Winamp had a market
share of 10-20% of on-line music listeners.
Page 138
Concerns
The Commission had some horizontal concerns, as the merger would bring together Time
Warner and - through AOL’s JV - Bertelsmann. The Commission feared that a single entity
controlling such a sizeable music catalogue could exercise market power.252 However, the main
focus of the analysis related to the ability of the merged company to “leverage” this power on to
the market for the on-line music market as well as the market for dial-up internet connections. In
particular, it was feared that the presence of “network effects” in these downstream markets
might increase the upstream concentration and create a dominant position downstream.
In the case of narrow band internet access, the Commission’s fears were as followed. First,
AOL would tie Time Warner’s and Bertelsmann’s music portfolio with internet access and
proprietary services and give its subscribers exclusive or preferential access to that content.
This “input foreclosure” would then mean that rival ISPs are less attractive, as they cannot
supply such a broad range of music, i.e., there is a “network” effect. Finally, this effect would be
amplified as the greater number of consumers which would be attracted to AOL would make
AOL more attractive to other upstream content providers, which would then attract more
consumers and so on. Thus, the Commission states that “competing record companies would
feel obliged to joint [sic] AOL, which would end up by having access to all the music.”253 The
Commission feared that this might amplify AOL’s share in the UK market for narrow band
internet access sufficiently to give it a dominant position.
The Commission had similar fears in the market for music players. It suggested that the merged
company would be able to “leverage” market power in the music publication market on to the
market for on-line music delivery and music players by foreclosing its downstream competitors
through technical tying. The Commission argued that network effects would then grant it a
dominant position in the downstream markets by reasoning as follows:
•
Technical tying of on-line delivery: The Commission suggested that the merged
company would develop a proprietary technical standard for on-line music delivery.
It would be able to impose this as industry standard, for example by threatening not
to license the technology to music players which also supported competing
technologies. This argument assumes an implicit network effect: music players
which do not support Time Warner/Bertelsmann’s music are not as attractive as
those that do. Competing record companies wishing to distribute their music would
then be required to use this proprietary technology. This would lead the new entity
to become dominant in the market for on-line music delivery and would allow it to
raise its competitors’ costs through excessive licence fees for the technology.
•
Technical tying of music players: Rather than developing a proprietary technology,
the merged company could make Time Warner and Bertelsmann on-line music
delivery compatible only with Winamp. The Commission argued that:
252
European Commission’s decision, §47.
253
European Commission’s decision, §84.
Page 139
“as a result, Winamp would become the only music player in the world able
to play virtually all the music available on the Internet. The other record
companies would have no incentive to format their music using proprietary
standards because they do not have a music-player software (such as
AOL’s Winamp), do not have an on-line music distribution network
comparable to AOL’s, and do not have the same critical mass in terms of
back-catalogue. By refusing to license its technology, the new entity would
impose Winamp as the dominant music player as no other player would be
able to decode the proprietary format of Time Warner and Bertelsmann
music. Given their technical limitations, competing music players will exert
no competitive constrain on the pricing of Winamp. Therefore, as a result of
the transaction the new entity will control the dominant player software and
could charge supra-competitive prices for it. “
Again, there is an implicit network effect behind this argument: the more technology
with which the music player is compatible, the more attractive it is.
Decision
The Commission approved the merger. However, in order to remedy the problems in the on-line
delivery of music, the parties severed all structural links with Bertelsmann. This also resolved
the problems in the ISP market, which removed the need for a separate assessment of the
issues raised there.
6.3.3
Analysis
Competitive concerns
The most common critique to leveraging theories – the “one-monopoly profit” principle – does
not apply to the case of AOL/Time Warner, for two reasons. First, AOL/Time Warner, even if
dominant in one market, lacked monopoly power. Since it could not earn the full monopoly profit
in the upstream market, in principle it could have increased its profits by monopolising a
downstream market. Second, ISPs have “secondary uses” besides music downloading.
Consumers use ISPs to access not only music, but also other contents, many of which were not
supplied by AOL/Time Warner. As a result, the profits of a hypothetical monopolist of the ISP
market are significantly larger than the monopoly profits that can be extracted from the market
for music downloads alone. (The limitations of the “one-monopoly profit” principle are discussed
in Section 2.3.3.)
However, this does not suffice to prove that AOL/Time Warner would have the incentive to tie
the products in order to leverage its market power. For example, reducing availability of
Bertelsmann/Time Warner content by allowing it to be accessed only through AOL or on
Winamp will simply reduce sales of Bertelsmann/Time Warner music. Similarly, making Winamp
or AOL compatible only with AOL’s proprietary standards will make these products less
Page 140
attractive to downstream consumers. Thus, establishing whether AOL/Time Warner had the
incentive to tie its product is a very difficult exercise.
But had AOL/Time Warner the ability to leverage its market power through tying? The addition
of network effects in the downstream markets creates an interesting twist which the Commission
believed would give the merged firm the means to gain dominance in the emerging market for
downloadable music or ISPs. The Commission feared that - having integrated into the
downstream market - the merged firm’s market share may be “amplified” through network
effects. As the firm gains more and more market share, it becomes more attractive and this
snowball effect may eventually drive other downstream firms out of the market.254
The fact that the markets were still emerging at the time of the enquiry meant that any
predictions about how it would have developed would have been necessarily highly speculative.
Nevertheless, the existence of such “snowball effects” does seem rather tenuous, and could
underestimate competitors’ ability to prevent AOL from gaining the critical mass required for
such effects to materialise.
Competing ISPs or music player producers (which already have the advantage of a larger
customer base) could, for example, develop their own proprietary standard. Competing music
publishers would have an incentive to support this standard over AOL’s, despite not being
vertically integrated, as they would recognise the importance of network effects The upstream
and downstream rivals may even consider responding with a counter merger.
The role of other players in the industry such as software producers and music retailers - many
of which already have substantial market power - may also be important in preventing AOL from
establishing a dominant position. Indeed, in this event it was Apple and Microsoft that
established standards for music delivery, in competition with one another. Apple introduced a
proprietary standard for its iTunes compatible only with its portable music player, the iPod.
Microsoft also introduced its Media Player (although it bundled this with its operating system
prompting the Commission’s intervention) and downloaded music encrypted with Windows
Media Audio format is compatible with a number of different portable players but not Apple's
iPod. The Windows format is supported by various on-line music stores which compete with
iTunes. Thus, to date, the market is supporting two competing standards and it seems that the
snowball effect has yet to materialise (although the development is still at an early stage and
may have been affected by further Commission intervention).
254
At times the Commission’s analysis does seem to slip back into arguments which contravene the basic Chicago
principle, suggesting, for example that AOL, having attained a dominant position, could use this “to restrict access to
[t]he on-line music market by favouring Time Warner and Bertelsmann and degrading the quality of access for
competing content providers” (§54). If AOL managed to obtain market power, it can extract all rents through charging
for access to its network, and generally has no further incentive to reduce the attractiveness of its service by degrade
access to competing content providers. This is all the more true given the network externalities in content provision.
Page 141
Efficiencies
The network externalities on which the Commission’s theory relies can be thought of as a
benefit to consumers. Access to a wider range of content, either in the general dial-up market or
the market for on-line music, increases the value of the service to consumers. Indeed, it is this
benefit which implies that competing firms will not be attractive and so be driven from the
market. The Commission acknowledges that the network effects provide a benefit to
consumers. For example, in relation to dial-up services it states that “content should become the
crucial factor determining an ISP’s competitiveness.”255 Assuming non-proprietary standards are
ruled-out for example by R&D spillovers or the hold-up problem, the consumer therefore obtains
a benefit from having a single provider which allows access to all content. Conversely, upstream
content providers benefit by being able to access a wide range of consumers from a single
provider. Of course, this benefit will be offset by the possibility that the dominant provider will
use its market power to raise prices. However, in theory at least, the benefit of having a single
provider may outweigh the increased prices, although it seems unlikely in practice and would be
almost impossible to test.
