-Should a merger control system take into account efficiency gains from horizontal mergers, and balance these gains against the anti-competitive effects of mergers?
If so, how should a system be designed to account for efficiency gains?
There are several reasons why efficiency gains from horizontal mergers are an
important issue today. Business conditions are changing rapidly, for example as a result
of the internal market, increased global competition, and the deregulation of many
industries.
The consequent need to adapt the industry structure has generated a wave
of mergers in Europe as well as in the rest of the world. The current wave is of
historical proportions.
All mergers with a so-called Community dimension must be notified to the
Commission and are subsequently reviewed under the Merger Regulation art.1
to Articles 2(3) of the Regulation, a concentration which “creates or strengthens a
dominant position as a result of which effective competition would be significantly
impeded” shall be prohibited.
According to Article 2(1)(b), the Commission shall, in making this appraisal take into account “the development of technical and economic progress provided that it is to
consumers’ advantage and does not form an obstacle to competition”.
so… does the Merger Regulation allow for an efficiency defence?
Can important cost savings (or other efficiencies) save an anticompetitive merger?
The Commission has, in policy statements, argued that, “(t)here is
no real legal possibility of justifying an efficiency defence under the Merger Regulation.
(Commission, 1996)”
Another reason why it is important to discuss an efficiency defence is the concept of joint dominance. According to the Merger Regulation, a merger can only be blocked if it creates or strengthens a dominant position.
Economic theory suggests that cost savings (and other efficiencies) makes the merger anti-competitive, the lower is concentration. Hence, the introduction of joint dominance makes it more natural to consider efficiencies today.
new computer based simulation techniques can be used as a
way to estimate the likely anti-competitive effects of a merger and, at the same time,
balance these effects against possible efficiency gains. Such techniques are starting to
be used in the USA and in Canada.
should E.U. merger control allow an efficiency defence, and if so, how it should be designed,?
THEORETICAL ANALYSIS
The effects from mergers on consumer, or total surplus, depend on the type of efficiency. We distinguish five categories, based on the concept of the production function:
(1) rationalisation of production, which refers to cost savings from reallocating
production across firms, without increasing the joint technological capabilities;
(2) economies of scale, i.e. savings in average costs associated with an increase in
total output;
(3) technological progress, which may stem from the diffusion of know-how or
increased incentives for R&D;
(4) purchasing economies or savings in factor prices such as intermediate goods or
the cost of capital;
(5) reduction of slack (managerial and X-efficiency).
Next, we review the anti-competitive effects that may arise from mergers, either
due to an increased unilateral market power, or due to an increased likelihood of
successful collusion.
Finally, we discuss both the price effects and total surplus effects arising from mergers, taking into account the role played by efficiencies.
internal efficiencies often need to imply sufficiently large savings in marginal costs for
price to decrease. Not all types of efficiencies guarantee that this will be the case. The
required amount of efficiencies for price to decrease depends on various variables,
such as the merging firms’ market share and the price elasticity of demand. For total
welfare to increase, all types of internal efficiencies may in principle be considered,
although the required amount may depend on which type of efficiency is realised due
to the merger. It is important to stress that the precise amount of efficiencies required
for price to decrease or total welfare to increase is dependent on the specific
assumptions one makes about how firms behave in the market, both before and after
the merger. This does not make it possible to present a simple and unique formula,
applicable to all mergers.
Typologies of efficiency gains
Efficiencies from mergers may come in a variety of ways. In order to obtain a clear and
systematic understanding of the consumer and welfare effects of mergers, discussed in
section 1.2, it is important to make a typology of the various kinds of efficiencies that
may be created. Naturally there are many different way in which one may categorise
efficiency gains from mergers. Different typologies are useful for the different
discussion in this report.
The first typology is based on production function.
· Rationalisation,
· Economies of scale,
· Technological progress,
· Purchasing economies,
· Slack.
This typology is useful for the identification of different types of efficiencies.
A second distinction that is often made in the context of merger analysis is:
· Real cost-savings,
· Redistributive (or pecuniary) cost-savings.
This distinction is important since typically only real cost-savings are considered
in an efficiency defence. Redistributive gains are cost-savings that the firms may
achieve for example in the form of lower taxes. However purchasing economies may
also be redistributive (see below). Real cost-savings are those savings that correspond
to some savings of productive resources in the economy. Rationalisation, economies of
scale, technological progress, and slack reduction are all real cost-savings. Also some
purchasing economies are real cost-savings.
A third distinction that is often made in the context of merger analysis is:
· Fixed costs,
· Variable costs.
This distinction is important since savings in variable costs, but not savings in
fixed costs, may almost immediately benefit not only the merging firms, but also the
consumers (see below). Savings in fixed costs normally come in the form of economies
of scale, technological progress, and purchasing economies. Savings in variable costs
may come in all five forms.
A fourth distinction that is useful in the context of mergers is:
· Firm level efficiencies,
· Industry level efficiencies.