The companies also believed that the merger of a content provider with a supplier of the new
delivery channel could provide other benefits. It envisaged that the economies of scope in
sharing human capital could have led to innovation which would drive the emerging market
forward. For example, Time Warner stated that “by joining the resources and talents of these
two highly creative companies, we can accelerate the development and deployment of a whole
new generation of interactive services and content”, and it also promised to “fundamentally
change the way people get information, communicate with others, buy products and are
entertained”.256
JVs and joint marketing agreements of the type that AOL had with Bertelsmann could have
been sufficient to achieve some of these goals. However, fears regarding spillovers, hold-up
problems, or negotiations over the appropriation of the gains may increase transaction costs
and prevent the companies from working as closely together as under a full merger.257
Moreover, as recognised in the Commission’s analysis, agreements of the kind instituted
between Bertelsmann and AOL may present much the same type of competitive problem.
In the event, however, many of these gains seem to have failed to materialise. There was some
cross-promotion but the dotcom bust which followed the merger saw a decrease in the
importance of ISPs in selling advertising space. More significantly, the promised new products
failed to materialise.258 There were cultural differences between the companies, which may
Page 142
255
European Commission’s decision, §77.
256
Time Warner Press Release “AOL & Time Warner Will Merge To Create World’s First Internet-Age Media &
Communications Company”, January 10, 2000.
257
Indeed, Bertelsmann suffered precisely such problems in its agreement with AOL. Prior to the AOL/Time Warner deal, it
had begun what it probably expected to be a long-term course of dealing with AOL by investing in building the AOL
Europe brand. As a private-held company Bertelsmann may have preferred partnership arrangements to transactions
involving change of control of its business. However, the EC’s remedy would have deprived Bertelsmann of the value of
that investment, illustrating the strategic risk that such arrangements entail (see Cook, 2001).
258
See e.g. http://news.bbc.co.uk/1/hi/business/1754702.stm.
have hindered the close relationships required to capitalise on the economies of scope
promised by companies’ pooled resources. 259 Finally, coloured by dotcom mania, the
envisioned gains may simply have been overly optimistic.
Conclusions
The case illustrates the difficulties which arise in assessing both efficiencies and competitive
harm of mergers in emerging markets. In this case, the Commission, as well as the integrating
firms, had to make some kind of judgement of the way in which the market would develop. This
complicates the analysis of the case. In particular, it is necessary to consider both the benefits
which the merger could have in driving the market forward with the risk that this foothold will
allow the merged firm to influence the market’s development and increase its market power. But
in making the assessment of the developing market it is important to bear in mind that it is the
impact on consumers rather than competitors that is important. The Commission’s analysis may
have taken too little account of rivals’ ability to respond.
As a remedy the Commission suggested that AOL sever its links with Bertelsmann. It could be
that the horizontal concerns were sufficient to justify this remedy. However, the highly
speculative non-horizontal arguments are not sufficiently robust to overturn the economic
presumption of pro-competitiveness. In particular, it was not clear that the rivals would be
unable to respond to the actions of the merging parties (i.e., Condition B of our proposed
conditions for assessing non-horizontal mergers was not satisfied.)
6.4
Tetra-Laval/Sidel
This section considers the acquisition of Sidel by Tetra Laval, which was notified to the
European Commission in 2001 once the concentration had been completed. 260 The European
Commission declared the merger incompatible with the common market and required Sidel to
be divested. Its concerns included the possible anti-competitive effects of tying and bundling.
However, the decision was appealed at the Court of First Instance (CFI) and was annulled in full
in 2002.261 The Commission re-examined and cleared the case but at the same time it appealed
the CFI’s ruling to the European Court of Justice (the ECJ). However, ECJ rejected the
Commission’s arguments. 262
6.4.1
The merger
The merging parties are active in the packaging market for liquid food products, in particular
plastic and carton packaging. Packaging products are manufactured and used with specialised
259
See e.g. http://news.bbc.co.uk/1/hi/business/2708965.stm.
260
Tetra Laval/Sidel Case No. COMP/M.2416 (2001).
261
Commission v Tetra Laval Case No. T-80/02 and T-5/02 (2002).
262
Judgment of the Court in case C-12/03 P, February 15, 2005.
Page 143
packaging machinery. In particular, plastic bottles are made using specialised Stretch Blow
Moulding (SBM) machines. Customers may purchase these bottles ready-made and fill them
using specialised filling machines. In other cases, customers only buy plastic “preforms” (empty
plastic tubes) and manufacture the bottle themselves using an in-house SBM machine.
Tetra Laval produces carton packaging machines and cartons and has more limited activities in
plastic packaging. Sidel is involved in the design and production of packaging equipment and
systems, and in particular the SBM machines for use in PET filling lines. The diagram below
shows a simplified version of the merged parties’ businesses.
Figure 6-5: Simplified representation of relevant Tetra Laval and Sidel businesses
Plastic packaging equipment
(incl. SBM machines)
Carton packaging equiplemt
Sidel
Third Parties
Plastic
packaging
Tetra Laval
Cartons
Tetra Laval
Third Parties
(Converters)
Preform
Tetra Laval
Bottles
Customers
Customers
The parties claimed the acquisition would enable them to provide the world's leading producers
of liquid food with state of the art plastics equipment.263 However, the Commission was
concerned about the effect of creating a new entity which was vertically integrated in both carton
and plastic packaging.264
Page 144
263
Tetra Laval Press Release January 14, 2003 “EU Commission approves Tetra Laval's acquisition of Sidel”
264
The merger also raised horizontal concerns in several areas of packaging technology.
6.4.2
Treatment by antitrust authorities
The European Commission considered that plastic and carton packaging material were
currently distinct relevant product markets. It believed that Tetra was dominant in carton
packaging, but Sidel was not found to be dominant in SBM machines.
However, the Commission expected that, in future, carton would compete with PET. In particular
it believed that there were certain “sensitive” products which were currently only suitable for
carton packaging but for which plastic technology may exist in the future. The Commission
therefore believed that the merger may eliminate potential competitors. In addition, it suggested
that the convergence in the market would allow Tetra to leverage its dominant position in carton
packaging into the market for PET packaging equipment. The Commission suggested the
merged company would be able to identify producers of “sensitive” goods which wanted to
switch from carton to PET packaging, and offer these customers renewed contract terms which
included the purchase of PET equipment from Sidel. Such bundling would allow it to gain a
dominant position in the market for PET packaging equipment, in particular the sale of SBM
machines to the sensitive segment, as a result of the dominant position in the carton market. 265
266
The CFI overturned the Commission’s theory of harm on two principle grounds. First, it believed
the Commission had overestimated the growth in the PET market and the convergence of the
two markets. This weakened the case regarding the elimination of potential competition and
also reduces the potential for the merged party to bundle sales of SBMs and cartons. Second,
the CFI believed the Commission’s decision did not contain sufficient evidence to justify the
definition of a distinct SBM-machine market for sensitive products. Without such a distinction,
Sidel would face competition from suppliers to the non-sensitive market, so that high sales in
the sensitive market alone could be presumed to confer market power. The ECJ’s judgement
reinforced these points.
In addition, the CFI also criticised the Commission for not taking account of commitments
offered by Tetra which specified that it would not engage in tying and bundling behaviour that
would amount to an abuse of its market power in cartons. In its appeal to the ECJ, the
Commission complained that it could not be expected to assess whether the threat of Article 82
legislation was strong enough to dissuade Tetra from engaging in such behaviour. The ECJ
agreed with the Commission that in general it could not be expected to assess whether or not
the merged party would engage in abuse of dominance, suggesting that reliance on Article 82 in
itself is not enough to clear a potentially problematic merger. However, given that Tetra had
offered behavioural commitments, the ECJ said that the Commission should have taken these
into consideration as remedies for the feared anti-competitive effect.
265
European Commission’s decision, §261, §331 and §343.