An example of efficiencies at the level of the industry is cost-savings due to a
reallocation of production from the merging firms to their competitors—a common
effect of mergers (see below). This distinction is important since it is mainly about the
first category that the merging firms have an informational advantage over competition
authorities. All the five categories above can occur at both the level of the firm, but
also at the level of the industry. Efficiencies at the level of the industry are discussed
separately in section 2.3, since they should be considered in a different way by
competition authorities.
Finally, one may distinguish between:
· Efficiencies in the relevant market,
· Efficiencies in other markets.
According to some, only the first category can be taken into account in an
efficiency defence. At a minimum, including also the second type introduces some
additional complications.
Rationalisation
Rationalisation of production refers to the cost savings that may be realised from
shifting output from one plant to another, without changing the firms’ joint production
possibilities. As the term indicates, rationalisation of production refers to an optimal
allocation of the production levels across the different plants of a firm.
Before the merger, the firms may differ in their marginal cost of production. This
may be the case for at least three reasons.
First, one of the firms may have a higher amount of physical capital.
Second, one firm may have some inherent competitive advantage, for example due to a patent or other superior knowledge.
Finally, when marginal cost is increasing in output due to capacity constraints, firms may differ in their marginal costs because they are producing at different output levels.
After the merger, the new company becomes a multi-plant firm, and cost savings
can be realised by shifting production from the plants with a high marginal cost to the
plants with a lower marginal cost. A firm fully rationalises its production if the
marginal costs at all its plants are equalised: in this case it no longer pays to further
reallocate production across plants. For example, when cost differences arise from
differing capacity constraints, output rationalisation implies that the firm, who is most
capacity-constrained, reduces its production in favour of the firm with more excess
capacity.
The most drastic case of rationalisation occurs when one firm operates at such a
low marginal cost (for all relevant levels of total production) that it is optimal to
reallocate all production to that firm. The merger then effectively involves the
shutdown of the other, less efficient firm. In this case, the merger also leads to an
elimination of the fixed set-up costs that are required to keep a plant operational.
Such fixed cost savings are discussed in the next sub-section on economies of scale.
Economies of scale (and scope)
A firm is said to have economies of scale when its average cost falls as output
increases. Economies of scope generalise the concept of economies of scale to the case
of the multiproduct firm. Economies of scale and scope are frequently
used as an argument to defend a proposed merger. To assess the validity of the
But in some cases plant closure may also lead to new irreversible costs argument in each case, it is important to understand the sources of economies of scale, and assess whether they cannot be realised otherwise.
Economies of scale, realised through a merger, may be the result of coordination of the (formerly separate) firms’ investments in physical capital – called long-run economies of scale. Other realisations of economies of scale may, however, come already in the short run (when physical capital is held fixed).
Short-run economies of scale
There are two types of short-run economies of scale that potentially can be realised
through a merger, namely the elimination of duplication of indivisible tasks, and a form
of rationalisation.
First, indivisibility occurs when it is technically impossible to scale down an input
below a certain minimum size, even when the level of output is very small. No matter
how small the size of a firm, some minimum expenditure on essential tasks is required
to keep the firm operational. These include certain administrative and support routines, such as the purchasing of materials, the billing of customers, personnel service etc.
These tasks involve fixed costs, i.e. costs that do not increase as total output increases.
Before the merger, all firms wastefully duplicate the fixed costs. After the merger, they
may be spread over the larger combined output of the merging firms. The merger then
realises scale economies by avoiding a duplication of fixed costs. Note that a spreading
of these fixed costs may be feasible both when the firms produce identical and when
they produce different products.
Second, short run economies of scale may be realised by a reallocation of
production between plants (rationalisation). We categorise such cost-savings as
economies of scale rather than as rationalisation, if the reason for the reallocation is
that short-run marginal cost are decreasing with higher output.
Such cost-savings could also be classified as rationalisation. We choose to classify them as
economies of scale since they reflect downward sloping marginal costs, which is in line with
the ordinary use of the term economies of scale. Moreover, the primary use of the term
rationalisation can be found in Farrell and Shapiro (1990). Their stability requirement in effect
assumes away the cost-savings that we discuss here, if they are sufficiently large.
Long-run economies of scale
Long-run economies of scale occur when a doubling of all inputs (including physical
capital) leads to more than a doubling of total output. Product-level (or specific)
returns to scale are related to the total production of a single product variety. Plantlevel returns to scale are related to the total production of all product varieties within a
plant. Firm-level returns to scale are related to economies realised by managing many
plants within the same firm.
Long-run economies of scale may arise for several reasons. First, when the
output of a firm is small, it is usually preferable to invest little, and operate at an
inferior technology with a higher marginal cost. As the production of the firm
increases, it becomes worthwhile to invest more in automated technologies that yield
lower marginal costs. Long-run economies of scale may also arise because of the
benefits from specialisation. Each worker can concentrate his or her efforts on certain
specific tasks that can be implemented more efficiently. Similarly, the energy
requirements for a large machine may be proportionally lower than those of a small
machine. Due to certain physical laws, material requirements may also be decreasing in
size.
To realise long-run economies of scale through a merger, it is essential that the
assets of the partners are combined and integrated. Such a restructuring may not be
desirable, in the short run: the plants are already built and it is costly to unbuild them,
reallocate capital and achieve the economies of scale. Therefore, in the short run,
adjustment costs may impede a full integration of the firms’ activities. In contrast, in
the long run, it may be less costly to integrate the future investment decisions within
the newly created firm. Future investments occur for two reasons. First, the firms’
current capital depreciates and old plants need renovation. This includes both physical
capital and intangible assets such as brand name. Second, new investment opportunities
may arise if the size of the market increases.
Economies of scale (Scherer et al., 1975; Panzar, 1989) also apply to a context
where the merging firms produce differentiated products. After a merger between firms
selling similar product lines, it may be desirable to concentrate the production of the
certain products within the same plant, rather than to have each former firm continue
to produce the whole product line within the same plant. Such a specialisation of
production allows the plants to reduce down time due to shifting production (runlength economies).
Long-run economies of scale may arise in both production and in research and
development. They may also arise in marketing activities. A single brand name may be
created to economise on advertising expenditures. The sales forces or the distribution
network may be combined (see, for example, Kitching, 1967).
The exploitation of economies of scale through a merger between firms that are
producing differentiated products may involve a reduction in product diversity. This
potential loss to consumers should also be taken into account when measuring the net
benefits from economies of scale
Economies of scope
Economies of scope arise when it is advantageous to produce goods that are related in
one way or another within the same plant. Economies of scope may for example arise
when multi-product production requires a common “public” input. For example, the
production of wool and mutton requires the common input sheep; the production of
beef and hides requires the common input cows. When economies of scope are
present, it may not be desirable to have plants specialise in the production of single
products even if there are product-specific economies of scale.
Technological progress
In the analysis of technological progress a distinction is often made between process
and product innovations. A process innovation reduces the cost of producing an
existing product; a product innovation increases the value (quality) of an existing
product.
Both types of innovation are essentially equivalent in that they imply an
improvement in the firms’ joint production possibilities frontier. In the discussion
below, we often refer to technological progress as process innovations with
An often cited example is the “cube-square rule”: a doubling in the surface area of a pipe, will
increase the volume by a factor of 4.
Tirole (1988, p. 389) uses the same definition for a process innovation. He defines a product
innovation somewhat differently as the creation of an entirely new product. He then remarks
that product innovation can be viewed as a special case of process innovation: the new product already existed prior to the innovation, and the process innovation reduces the cost of
production should that it can be profitably supplied.
corresponding cost savings. However, it should be clear that similar conclusions can be
drawn for product innovations and corresponding quality improvements.
Diffusion of know-how
Firms may have different technological or administrative capabilities due to different
patents, different experiences, a different management or organisation, etc … A merger
between firms with different characteristics may then lead to a diffusion of know-how
across the participants. This can bring the firms closer to their joint production
possibilities frontier, without shifting the frontier itself.
First, it may be the case that one of the merging firms has superior know-how in
all dimensions. The merger then allows its partner(s) to learn and potentially adopt all
skills of the firms with superior know-how. For example, the better management may
teach the worse management….Alternatively, the better management may replace the old management and make all decisions by itself. In these examples, the diffusion of know-how goes in one single direction, and the superior firm does not learn.
Second, there may be two-way diffusion of know-how. In this case, both firms
can benefit from the merger and improve their technological or administrative
capabilities. A two-way diffusion of know-how is possible when firms have
complementary skills or own some other complementary assets. For example, the
merging firms may own complementary patents, which taken together further improve
the production process on the quality of the product.
A merger in this instance effectively implements a cross-licensing agreement. Similarly, the management of each firm may have built up different experience or expertise. As another example, consider a relatively young R&D-intensive firm that has developed a superior product but lacks the marketing know-how or a distribution network. A merger with one of the existing competitors, with an established brand name and a solid distribution network, may then ensure a fast diffusion of the new (or improved) product.
Two-way diffusion of know-how may also occur when firms have different
capabilities in the production of intermediate outputs. For example, consider two car
manufacturers which both produce (or purchase from suppliers) several intermediate
outputs such as brakes or gears. One of the firms may be more efficient in producing
brakes (or can buy them cheaper from a supplier), whereas the other firms may be
more efficient in producing (or purchasing) gears.
Specialisation then leads to a reduction in costs, below either of the firms’ costs before the merger. As a final example of two-way diffusion of know-how, consider an industry in
which there is learning by doing. Learning by doing means that the firms’ average costs
are declining in their cumulative (past) output (as a measure of experience). It is
sometimes referred to as “dynamic economies of scale”. A merger may facilitate
learning by doing spill-overs, allowing firms to better learn from each other’s
experience.
Incentives for research and development
An important activity of many firms involved in mergers is research and development
(R&D), in either cost-reducing production processes or in product improvement or in
the development of new products. As discussed above, the integration of investment
and R&D activities after the merger may sometimes create significant economies of
scale. In addition, a merger may alter the incentives for R&D expenditures.