266
There were also concerns regarding the vertical foreclosure of third party converters through the tying of SBM and
preforms but these were allayed by Tetra’s undertakings to divest its perform business.
Page 145
6.4.3
Analysis
The merged entity could derive efficiency benefits from operating in separate but closely related
markets of plastic and carton packaging. For example, Tetra Laval enjoyed a good reputation in
aseptic cartons, which it could transfer to the market for PET packaging. Indeed, the European
Commission decision concluded that reputation and proven ability were the most important
factors in the market for aseptic machinery and packaging.267 Tetra also had a number of very
successful selling strategies which could be applied to Sidel’s products. Finally, the merged
entity could be expected to have a number of common customers who may benefit from
reduced transaction and marketing costs.
However, despite these efficiency gains, the European Commission believed that the merger
would result in competitive harm through the elimination of potential competition and leveraging
by way of bundling and tying. There was also some suggestion that these problems would be
reinforced by the efficiency gains. For example paragraph 9 of the ECJ’s judgement states that,
according to the Commission:
“Tetra would be helped in this by its close and sustained relationship with its customers,
its financial strength, its know-how and its reputation in the aseptic and ultra-clean
sector, by Sidel's current strength, technology and reputation for quality and by the
vertical integration from which the entity emerging from the notified merger (the merged
entity') will profit in relation to the three packaging systems (carton, PET and HDPE).”
In the main, however, the Commission’s theory of anti-competitive harm rested on speculative
predictions regarding the long-term development of the market, which were insufficient to
overturn the economic presumption of pro-competitiveness. This was clearly recognised by the
ECJ’s judgement which stated that:
“the analysis of a ‘conglomerate-type’ concentration is a prospective analysis in which,
first, the consideration of a lengthy period of time in the future and, secondly, the
leveraging necessary to give rise to a significant impediment to effective competition
mean that the chains of cause and effect are dimly discernible, uncertain and difficult to
establish. That being so, the quality of the evidence produced by the Commission in
order to establish that it is necessary to adopt a decision declaring the concentration
incompatible with the common market is particularly important since that evidence must
support the Commission’s conclusion that, if such a decision were not adopted, the
economic development envisaged by it would be plausible.”
In particular, the lack of convergence in the two markets and the non-existence of a distinct
market for consumers of “sensitive” products suggest that competitors are unlikely to have been
marginalised by any bundling or tying. That is Condition C of our proposed assessment of nonhorizontal mergers was not satisfied and the merger should have been cleared. This conclusion
can only have been reinforced by Tetra’s undertakings not to engage in bundling and tying.
267
Page 146
European Commission’s decision, §382 and §383
Bibliography
Adams, William J. and Yellen, Janet L. (1976), “Commodity bundling and the burden of monopoly”, Quarterly Journal
of Economics, Vol.90, pp.475-498
Adelman, M. A (1955), “Concept and statistical measurement of vertical integration”, in: Stigler, George J., ed.,
Business concentration and price policy, (Princeton, New Jersey: Princeton University Press)
Alfter, Mette, Bishop, Simon and Mahr, Eric (2004), “The Commission’s horizontal merger guidelines – Plugging the
gap or opening the floodgates?”, Business Law International, forthcoming
Armstrong, Mark and Sappington, David (2003), “Recent developments in the theory of regulation”, in: Armstrong,
Mark and Porter, Rob, eds., Handbook of Industrial Organization, Vol. III, (Amsterdam, North-Holland:
Elsevier)
Arrow, Kenneth. J. (1975), “Vertical integration and communication”, Bell Journal of Economics, Vol.6, pp.173-183
Avenel, Eric and Barlet, Corinne (2000), “Vertical foreclosure, technological choice, and entry on the intermediate
market”, Journal of Economics & Management Strategy, Vol.9, pp.211-230
Bakos, Yannis and Brynjolfsson, Erik (1999), “Bundling information goods: Pricing, profits and efficiency”,
Management Science, Vol.45, pp.1613-1630
Banerjee, Samiran and Lin, Ping (2003) “Downstream R&D, Raising rivals’ costs, and input price contracts”,
International Journal of Industrial Organization, Vol.21, pp.79-96
Bernanke, Ben S. and Gertler, Mark (1986), "Agency costs, collateral, and business fluctuations", American
Economic Review, Vol.79, pp.14-39
Bernanke, Ben S. and Gertler, Mark (1987), "Financial fragility and economic performance", Quarterly Journal of
Economics, Vol.105, pp.87-114
Bernheim, Douglas B. and Whinston, Michael D. (1990), “Multimarket contact and collusive behavior”, RAND Journal
of Economics, Vol.21, pp.1-26
Berry, Steven T. (1994), "Estimating discrete choice models of product differentiation", RAND Journal of Economics,
Vol.25, pp.242-262
Bishop, Simon and Lofaro, Andrea (2004), “A legal and economic consensus? The theory and practice of
coordinated effects in EC merger control”, Antitrust Bulletin, March, pp.195-242
Bishop, Simon and Walker, Mike (2002), The economics of EC competition law, 2nd ed, (London: Sweet & Maxwell)
Bolton, Patrick and Bonanno, Giacomo (1988), “Vertical restraints in a model of vertical differentiation”, Quarterly
Journal of Economics, Vol.103, pp.555-570
Bolton, Patrick and Whinston, Michael (1993), “Incomplete contracts, vertical integration and supply assurance”,
Review of Economic Studies, Vol.60, pp.121-148
Bonanno, Giacomo and Vickers, John (1988), “Vertical separation”, Journal of Industrial Economics, Vol.36, pp.25765
Bork, Robert (1978), Antitrust paradox, (New York: Basic Books)
Bresnahan, Timothy F. (2001) “Network effects and Microsoft”, Stanford Institute for Economic Policy Research
Discussion Paper No. 00-51, http://www.stanford.edu/~tbres/Microsoft/Network_Theory_and_Microsoft.pdf
Burstein, Meyer L. (1960), “A theory of full-line forcing”, Northwestern University Law Review, Vol.55, pp.62-95
Butz, David A. and Kleit, Andrew N. (2001), “Are vertical restraints pro- or anti-competitive? Lessons from Interstate
Circuit”, Journal of Law & Economics, Vol.44, pp.131-59
Carbajo, Jose, de-Meza, David and Seidmann, Daniel J. (1990), "A strategic motivation for commodity bundling",
Journal of Industrial Economics, Vol.38, pp.283-298
Carlton, Dennis W. (1979), “Vertical integration in competitive markets under uncertainty”, Journal of Industrial
Economics, Vol.27, pp.189-209
Carlton, Dennis W. and Waldman, Michael (2002), “The strategic use of tying to preserve and create market power in
evolving industries”, RAND Journal of Economics, Vol.33, pp.194-220
Chandler, Alfred D. Jr. (1969), Strategy and structure: Chapters in the history of American industrial experience,
(Cambridge, Massachusetts: MIT Press)
Chen, Yongmin (1997), "Equilibrium product bundling", Journal of Business, Vol.70, pp.85-103
Page 147
Chen, Yongmin (1999), “Oligopoly price discrimination and resale price maintenance”, RAND Journal of Economics,
Vol.30, pp.441-455
Chen, Yongmin (2001), “On vertical mergers and their competitive effects”, RAND Journal of Economics, Vol.32,
pp.667-685
Chipty, Tasneem (2001), “Vertical integration, market foreclosure and consumer welfare in the cable television
industry”, American Economic Review, Vol.91, p.428-453
Choi, Jay Pil (2001), “A theory of mixed bundling applied to the GE-Honeywell merger,” Antitrust, Fall Issue, pp.32-33
Choi, Jay Pil (2003), “Antitrust analysis of mergers with bundling in complementary markets: implications for pricing,
innovation and compatibility choice”, NET Institute Working Paper #03-02, http://www.netinst.org/Choi.pdf
Choi, Jay Pil (2004), “Tying and innovation: A dynamic analysis of tying arrangements”, Economic Journal, Vol.114,
pp.83-101
Choi, Jay Pil and Stefanadis, Christodulos (2000), “Tying, investment, and the dynamic leverage theory”, RAND
Journal of Economics, Vol.32, pp.52-71
Choi, Jay Pil and Yi, Sang-Seung (2000), “Vertical Foreclosure with the choice of input specifications,” RAND Journal
of Economics, Vol.31, pp.717-743
Church, Jeffrey (2004), The impact of vertical and conglomerate mergers on competition, Report for DG Competition,
European Commission, http://europa.eu.int/comm/ competition/mergers/others/#study
Coase, Ronald H. (1937), “The nature of the firm”, Economica, 4, pp.386-405; reprinted in: Stigler, George J. and
Boulding, Kenneth E., eds., Readings in price theory, (Homewood, Illinois: Richard D. Irwin)
Cook, Robert N. (2001), “AOL/Time Warner: The close antitrust call and the changing of the guard at the FTC”,
Antitrust Report, May
Cooper, James C., Froeb, Luke M., O’Brien, Dan and Vita, Michael G.(2005), “Vertical Antitrust Policy as a Problem
of Inference”, International Journal of Industrial Organization, forthcoming
D’Aspremont, Claude and Jacquemin, Alexis (1988): “Cooperative and Noncooperative R&D in Duopoly with
Spillovers”, American Economic Review, Vol.78, pp.1133-1137
De Bondt, Raymond (1996), “Spillovers and innovative activities”, International Journal of Industrial Organization,
Vol.15, pp.1-28
Delbono, Flavio and Denicolo, Vincenzo (1990): "R&D Investment in a Symmetric and Homogeneous Oligopoly",
International Journal of Industrial Organization, Vol.8, pp. 297-313
Denicolo, Vincenzo (2004), “Compatibility and bundling with generalist and specialist firms”, Journal of Industrial
Economics, Vol.48, pp.177-188
Diamond, Douglas (1984), “Financial intermediation and delegated monitoring”, Review of Economic Studies, Vol.51,
pp.393-414
Dobson, Paul, Waterson, Michael and Chu, Alex (1998), “The welfare consequences of the exercise of buyer power”
Office of Fair Trading Research Paper 16, London, http://www.oft.gov.uk/NR/rdonlyres/9A4F0B82-15144344-9C1F-39621912E9DE/0/oft239.pdf
Economides, Nicholas (1993), "Mixed bundling in duopoly," Discussion Paper EC-93-29, Stern School of Business,
N.Y.U., New York
Economides, Nicholas, and Salop, Steven C., (1992), "Competition and integration among complements, and
network market structure," Journal of Industrial Economics, Vol.40, pp.105-123
Emons, Winand (1996), “Good times, bad times and vertical upstream integration”, International Journal of Industrial
Organisation, Vol.14, pp.465-484
European Commission (1997), Green paper on vertical restraints in EC competition policy, COM(96) 721, January,
http://europa.eu.int/comm/off/green/index.htm
European Commission (2000), “Guidelines on vertical restraints”, Official Journal of the European Communities C
291, 13.10.2000, pp.1-44
European Commission (2001), “Guidelines on the applicability of Article 81 to horizontal co-operation agreements”,
Official Journal of the European Communities C 3, 06.01.2001, p.2
European Commission (2004a), “Guidelines on the assessment of horizontal mergers under the Council Regulation
on the control of concentrations between undertakings”, Official Journal of the European Union C 31,
05.02.2004, pp.5-18
European Commission (2004b), “Guidelines on the application of Article 81(3) of the Treaty”, Official Journal of the
European Union C 101, 27.4.2004, pp.97-118
Page 148
Farrell, Joseph and Katz, Michael L (2000), "Innovation, rent extraction, and integration in systems markets", Journal
of Industrial Economics, Vol.48, pp.413-432
Farrell, Joseph and Shapiro, Carl (1990), “Horizontal mergers: An equilibrium analysis”, American Economic Review,
Vol.80, pp.107-126
Farrell, Joseph and Shapiro, Carl (2001), “Scale economies and synergies in horizontal merger analysis”, Antitrust
Law Journal, Vol.68, pp.685-710
Farrell, Joseph and Weiser, Phil (2003), “Modularity, vertical integration, and open access policies: Towards a
convergence of anti-trust and regulation in the internet age”, Harvard Journal of Law and Technology,
Vol.17, pp.86-134
Fried, Ian and Kawamoto, Dawn (2002), “Two palms better than one?” CNET NEWS.COM February 4,
http://news.com.com/2100-1040-828446.html
Gal-Or, Esther (1991), “Vertical restraints with incomplete information”, Journal of Industrial Economics, Vol.39,
pp.302-312
Gal-Or, Esther (1999), “Vertical integration and separation of the sales function as implied by competitive forces”,
International Journal of Industrial Organization, Vol.17, pp.641-662
Gaudet, Gerard and Long, Ngo Van (1996), “Vertical integration, foreclosure, and profits in the presence of double
marginalization”, Journal of Economics and Management Strategy, Vol.5, pp.409-432
Gow, Hamish R. and Swinnen, Johan F.M. (1998), “Up and downstream restructuring foreign direct investment and
hold-up problems in agricultural transition”, European Review of Agricultural Economics, Vol.25, pp.331350
Göx, Robert F. (1998), “Strategic transfer pricing, absorption costing and vertical integration”, Management
Accounting Research, Vol.11, pp.327-348
Green, Jerry (1986), “Vertical integration and the assurance of markets”, in: Stiglitz, Joseph E. and Mathewson,
G.Frank, eds., New developments in the analysis of market structure, (New York; MacMillan). Also
published as Discussion Paper 383, Harvard Institute of Economic Research, October 1974
Grossman, Sanford J. and Hart, Oliver (1986), “The costs and benefits of ownership: A theory of vertical and lateral
integration”, Journal of Political Economy, Vol.94, pp.691-719
Hart, Oliver (1995), Firms, Contracts, and Financial Structure, (Oxford: Oxford University Press)
Hart, Oliver and Holmstrom, Bengt R. (2002), “A theory of firm scope”, MIT Department of Economics Working Paper
No. 02-42, Cambridge, Massachusetts
Hart, Oliver and Moore, John (1990), “Property rights and the nature of the firm,” Journal of Political Economy,
Vol.98, pp.1119-1158
Hart, Oliver and Tirole, Jean (1990), “Vertical integration and market foreclosure”, Brookings Papers on Economic
Activity. Microeconomics 1990, pp.205-76 and pp.285-286
Hausman, Jerry and MacKie-Mason, Jeffrey (1988), “Price discrimination and patent policy”, RAND Journal of
Economics, Vol.19, pp.253-265
Hay, Gordon (1973) “An economic analysis of vertical integration”, Industrial Organization Review, Vol.1, pp.188-198
Hennessy, David A. (1996), “Information asymmetry as a reason for food industry vertical integration”, American
Journal of Agricultural Economics, Vol.78, pp.1034-1043
Hennessy, David A. (1997), “Information asymmetry as a reason for vertical integration”, in: Caswell, Julie A. and
Cotterill, Ronald W. eds., Proceedings of the NE-165 Conference titled Strategy and Policy in the Food
System: Emerging Issues, Washington, D.C., June 20–21, 1996, Food Marketing Policy Center, Univ. of
Connecticut, pp.39–51
Higgins, Richard S. (1999), “Competitive vertical foreclosure”, Managerial and Decision Economics, Vol.20, pp.229237
Holmstrom, Bengt R. and Roberts, John (1998), “Boundaries of the firm revisited,” Journal of Economic Perspectives,
Vol.12, pp.73-94