R&D decisions are frequently taken strategically, i.e. dependent on the actions of competing
firms. It is often argued that the presence of too much competition destroys the
incentives to spend money on R&D. The most popular argument is that the outcome of
R&D is highly non-proprietary, due to imitation or information spill-overs of the R&D
results (d’Aspremont and Jacquemin, 1989). If this is indeed important, a merger could
help to internalise the benefits from R&D among the participating firms, thereby
creating an increased incentive for R&D.
Even in the absence of R&D spill-overs the intensity of competition influences
incentives for R&D. The industrial economics literature has addressed the question
whether a large dominating firm still has sufficient incentives to spend money on R&D,
compared to its smaller rivals. The answer depends on the expected payoffs from R&D
(assuming no spill-overs). Consider first the case in which the outcome of R&D
involves little risk, so that R&D looks much like traditional investment. In this case, a
ie,the inferior firm’s production frontier has expanded. we classify this as technological progress and not as rationalisation. First, the cost-savings correspond to an expansion of the firms’ joint production possibility frontier. Second, the cost dominating firm would have greater incentives to invest in R&D, so as to retain its competitive advantage and associated monopoly rents. In contrast, if R&D are very risky, then the dominating firm cannot guarantee success even for disproportionately large amounts spent on R&D. As a result, the large firm would rather “rest on its laurels”, enjoy the current monopoly rents and accept the risk of being leapfrogged.
Purchasing economies
Cost savings may also be involved in the merger because of the presence of imperfectly
competitive factor markets. Small firms often need to purchase their inputs such as
materials and energy at prices above marginal costs. When the firms merge, their
bargaining power may increase and more pressure can be put on upstream input
suppliers to cut their prices and obtain quantity discounts. In the automobile industry,
for example, manufacturers sometimes form alliances to purchase their components in
larger amounts to increase their discounts. Similarly, significant advertising discounts
may be obtained when large contracts can be made.
To assess the social effects from increased bargaining power towards suppliers, it is important to know the degree of power at the supplier side. If there is little power on the supplier side, the increased bargaining power of the merging firm may be socially harmful. If, however, the
increased bargaining power forms a form of countervailing power to an already strong
supply side, then the private benefits from the merging firm may coincide with the
social benefits. Unfortunately, economic theory has not yet analysed these issues in
detail.
Note that mergers may also create discounts when there is no increase in
bargaining power (i.e. when the upstream supplier would retain all the bargaining
power to set prices). This happens when suppliers offer two-part tariffs, consisting of a
fixed fee, and a price per unit (or some other non-linear pricing schemes). Such tariffs
are used to price-discriminate between low and high users. When firms merge, they
saving is achieved through reallocation of production of intermediary as opposed to final
products. The first effect has been called the “efficiency effect” in the literature, referring to the fact that a monopoly enjoys higher profits than do two duopolists jointly (see e.g. Gilbert and Newberry, 1982, 1984). The second effect has been called the “replacement effect” (Arrow, 1962). Reinganum (1983) and Tirole (1988) analyse the role of uncertainty in the R&D outcome to
assess the relative importance of both effects.
become high users, and thereby can spread the fixed fee and obtain a lower average
price for their supplies. A merger may also lead to a lower cost of capital since capital markets do not function perfectly. For a variety of reasons, such as asymmetric information about risk
and expected return, firms cannot always borrow at a competitive interest rate.
Especially small and expanding firms often face stringent liquidity constraints, while
large corporations usually have better access to the outside capital markets, and big
firms in declining industries may even generate “excess” liquidity. A small firm that
joins a large corporation, or is being bought by a firm with limited possibilities for
internal expansion obtains new possibilities in raising capital.
Slack
Large public corporations are characterised by a separation of ownership and control.
This creates problems of asymmetric information between the shareholders of the firm
and the management to whom control is delegated. A failure by the management to
maximise the profits of the firm leads to internal inefficiency, sometimes called Xinefficiency.
Why may the goals of management diverge from profit maximisation,
despite the fact that they are often given profit maximising incentives through profit
sharing, stock option plans, etc…? Other interests of management may be the personal
ambitions to obtain power, become the leader of a big or growing company; or not to
change a chosen strategy and thereby admitting old mistakes; or to avoid to fire excess
personnel. These goals conflict with the shareholders’ goals.
The shareholders can exercise some control over the management through the board of directors. However, the management is usually much better informed about its projects, and the board may have only limited possibility to challenge the management’s decisions. To collect all the necessary information is not necessarily profitable. After all, the whole idea to delegate power to the management is to avoid these information costs.
The firm’s internal efficiency is partly determined by the management
(administrative) techniques, such as incentives systems, used to mitigate the problems
caused by the separation of ownership and control. We view such techniques as
For preliminary discussions of the role of countervailing power, see Galbraith (1952), von
Ungern-Sternberg (1995), and Snyder (1996).
defining the firm’s production possibility frontier, and hence improvements in such
techniques as technological progress. A merger may lead to improvements in such
techniques (for example if one firm can learn them from the other). Such gains, we
classify as technological progress, and not slack reduction.