Hovenkamp, Herbert (2001), “Post-chicago antitrust: a review and critique”, Columbia Business Law Review, Vol.2,
pp.257-332
Hubbard, Robert Glenn (1998), “Capital-market imperfections and investment”, Journal of Economic Literature,
Vol.36, pp.193-225
Hubbard, Robert Glenn and Palia, Darius (1999), “A re-examination of the conglomerate merger wave in the 1960s:
An internal capital markets view”, Journal of Finance, Vol.54, pp.1131-1152
Page 149
Huck, Steffen, Konrad, Kai A. and Müller, Wieland (2004), “Profitable horizontal mergers without cost advantages:
the role of internal organization, information, and market structure”, Economica, Vol.71, pp.575-587
Hughes, John S. and Kao, Jennifer L. (2001), “Vertical integration and proprietary information transmission”, Journal
of Economics and Management Strategy, Vol.10, pp.277-299
Inderst, Roman and Müller, Holger M. (2003), “Internal vs. external financing: An optimal contracting approach”,
Journal of Finance, Vol.58, pp.1033-1062
John, George and Weitz, Barton A. (1988), “Forward integration into distribution: an empirical test of transaction cost
analysis”, Journal of Law, Economics and Organisation, Vol.4, pp.337-356
Jorde, Thomas M. and Teece, David J. (1990), “Innovation and cooperation: Implications for competition and
antitrust”, Journal of Economic Perspectives, Vol.4, pp.75-96
Joskow, Paul L. (1985), “Vertical integration and long-term contracts: The case of coal-burning electric generating
plants”, Journal of Law Economics and Organisation, Vol.1, pp.33-80
Joskow, Paul L. (1988), “Asset specificity and the structure of vertical relationships: empirical evidence”, Journal of
Law, Economics and Organisation, Vol.4, pp.95-177
Kamien, Morton I., Mueller, Eitan and Zang, Israel (1992), “Research joint ventures and R&D cartels”, American
Economic Review, Vol.82, pp.1293-1306
Klass, Michael and Salinger, Michael (1995), “Do new theories of vertical foreclosure sound guidance for consent
agreements in vertical merger cases?”, Antitrust Bulletin, Vol.40, pp.667-698
Klein, Benjamin, Crawford, Robert G. and Alchian, Armen A. (1978), “Vertical integration, appropriable rents and the
competitive contracting process”, Journal of Law and Economics, Vol.21, pp.297-326
Kolasky, William J. (2004), Proceedings of the FTC/DOJ Joint Workshop on Merger Enforcement, Washington DC,
19.02.2004, http://www.usdoj.gov/atr/ public/ workshops/docs/202670.pdf
Krattenmaker, Thomas G. and Salop, Steven C. (1986), “Anticompetitive exclusion: Raising rivals’ costs to achieve
power over price”, Yale Law Journal, Vol.96, pp.209-293
Krickx, Guido A. (1995), “Vertical integration in computer main frame industry: A transaction cost interpretation”,
Journal of Economic Behaviour and Organisation, Vol.26, pp.75-91
Kühn, Kai-Uwe (2001), “Fighting collusion: The role of regulation of communication between firms”, Economic Policy,
Vol.32, pp.169-197
Kumpe, Ted and Bolwijn, Piet T. (1988), “Manufacturing: The new case for vertical integration”, Harvard Business
Review, Vol.66, pp.75-81
Laffont, Jean-Jacques and Tirole, Jean (1999), A theory of incentives in procurement and regulation, (Cambridge,
Massachusetts: MIT Press)
Lazear, Edward P. (1986), "Retail pricing and clearance sales", American Economic Review, Vol.76, pp.14
Levy, David and Haber, Lawrence J. (1986), “An advantage of the multiproduct firm: the transferability of firm-specific
capital”, Journal of Economic Behaviour and Organisation, Vol.9, pp.153-185
Lieberman, Marvin B. (1991), “Determinants of vertical integration; An empirical test”, Journal of Industrial
Economics, Vol.39, pp.445-450
Lipczynski, John and Wilson, John (2001), Industrial organisation: An analysis of competitive markets, (Harlow, New
Jersey: Prentice Hall)
Lyons, Bruce R. (2002), “Could politicians be more right than economists? A theory of merger standards”, Revised
CCR Working Paper, CCR 02-1, University of East Anglia (U.K.). http://www.ccr.uea.ac.uk/workingpapers/
ccr02-1.pdf
Ma, Ching-to A. (1997), “Option contracts and vertical foreclosure”, Journal of Economics and Management Strategy,
Vol.12, pp.725-753
Mallela, Parthasaradhi and Nahata, Babu (1980), “Theory of vertical control with variable proportions”, Journal of
Political Economy, Vol.88, pp.1009-1025
Mason, Charles F. and Philips, Owen R. (2000), “Vertical integration and collusive incentives: An experimental
analysis”, International Journal of Industrial Organization, Vol.18, pp.471-496
Masten, Scott E. (1984), “The organisation of production: Evidence from the aerospace industry”, Journal of Law and
Economics, Vol.21, pp.297-326
Mathewson, Frank and Winter, Ralph A. (1998), “The law and economics of resale price maintenance”, Review of
Industrial Organization, Vol.13, pp.57-84
Page 150
Matutes, Carmen, and Regibeau, Pierre (1988), "Mix and match: Product compatibility without network externalities,"
Rand Journal of Economics, Vol.19, pp. 219-234
McAfee, Randolph P. (1999), “The effects of vertical integration on competing input suppliers”, Federal Reserve Bank
of Cleveland Economic Review, Vol.35, pp.2-8
McAfee, Randolph P. and Schwartz, Marius (1994), “Opportunism in multilateral vertical contracting:
Non-discrimination, exclusivity, and uniformity”, American Economic Review, Vol.84, pp.210-30
McAfee, Randolph P., McMillan, John and Whinston, Michael D, (1989), “Multiproduct monopoly, commodity
bundling, and correlation of values”, The Quarterly Journal of Economics, Vol.104, pp.371-83
McKenzie, Lionel W. (1951), “Ideal output and the interdependence of firms”, Economic Journal, Vol.61, pp.785-803.
Milgrom, Paul and Roberts, John (1992), Economics, organization and management, (Harlow, New Jersey: Prentice
Hall)
Milliou, Chrysovalantou (2004), “Vertical integration and R&D spillovers: Is there a need for ‘firewalls’?”, International
Journal of Industrial Organization, Vol.22, pp.25-43.
Mitchell, Mark and Mulherin, Harold (1996), “The impact of industry shocks on takeover and restructuring activity”,
Journal of Financial Economics, Vol.41, pp.193-229
Mock, Dave (2005), The Qualcomm equation, (New York: Amacom)
Monteverde, Kirk and Teece, David J. (1982), ”Supplier switching costs and vertical integration in the automobile
industry”, Bell Journal of Economics, Vol.13, pp.206-213
Montgomery, Cynthia A. and Hariharan, S. (1991), “Diversified expansion by large established firm”, Journal of
Economic Behaviour and Organisation, Vol.15, pp.71-89
Morse, Howard (1998), “Vertical mergers: recent learning”, Business Lawyer, Vol.53, pp.1217-1248
Nalebuff, Barry (1999), “Bundling”, Yale ICF Working Paper Series No.99-14, New Haven, Connecticut
Nalebuff, Barry (2000), “Competing against bundles”, in: Hammonds, Peter and Myles, Gareth D. eds., Incentives,
organization, and public economics, (Oxford: Oxford University Press)
Nalebuff, Barry (2003), “Bundling, tying and portfolio effects, Part 1: Conceptual issues”, Department of Trade and
Industry Economics Paper 1, Part 1, London, U.K., http://www.dti.gov.uk/ccp/topics2/pdf2/bundle1.pdf
Nalebuff, Barry and Lu, Shihua (2001), “A bundle of trouble - Bundling and the GE-Honeywell merger”, Yale School
of Management Working Paper
Nalebuff, Barry and Majerus, David (2003), “Bundling, tying and portfolio effects, Part 2: Case studies”, Department
of Trade and Industry Economics Paper 1, Part 2, London, U.K., http://www.dti.gov.uk/ccp/topics2/pdf2/
bundle2.pdf
Nocke, Volker, and White, Lucy (2003), “Do vertical mergers facilitate upstream collusion”, Penn Institute for
Economic Research Working Paper 03-33, Philadelphia, Pennsylvania
OECD (2002), “Portfolio effects in conglomerate mergers”, OECD Journal of Competition Law and Policy, Vol. 4, pp.