However, the firm’s internal efficiency is also determined by other factors, and
these other factors may be affected as a result of a merger. One commonly mentioned
example is that mergers are an important part of the disciplining power of the capital
market (discussed in sub-section 1.1.5.1).
Another example, pointing in the opposite direction, is that a horizontal merger reduces competition in the product market and hence the disciplining power of the product market on firm efficiency (discussed in sub-section 1.1.5.2).
A third example is that a merger may increase the possibilities for the merged entity to use relative performance evaluations between the formerly separately owned plants.11 If that would happen, slack would be reduced as a result of the merger.
The market for corporate control
A slack makes a firm undervalued and lowers the firm’s stock price. This may induce
another company to buy the firm, re-organise and bring the firm back to profit
maximising behaviour. Actually, Manne (1965) and Marris (1964) argue that the mere
threat of a take-over can be sufficient to discipline the current management, despite the
presence of asymmetric information between shareholders and management.
The threat of corporate take-overs may then serve as a disciplining device to the
management. Yet management is usually not punished after a take-over; they often
even obtain large compensations (“golden parachutes”).
The punishment mainly lies in the loss of the enjoyed managerial rents, including prestige or on-the-job consumption, such as private jets and excessive representation. See Scharfstein (1988) for a detailed analysis. Provided there is a sufficiently high punishment threat to management involved, a well-functioning market for corporate control could then ensure that
managerial inefficiencies could not be long lived.
The term X-efficiency, introduced by Leibenstein (1966), was originally used in a broader
sense than just to describe issues related to the separation of ownership and control.
11 For a discussion of relative performance evaluations, see Holmström (1982).
The disciplinary power of take-overs is, however, limited for several reasons.
First, there is a free-rider problem involved in disciplining the management, as pointed
out by Grossman and Hart (1980). A raider needs to collect costly information to
identify managerial inefficiencies. The raider can then make a profit only if the tender
price of the shares (at which he buys) is lower than the post-raid price.
However, each single current shareholder may not be willing to sell at a tender price, in anticipation of a higher stock price after the raid. One solution to this free rider problem is dilution, which allows a majority shareholder (the raider) to sell part of the firm to another
company he owns at terms that are disadvantageous to minority shareholders. Shleifer
and Vishny (1986) point out that a raid may be successful if it is done by a large
existing shareholder, who than at least enjoys an increase in the value of its own shares
(the other shareholders would then benefit more, of course).
A second limit to take-overs arises because of possible actions by the existing
management in response to take-over bids. For example, poison pills are sometimes
used as a defence. These are preferred stock rights that may be used in the event of a
tender. Scherer (1980, cited in Holmström and Tirole, 1989) also questioned the
strength of the disciplining power of take-overs. Take-overs are very costly so that
they may be used only in cases of severe mismanagement.
The theory of the disciplining role of take-over threats is also problematic in that
it is difficult to test empirically. Studies of actual take-overs cannot provide sufficient
information, since the theory is about the disciplining threat, which is difficult to
measure as long as it is not realised.
As discussed by Manne (1965), the theory of the disciplining role of take-over
threats is potentially important for the design of merger control. Unfortunately, to our
knowledge, there has been little or no research on this topic. Nevertheless, a few
general remarks can be made. By making take-overs more difficult, competition policy
may reduce the disciplining role of the take-over threat. Presumably, the firm’s closest
competitors form the most effective take-over threat, since they are likely to possess
the best information about mismanagement. Without the allowance of an efficiency
defence, such disciplinary mergers may not be carried out. As a compromise, one may
argue that a take-over should be temporarily allowed, to replace current management;
However, see Bagnoli and Lipman (1988), for a critique to the Grossman and Hart argument.
after the re-organisation the target should be sold again as soon as possible. The
problem with such as compromise is however that a raider would be strengthening the
position of its own competitors, so that the raid would not be carried out in the first
place. Hence, if a disciplinary threat of mergers mainly comes from competing firms,
there may be a real trade-off between efficiency gains and anti-competitive effects.
Product market competition and the internal efficiency of firms
When product market competition is soft, management and employees exert low effort
and production costs are high. Moreover, the low efforts and the high costs are too
low and too high respectively, from a social welfare point of view. These ideas are
widespread among economists as well as policy makers. Actually, these ideas motivate
policies to promote competition such as deregulation and trade liberalisation.
For example, the European Commission (1988) argued that “…the new competitive
pressures brought about by the completion of the internal market can be expected to
…produce appreciable gains in internal efficiency…[which will] constitute much of
what can be called the dynamic effects of the internal market…” According to Scherer
and Ross (1990) the empirical evidence is fragmentary but points in the same general
direction: x-efficiency is more apt to be low when competitive pressures are strong
than when firms enjoy insulated market positions.
Moreover, these X-inefficiencies are at least as large as the welfare losses from resource misallocation. There is a small literature that has analysed the different linkages between
product market competition and firm efficiency. Some studies focus on financial
linkages between competition and firm efficiency. Grossman and Hart (1982) argue
that if there exists a bankruptcy-risk, and a receiver who is able to recover all funds
that are not invested, leaving the manager with no perquisites in case of bankruptcy,
the manager invests more to reduce the risk of bankruptcy.
Any changes in product market competition that affect the risk of bankruptcy also affect managerial incentives. Schmidt (1997) makes this intuition more precise. In his model, increased competition has two effects on managerial incentives: it increases the probability of liquidation, which increases managerial effort, but it also reduces the firm’s profits, which may
make it less attractive to induce high effort. Stennek (forthcoming) argues that limited
liability may serve as a disciplining device on the internal efficiency of a firm, and the
tougher the product market competition, the higher the disciplining power.13 However,
even if policies that promote competition enhance X-efficiency, the social gain may be
outweighed by a less efficient allocation of risk.
Other studies focus on informational linkages between competition and firm efficiency. Holmström (1982) and Nalebuff and Stiglitz (1983) argue that cost shocks normally should be positively correlated between competitors. Then, by using relative performance evaluations, the moral hazard problem is mitigated within each firm. However, even if more firms increases the amount of available information and hence increases the overall efficiency of the firms,
the effect on effort is ambiguous.14 Hermalin (1992) focus on wealth linkages between
competition and firm efficiency.
He argues that managers often have substantial bargaining power in negotiations over their employment contracts. Since increased competition decreases the “pie” over which the principal and agent bargain, increased competition also reduces the manager’s wealth. If “shirking” is a normal good, then competition reduces shirking. Finally, some papers focus on output and strategic linkages between competition and firm efficiency. Martin (1993) and Horn, Lang, and Lundgren (1994, 1995) argued that a reduction in marginal cost saves more money for larger firms. Since the size of firms decreases with competition, managers exert more
effort into cost-reduction under soft competition – an output effect. Horn, Lang, and
Lundgren (1994) also demonstrate a strategic effect when the compensation scheme is
public information.
Perhaps surprisingly, the emerging picture from these studies is that the effect of
competition on internal efficiency may be both positive and negative. Moreover, even
in the cases the effect is positive, the welfare consequences may be negative. The
picture is rather complex. Unfortunately, none of the studies explicitly examines the
effect of merger on X-efficiency. The reason for the change in competition in these
studies is rather lowered entry barriers or trade liberalisation. Hence, it is not clear to
what extent the studied linkages between competition and X-efficiency also are if competition is tough, firm revenues are low. Hence, in case costs are firms are forced to pay
low wages (including fringe benefits, courses that increase human capital, and so on). But then, to guarantee that employees receive their expected utility, the firms must pay high wages in case costs are low. Hence, in the presence of limited liability, if competition is tough, the market forces firms to give their employees an incentive contract (wages contingent on cost realisation).
In Hart (1983) and Scharfstein (1988) the information is transmitted via the market price.
Also Brander and Spencer (1989) establishes the existence of a negative relation between
competition and managerial incentives.
whether and how mergers, troughmits effect on competition, affect the X-efficiency of firms is still an open question.
Anti-competitive effects from merger
The prime reason why competition policy authorities are concerned with horizontal
mergers is that they reduce competition, which may have many unwanted
repercussions in the affected markets. The most well-known effect of a reduction of
competition is that prices may increase. This effect will be discussed in detail below.
Another effect is that a reduction of competition may lead to X-inefficiency (or slack)
in the firms (see sub-section 1.1.5 above). Reduced competition may also reduce firms’
incentives to provide product diversity and to innovate (see sub-section 1.1.3.2 above).
Finally we should mention an aspect which is often mentioned but not yet studied in
any detail. Firms’ managers have only limited cognition (or bounded rationality). As an
unavoidable consequence, many of their decisions are based on their beliefs and
prejudices, and not on facts and reasoning.
As a result of these individual imperfections, competition in the market place has an important role to fill as a selection device. If competition is intense, only those firms survive and expand that are run by managers who (by coincidence) happen to be equipped with the most accurate
beliefs and prejudices. These are the firms that produce the product varieties that
consumers want at a low cost. In contrast, if competition is soft, also high-cost firms
that produce inferior product varieties may survive.
As already argued, the most well-known effect of a reduction of competition is
that prices may increase. There are two possible reasons to be concerned with price
increases. First, there is the obvious fact that a price increase implies a transfer of
wealth from consumers to producers. This is a distributional consideration.
Second, an increase in the price of a product above its marginal cost creates (or strengthens) an allocative inefficiency, also called the dead-weight loss. Indeed, whenever the price of
a product exceeds its marginal cost, it would be socially desirable to increase
production. The social value of increasing production by one unit equals the current
price minus the marginal cost (the difference between what consumers are willing to
pay for an additional unit, and what this additional unit costs to producers).