54-147.
Office of Fair Trading (2000), “The supply of beer”, Report No.317, London, U.K., http://www.oft.gov.uk/NR/rdonlyres
/34960586-5561-468F-ADB7-1A0EB8A6F367/0/oft317.pdf
Pautler, Paul A. (2001), “Evidence on mergers and acquisitions”, Federal Trade Commission, Bureau of Economics
Working Paper No.244, Washington, DC
Perry, M.K. (1989), “Vertical integration: Determinants and effects”, in: Schmalensee, R. and Willig, R.D., eds.,
Handbook of Industrial Organisation, Vol.I, (Amsterdam, North-Holland: Elsevier)
Porter, M. E. (1980), Competitive strategy, (New York: Free Press)
Posner, Richard (1976), Antitrust law, (Chicago: University of Chicago Press)
Prescott Edward C. and Visscher, Michael (1980), “Organization capital”, Journal of Political Economy, Vol.88,
pp.446-461
Quirmbach, Herman C. (1986), “The path of price changes in vertical integration”, Journal of Political Economy,
Vol.94, pp.1110-1119
Ragan, Kasturi and Bell, Jerry A. (1988), “Merck-Medco: Vertical integration in the pharmaceutical industry”, Harvard
Business School, Case 9-598-091
RBB Economics (2002), “A bridge too far? Substitutes, complements and theories of exclusion in EC merger control”,
RBB Brief 1. http://www.rbbecon.com/publications/downloads/rbb_brief01.pdf
Page 151
Reiffen, David and Vita, Michael (1995), “Is there a new thinking on vertical mergers?”, Antitrust Law Journal, Vol.63,
pp.917-942
Rey, Patrick and Stiglitz, Joseph (1995), “The role of exclusive territories in producers' competition”, RAND Journal of
Economics, Vol.26, pp.431-451
Rey, Patrick and Tirole, Jean (1986a), “The logic of vertical restraints”, American Economic Review, Vol.76, pp.921939
Rey, Patrick and Tirole, Jean (1986b), “Vertical restraints from a principal-agent viewpoint”, in: Pellegrini, L. and
Reddy, S.K., eds., Marketing channels: Relationships and performance, (Lexington, Massachusetts:
Lexington Books)
Rey, Patrick and Tirole, Jean (2003), “A primer on foreclosure”, in: Armstrong, Mark and Porter, Robert, eds.,
Handbook of Industrial Organization, Vol. III, (Amsterdam: Elsevier North-Holland), forthcoming
Riordan, Michael (1998), “Anticompetitive vertical integration by a dominant firm”, American Economic Review,
Vol.88, pp.1232-1248
Riordan, Michael (1998), “Anticompetitive vertical integration by a dominant firm”, American Economic Review,
Vol.88, pp.1232-1248
Riordan¸ Michael and Salop, Steven C. (1995), “Evaluating vertical mergers: A post-Chicago approach”, Antitrust
Law Journal, Vol. 63, pp.513-568
Romano, Richard E. (1994), “Double moral hazard and resale price maintenance”, RAND Journal of Economics,
Vol.25, pp.455-466
Salinger, Michael A. (1988), “Vertical mergers and market foreclosure”, Quarterly Journal of Economics, Vol.103,
pp.345-356
Salinger, Michael A. (1989), “The meaning of ‘upstream’ and ‘downstream’ and the implications for modelling vertical
mergers”, Journal of Industrial Economics, Vol.37, pp.373-387
Salinger, Michael A. (1995), “A graphical analysis of bundling”, Journal of Business, Vol. 68, pp.85-98
Salinger, Michael A. (1998) “Vertical mergers and market foreclosure”, Quarterly Journal of Economics, Vol.77,
pp.345-356
Salop, Steven C. (1995), “Vertical mergers and efficiency: A reply to Reiffen and Vita”, Antitrust Law Journal, Vol.63,
pp.943-950
Salop, Steven C. and Scheffman, David T. (1987), “Cost-raising strategies”, Journal of Industrial Economics, Vol.36,
pp.19-34
Scheffman, David T. and Higgins, Richard S. (2004), “20 years of raising rivals’ costs: History, assessment and
future”, George Mason University Law Review, http://www.ftc.gov/be/rrcgmu.pdf
Schmalensee, Richard (1973), “Output and welfare implications of monopolistic third-degree price discrimination”,
American Economic Review, Vol.71, pp.242-247
Schmalensee, Richard L. (1984), “Gaussian demand and commodity bundling”, Journal of Business, Vol. 57,
pp.S211-S230
Seidmann, Daniel J. (1991), "Bundling as a facilitating eevice: A reinterpretation of leverage theory", Economica,
Vol.58, pp.491-499
Shaffer, Greg (1991), "Capturing strategic rent: Full-line forcing, brand discounts, Aggregate rebates, and maximum
resale price maintenance”, Journal of Industrial Economics, Vol. 39, pp.557-575
Slutsky, Allan and Caves, Richard E. (1991), “Synergy agency and the determinants of premia paid in mergers”,
Journal of Industrial Economics, Vol.39, pp.277-296
Spiller, Pablo T. (1985), “On vertical mergers”, Journal of Law, Economics and Organisation, Vol.1, pp.285-312
Stigler, George. J. (1963), “United States v. Loew’s Inc.: A note on block booking”, Supreme Court Review, Vol.152,
pp.152-157
Stole, Lars A. (2003), “Price discrimination and imperfect competition”, in: Armstrong, Mark and Porter, Rob, eds.,
Handbook of Industrial Organization, Vol. III, (Amsterdam: Elsevier North-Holland), forthcoming
Tan, Guofu (2001), “The economic theory of vertical restraints”, The Competition Bureau, Canada,
http://strategis.ic.gc.ca/epic/internet/incb-bc.nsf/en/ct02318e.html#ii
Teece, David J. (1980), “Economies of scope and the scope of the enterprise”, Journal of Economic Behaviour and
Organization, Vol.1, pp.223-47
Teece, David J. (1981), “Internal organisation and economic performance: an empirical analysis of profitability of
principal firms”, Journal of Industrial Economics, Vol.30, pp.173-99
Page 152
Teece, David J. (1982), “Towards an economic theory of the multiproduct firm”, Journal of Economic Behaviour and
Organisation, Vol.3, pp.39-63
Thomas, Charles J. (1996), “Market structure and the flow of information in repeated auctions”, Federal Trade
Commission,
Bureau
of
Economics
Working
Paper
No.213,
Washington,
DC,
http://www.ftc.gov/be/workpapers/wp213.pdf
Tirole, Jean (1988), The theory of industrial organization, (Cambridge, Massachusetts: MIT Press).
Varian, Hal (1989), ‘‘Price discrimination’’, in: Schmalensee, Richard and Willig, Robert D., eds., Handbook of
Industrial Organization, Vol. I, (Amsterdam: Elsevier North-Holland)
Vernon, John M. and Graham Daniel A. (1971), “Profitability of monopolization by vertical integration”, Journal of
Political Economy, Vol.79, pp.924-925
Walker, Gordon and Weber, David (1984), “A transaction cost approach in make-or-buy decision”, Administrative
Science Quarterly, Vol.29, pp.373-391
Warren-Boulton, Frederick R. (1974), “Vertical control with variable proportions”, Journal of Political Economy,
Vol.82, pp.783-802
Warren-Boulton, Frederick R. (2002), “The contribution of the merger guidelines to the analysis of non-horizontal
mergers”, The 20th anniversary of the 1982 merger guidelines to the evolution of antitrust doctrine,
http://www.usdoj.gov/atr/hmerger.htm#papers
Waterman, David and Weiss, Andrew (1996), “The effects of vertical integration between cable television systems
and pay cable networks”, Journal of Econometrics, Vol.72, pp.357-395
Waterson, Michael (1980), “Price-cost margins and successive market power”, Quarterly Journal of Economics,
Vol.94, pp.135-150
Waterson, Michael (1982), “Vertical integration, variable proportions and oligopoly”, Economic Journal, Vol. 92.
pp.129-144
Westfield, Fred M. (1982), “Vertical integration: does product price rise or fall”, American Economic Review, Vol.71,
pp.334-346
Whinston, Michael (1990), “Tying, foreclosure, and exclusion”, American Economic Review, Vol.80, pp.837-859
White, Lawrence J. (1989), “The revolution in Antitrust Analysis of vertical relationships: How did we get from there to
here?”, in: Larner, Robert J. and Meehan, James W., eds., Economics and Antitrust Policy, (New
York, London: Quorum).
Williamson, O.E. (1975), Markets and hierarchies: Analysis and antitrust implications, (New York: Free Press)
Williamson, O.E. (1979), “Transaction cost economics: The governance of contractual relations”, Journal of Law and
Economics, Vol.22, pp.233-261.
Williamson, O.E. (1989), “Transaction Cost Economics”, in Schmalensee, Richard. and Willig, Robert D., eds.,
Handbook of Industrial Organisation, Vol. I, (Amsterdam: Elsevier North-Holland)
Winter, Ralph A. (1993), “Vertical control and price versus non-price competition”, Quarterly Journal of Economics,
Vol.108, pp.61-76
Winter, Ralph A. (1997), ‘‘Colluding on relative prices’’, RAND Journal of Economics, Vol.28, pp.359–371
Page 153
Appendix A: Intensity
Marginalisation
of
Competition
and
Double
This appendix makes use of simple theoretical economic models to prove that the lower the
intensity of competition at the downstream level, the higher the benefits from the elimination of
the double marginalisation problem that can be achieved through vertical integration. Moreover,
we show that in the extreme case of perfect competition in the downstream market, the
competitive effects brought about by the vertical merger are neutral as no double
marginalisation occurs pre merger.
A.1
Model
A.1.1
Assumptions
Our model is based on the following assumptions:
1. In the industry there are n>0 firms operating downstream and one operating upstream;
2. The monopolist produces each unit of output at cost C and it sells it to the downstream
firms at price c. Its demand function is given by the output of the downstream firms;
3. The n downstream firms produce a perfectly homogeneous product;
4. Downstream firms use one unit of input (which is the output of the upstream firm) in
order to produce one unit of output. They also face an additional cost x per unit sold.
n
5. Demand function in the downstream market is
p = a − bQ , where Q = ∑ qi is the
i =1
aggregate output, and qi is firm i’s individual output.
Given assumptions 1-5, each downstream firm’s profit function can be written as
n
⎛
⎞
Π i = ⎜ a − b∑ qi − c − x ⎟qi
i =1
⎝
⎠
(1)
and the upstream firm’s profit function is
Page 155
n
π = (c − C )∑ qi
(2)
i =1
A.1.2
Time structure of the model
As common in the literature on vertical restrictions, the timing in our analysis is as follows.
•
First stage. The upstream monopolist chooses c such that (2) is maximised.
•
Second stage. Downstream firms observe c and compete in quantities à la Cournot.
In the special case n=1, the downstream firm simply chooses the output that
maximises its profit.
Since this is a dynamic game with complete and perfect information, the relevant concept of
equilibrium is the subgame perfect equilibrium. Therefore, the game must be solved backwards
starting from the second stage, in which for a given value of c, the n downstream firms compete
à la Cournot. Then in the first stage, the upstream firm, anticipating the outcome of the next
stage, will choose the optimal value of c that maximises its own profit.
A.2
Solution in the case of non-integration
In the second stage, each downstream firm i maximises its individual profit, given by (1), with
respect to qi . The solution is
qi =
a−c− x
.
b(1 + n )
(3)
Then, in the first stage, the upstream firm maximises its profit, given by (2), with respect to c.
That is,
(c − C )∑ qi = n(c − C ) a − c − x
b(1 + n )
i =1
n
max
c
This leads to
c=
Page 156
a− x+C
2
Industry output is then:
n
Q = ∑ qi = n
i =1
A.3
a − x−C
2b(1 + n )
Solution in the case of vertical integration
Suppose now that the n downstream firms are acquired by the upstream firm. In what follows,
we solve the two-stage model described above in the new scenario and then we compare the
outcomes with the previous model.
As the upstream firm now owns the downstream firms, the only choice that it needs to make is
how much output to produce and sell in the downstream market. In other words, the monopolist
maximises:
max Π = (a − bQ − C − x )Q
Q
Note that we have assumed that the monopolist’s cost is equal to C+x. In other words, we have
assumed that the merger does not give rise to any cost synergies.
The solution to the above maximisation problem is:
Q ** =
A.4
a−C − x
2b
Benefits from vertical integration
The benefit from vertical integration can be defined as the difference between
Q ** and Q .
Indeed, to the extent that vertical integration brings about a social welfare improvement, this will
result in an increase in the level of output (or equivalently a decrease in the level of price). The
difference between the levels of output in the two scenarios is:
Q ** − Q =
a − x −C 1
2b 1 + n
It is easy to prove that the above difference is always positive (i.e.,vertical integration always
leads to a higher output), is increasing in the demand intercept a and decreasing in the cost
Page 157
parameters x and C , the demand parameter b , and most importantly the number of firms n .
The latter result implies that the larger the number of firms, the lower is the benefit from vertical
integration. This is due to the fact that as the number of downstream firms increases,
competition also increases. Therefore the double marginalisation problem tends to disappear.
Interestingly, as the number of firms n tends to infinite, the benefit from vertical integration
disappears.
Figure A.1 below illustrates the behaviour of the relationship between
Q ** − Q and the number
of firms n assuming that ( a − x − C ) 2b = 20 holds.
Figure A.1
Benefits from vertical integration as a function of the number of
downstream firms
12.0
Additional Output from Vertical Integration
10.0
8.0
6.0
4.0
2.0
0.0
1
2
3
4
5
6
7
8
9
10
11
Number of firms
Page 158
12
13
14
15
16
17
18
19
20
Appendix B: Glossary
Adverse Selection: The tendency for a contract to attract parties which are most likely to
benefit from it. For example, people that have higher risk of illness are more likely to take out
health insurance.
Agency Costs: The costs incurred by a party (the principal) for incentivising or monitoring the
performance of a firm (the agent) whose goods or services it has contracted.
Agent: A party whose goods or services have been contracted by another party (the principal).
Arbitrage: Buying a good at one price in one market and selling it in another market at a higher
price (e.g. this may give rise to parellel imports or “grey trade”).
Asymmetric Information: A situation where economic agents have different information.
Bertrand Competition: A model of economic conduct in which firms compete in prices.
Chicago School (on Antitrust): The school of thought developed during the 1970s by
economists such as Bork and Posner which suggests that non-horizontal mergers (and selfinterested actions more generally) do not cause anti-competitive effects.
Closed System: See Tying
Closed Territory Distribution: A vertical restraint whereby a manufacturer commits to supply
only one retailer in each territory.
Commercial Tying: See Pure Bundling.
Complements: Two products are termed complements if a rise in the price of one good induces
a decrease in demand for the other. A simple example of a complementary relationship might
be bread and butter. Complementary goods have a negative cross-price elasticity.
Complete Contract: A contract specifying the parties’ undertakings under all possible
circumstances.
Conglomerate Merger: Concentrations involving companies that act in different markets, which
are not horizontally or vertically related.
Consumer Surplus: The benefit a consumer gains from a good in excess of what is paid for it.