Competition authorities in most countries have been primarily concerned with the
distributional effects of price increases. Economists, in contrast, tend to argue that one
should focus on total welfare, and therefore focus on the allocative inefficiency (or
dead-weight loss) caused by the merger. Whatever the policy concern behind price
increases, a proper assessment of the competitive effects from mergers requires a good
understanding of the nature of competitive interaction in the industry.
Generally speaking, oligopoly theory confirms the common intuition that prices
increase as the number of firms is reduced.
Abstracting from cost reductions, the fear for price increases following merger may thus be justified. Two reasons for price increases may be distinguished.
First, a merger between two or more firms may increase the firms’ unilateral market power. Before the merger, the firms compete and do not take into account the effect of their quantity or price decisions on the profits of their competitors. After the merger, the firms maximise their joint profits, and thereby take into account the detrimental effect of quantity increases or price cuts on the market share of each others’ products.
Second, a merger may shift the nature of conduct from competitive to collusive
behaviour, or facilitate collusion at a higher price level. As the number of firms
decreases, it may become easier to sustain implicit cartel agreements, for example
because it becomes easier to monitor cheating. When a shift in conduct takes place, the
merger increases the joint market power of the firms in the industry.
The risk for price increases following mergers may be limited for several reasons.
The presence of actual competitors producing similar products is an obvious first
constraint.
Second, the possibility of entry in the long run may effectively constrain the
firms’ willingness to raise prices. Third, especially in intermediate goods markets,
strong buyers may exercise countervailing bargaining power that may limit the merging
firms’ potential to raise price. Finally, it may be the case that one of the merging firms
is failing*, perhaps due to a drop in demand for its product. In the absence of a merger
*This relates to the failing firm defence… In the EU the additional condition [to the company would be closing down] is that the market share of the failing firm would inevitably accrue to the acquiring firm even without the merger.
such as firm would eventually have to leave the market, so that market power would
increase irrespective of the merger taking place.
In the following discussion, we will assume that entry is difficult, and that none of the
firms is failing. Efficiency considerations are usually only made when these
assumptions hold true. We will also assume that buyers take prices as given, and hence
do not exercise countervailing power. This assumption is an unfortunate consequence
of the fact that economic theory is lagging behind. The theory of oligopoly that is used
to assess the competitive effects of mergers has generally assumed that one side of the
market (typically the consumer side) is price taking, and that only the other side of the
market exercises market power.
Welfare Effects of Mergers and the Role of Efficiencies
the previous section sketched some general considerations on the competitive effects
of mergers. This section goes into more detail, and in particular examines the role of
efficiencies. In order to assess the overall effects of mergers it is necessary to first
make explicit the policy goals of merger regulation. In the policy debate several
objectives are frequently mentioned.
· Consumer surplus. A first objective upon which merger analysis may
be based is the protection of consumer interests. If this is the case, the
central focus of merger analysis is on the competitive, or price effects, of
mergers
· Total surplus. Another objective may be to further both consumer and
producer interests. Total surplus may operationally be defined as the sum
of consumer and producer surplus. More generally, one may give different
weights to consumer and producer surplus.
· Other objectives. These include the promotion of European integration,
employment, regional balance, viability of small firms, and competitiveness
of national firms on international markets. The concern with the
preservation of employment has often been a political concern in proposed
mergers.
In principle, employment considerations should be taken into
account in a full cost-benefit analysis of mergers. In practice, a merger
policy designed to preserve old production structure is presumably not the
best way to deal with the employment objective in the long run. See for
example Jenny (1997) and Crampton for more on the employment
objective.
In the analysis below we focus on the first two policy goals. This makes it natural
to split our analysis into two distinct parts. The first part analyses the price effects of
mergers. Of particular interest is the role played by efficiencies. Is the relationship
between efficiency gains and consumer interests necessarily a trade-off? Or are there
circumstances in which efficiency gains also benefit consumers? To answer these
questions, the typology of efficiency gains provided above will prove to be very useful.
Depending on the specific type of efficiency, a merger may sometimes lower prices, to
the benefit of consumers. To assess the price effects of a merger, it is therefore crucial
to identify the specific types of efficiencies that are involved, and, less obviously, to
quantify the magnitude of these efficiencies.
The second part analyses the total welfare effects of mergers. To focus ideas, we
assume in this part that the efficiency gains are insufficient to ensure price reductions
(pro-competitive effects). We then follow a trade-off approach in which the possible
efficiency gains need to be weighted against anti-competitive effects from the merger.
Once again, the typology of efficiency gains proves useful to assess the trade-offs
involved.
Price effects of horizontal mergers
This section considers the price effects of horizontal mergers in the presence (or
absence) of cost savings. In practice horizontal mergers may also generate product
(quality) improvements. In this case, consumers may benefit from a merger even
without price decreases, provided that quality increases sufficiently. The discussion in
this section may thus be rephrased in terms of “quality-adjusted” price effects of
horizontal mergers (e.g. Rosen, 1974). The spirit of the various results will therefore
also apply to mergers with product (quality) improvements.