Page 159
Coordinated Effects: An increased incentive for all firms in a market to coordinate on a price
rise after a merger.
Cost Function: A function showing the cost of producing a good, in particular, in relation to
quantity.
Cournot Competition: A model of oligopolistic competition where firms compete in quantities.
Cournot Effect: An inefficiency arising from two or more firms producing complementary goods
priced at a mark-up over the marginal cost of production; see Section 4.2.1.
Customer Foreclosure: An anti-competitive effect which arises if an integrated firm denies its
non-integrated upstream competitors access to the downstream market, such that these rivals
are unable to compete. This strategy is also known as “reducing rivals’ revenues”.
Demand Function: A function showing the demand for a good in relation to its price.
Double Marginalisation: An inefficiency arising from two or more firms on different levels of a
supply chain pricing at a mark-up over the marginal cost of production; see Section 3.2.1.
Economies of Scale: Reductions in production costs arising from an expansion of output of a
good.
Economies of Scope: Reductions in production costs arising from parallel production of
different products by a single firm.
Elasticity of Demand, Cross-Price: The proportionate decrease in demand for good 1 which
occurs with a unit increase in price of good 2.
Elasticity of Demand, Own-Price: The proportionate decrease in demand for a good due to a
unit increase in its price.
Exclusive Dealing: Vertical restraints in which two parties commit only to deal with each other.
Externality: The benefit or adverse effect which one firm’s actions have on another firm. For
example, an increase in advertising by one distributor may benefit all distributors of that good.
Firewall: A provision that confidential information revealed by other parties while dealing with
one division of a vertically integrated firm cannot not be passed on to its other divisions.
Fixed Cost: Costs which do not vary with the level of output.
Page 160
Foreclosure (Vertical Foreclosure): Denying or reducing horizontal competitors’ access to a
vertically related good, such that they are unable to compete. This can take the form of
customer or input foreclosure.
Free rider: A firm which benefits from the investment of another firm, without contributing to the
cost. For example, a distributor may benefit from the advertisements of another distributor of the
same good (horizontal free riding) or from improvements in quality by the manufacturer (vertical
free riding).
Hold-up Problem: A situation in which an investment that benefits two vertically or
complementary related firms is not undertaken because of fear of ex-post opportunistic
behaviour (see Section 3.5.2).
Horizontal Merger: Concentrations involving companies that act at the same level of the supply
chain.
Incomplete Contract: Contracts which do not specify the parties’ undertakings under all
possible future circumstances, usually because not all behaviour is measurable or the number
of outcomes makes such a contract intractable.
Input Foreclosure: An increase in the price of inputs by a vertically integrated firm, such that
downstream rivals are put at a disadvantage and are unable to compete. This strategy is also
described in the literature as “raising rivals’ costs”. Complete input foreclosure is realised when
the integrated firm refuses to supply the good to independent firms.
Input Substitution: An inefficiency that arises due to replacement of (a part) of a marked-up
input and towards an alternative, substitute input produced under more competitive conditions
(and consequently sold at a lower price). See Section 3.2.2.
Inter-brand competition: Competition between products sold under different brands, as
opposed to inter-brand competition (see below).
Intermediate Good: A good that is used as an input in production and not as a final good.
Internalizing Complementary Efficiencies (ICE): The idea, developed by Farrell and Weiser
(2003), that a monopolist will actively promote the most efficient organization for the
complementary or vertically related (i.e., upstream or downstream) industry; see Section 2.3.2.
Intra-brand competition: Competition between different outlets for sales of one brand, as
opposed to inter-brand competition (see above).
Marginal Cost: The additional cost of producing an extra unit of good.
Page 161
Metering: A form of price discrimination whereby a producer can charge consumers according
to intensity of use. For example, a manufacturer of electronic printers can price discriminate
across end-users by charging a lower price for the printer and a higher price for ink cartridges,
so that more intensive users effectively pay a higher prices. The practice of tying
complementary goods is often used to this purpose.
Mixed Bundling: A commercial practice in which two goods are offered to customers as a
package in parallel to being sold individually. The package, or bundle, is sold at a discount
compared to the sum of the prices of individual components.
Moral Hazard: The danger that, having entered into a contract, a party will change its behaviour
to act opportunistically. A classic example of moral hazard are consumers, who, having insured
their goods, have a lower incentive to take care of them than if they were uninsured.
Neighbouring Market: Unrelated goods which use similar distribution channels. For example
banking services and financial services may be neither complements nor substitutes; however,
they may be aimed at the same group of customers, and so utilise similar distribution channels.
Non-Coordinated Effects: See Unilateral Effects
Non-Horizontal Mergers: Vertical or conglomerate mergers.
One-Monopoly Profit: The idea that a firm cannot increase its total market power by extending
it to vertically or complementary markets.
Open System: See Tying
Perfect Price Discrimination: Setting prices such that each consumer is charged the
maximum amount he or she is willing to pay for the good.
Portfolio Power: A firm’s benefit that is due to offering a larger range of products.
Price Discrimination: Charging different prices to different consumers for the same good.
Principal: A party contracting the provision of some service or good to a second party, the
agent.
Process Innovation: An innovation leading to more efficient production of the same good (see
Product Innovation).
Product Innovation: An innovation leading to a new product, or a quality improvement of an
existing product (see Process Innovation).
Page 162
Pure Bundling: The commercial practice where two products are only sold together, but not
individually.
Quantity Forcing: A vertical restraint in which the upstream manufacturer fixes a minimum
quantity of purchases to be made by a downstream retailer.
Raising Rivals’ Costs (RRC): See Input Foreclosure.
Reducing Rivals’ Revenues: See Customer Foreclosure.
Resale Price Maintenance (RPM): A vertical restraint in which the upstream manufacturer
limits the ability of downstream firms (e.g. retailers) to choose the price for downstream sales. It
can take the form of a minimum RPM (i.e., a price floor), a maximum RPM (i.e., a price ceiling),
or a specific price.
Residual rights: All the rights of control over the parties’ assets not specifically given away in a
contract (see Grossman and Hart, 1986).
Specific rights: Rights of control which are specified in a contract (see Grossman and Hart,
1986).
Substitutes: Two products are termed substitutes if a rise in the price of one good induces an
increase in demand for the other. For example, if the price of butter rises, consumers may
consider buying alternative spreads, such as margarine. The demand for these alternative
products will consequently rise. Substitute goods have positive cross-price elasticity.
Sunk Costs: Costs which cannot be recovered if a firm ceases operation.
Technical Bundling: See Tying.
Technological Synergies: Production costs savings arising because the technical conditions
required to produce two goods are similar.
Tied Market: A market into which market power could potentially be leveraged through tying or
bundling of the good with a complementary product that is sold in a market in which the
producer already has market power.
Transaction Cost: The cost of market transactions with other firms; see Section 3.5.
Two-Part Tariff: A pricing system in which customers pay a fixed fee and an additional unit
price for their purchases.
Page 163
Tying Market: A market in which a firm has market power which it could potentially leverage
into a complementary market through tying or bundling.
Tying: This term denotes the practice of making two products technically incompatible with
those of other producers. Here, the notion of “incompatibility” implies that the system good
obtained by combining incompatible components does not function, or works less effectively
than a system made of compatible products. Tied goods are also described as forming a closed
system, as opposed to open systems where both components can be freely chosen on the
market between competing brands, and then matched together.
Unilateral Effects: An increased incentive for a firm to unilaterally raise prices, post merger,
regardless of the actions of other firms.
Unrelated Goods: Products that are neither complements nor substitutes.
Vertical Externality: The benefit a firm obtains when a competitor in vertical (or
complementary) related market lowers its price or increases the quality of the good; see Section
2.2.2.
Vertical Merger: Concentrations involving companies that act at different levels of the supply
chain. Common examples are mergers between a wholesaler and its retailer, or a manufacturer
and its input supplier.
Page 164