Since horizontal mergers reduce the number of competing firms in the industry,
the common view is that mergers tend to increase price. However, to obtain a
thorough understanding into the price effects of mergers, it is necessary to examine this
common view under various modes of competition and alternative types of efficiency.
First, consider a simple industry in which all firms have identical and constant
unit costs. Consider a merger with no efficiencies gains. In this case, most theories of
oligopoly imply that the price will increase.17 The only exceptions are if firms have a
“perfect” cartel before the merger, if one firm is failing, or if the merger triggers
immediate entry. In these cases, the price would be unaffected by the merger. What,
then, is the role played by internal efficiencies? To which extent can they ensure that
price decreases will take place after merger?
Non-collusion theories
Farrell and Shapiro (1990) consider a Cournot model, in which firms compete by
setting quantities. To simplify, they assume that all firms produce the same
homogeneous good. In this set-up they analyse the nature and the magnitude of
internal efficiencies that are required for a merger to reduce price and increase output
to the benefit of consumers.
From their analysis it is possible to draw the following conclusions:
· For price to decrease after a merger, the merged firm must realise a
substantially lower marginal cost than did either of its constituent firms
before the merger. Farrell and Shapiro (1990) provide a more precise
formula for the required reduction in marginal costs.
· If the internal efficiencies only consist of output rationalisation or fixed
cost savings, then there will be a price increase after the merger.18
Rationalisation, but not fixed cost savings, may combined with other cost
savings lead to lower prices. Given the negative result on rationalisation and fixed cost savings, one may wonder which type of efficiencies may be responsible for price reductions. The
typology of efficiency gains described in section 1.1 provides an answer.
The following types of efficiencies may ensure price reductions after mergers, at least provided that
they are “sufficiently large”:
(1) long-run economies of scale, and product-specific economies of scale
(2) technological progress, either achieved by a transfer of know-how, or by
increased incentives for R&D
(3) purchasing economies.
All these efficiencies have in common that they may lead to a (long term)
reduction in the marginal costs below those of the formerly separate firms. The
question is of course how large the reduction in marginal costs is required to be.
Interestingly, Farrell and Shapiro show that the required reductions in marginal cost
depend on relatively easy to observe variables, i.e. the firms’ pre-merger market shares
and the elasticity of demand. We return to this question in more detail in Part 5, where
we discuss a framework for merger analysis, including operational criteria for
determining the likelihood that mergers will not increase prices.
Farrell’s and Shapiro’s analysis, as most static theories of oligopolies, stresses the
importance of marginal costs rather than fixed costs in reducing prices. In a dynamic
setting, when new entry may occur, it may be possible that also fixed cost savings have
an impact on prices. At this point, there has however been little theoretical work on
this issue.
The analysis of Farrell and Shapiro establishes general results for the Cournot
model. Some industries are however better described by the Bertrand model, in
which firms compete by setting prices rather than quantities. Does Farrell’s and
Shapiro’s main result generalise to this alternative setting? In particular, can a price
reduction after a merger only be achieved through a significant reduction in the
Collusion theories
Firms may, under certain circumstances, sustain a cartel-like agreement also absent the
opportunity to write legally binding contracts. A pre-condition is that the firms are
able to detect and punish any firm under-cutting the agreed upon price, without the
help of the legal system. Firm organised punishment may, for example, take the form of
a price war that is limited in time.
Economic models of collusion (notably the supergame literature) attempt to delineate the exact conditions under which such cartel-like agreements can be sustained by the firms. These models may also allow us to assess whether mergers affect the degree of joint market power in the industry.
A first approach to analyse the role of mergers for collusion assumes that there is
initially no collusion, and asks whether and under which circumstances collusion is
likely to be facilitated as a result of the merger.
Recent theories of collusive behaviour predict that there exists a positive relationship between market concentration and the likelihood of collusion. Formal and informal contributions by Stigler (1964), Friedman (1971), Davidson and Deneckere (1984), Oliner (1982), among others, indicate that
more concentrated market structures facilitate collusion for a variety of reasons.
Increased concentration implies reduced profits from cheating by “stealing”
competitors’ market shares; increased possibilities to detect cheating; and less coordination problems. Compte, Jenny and Rey (1998) emphasise the role of capacity
constraints for the sustainability of collusion. A firm’s capacity constraint determines
how attractive it is for the firm to under-cut the collusive price. The capacity constraint
also determines how easy it is to flood the market to punish other firms that deviate.
Compte, Jenny and Rey show how merger-induced asymmetries in capacity may affect
the sustainability of collusion.
differentiation, since it implies a change in product diversity. Moreover, the welfare analysis is
complicated by any change in product variety.
assume there are two firms, one with a low (constant) marginal cost, and another with at a high (constant) marginal cost. In a fully collusive agreement firms will bargain to allocate production such that the price lies in between two hypothetical monopoly prices: the price that the low cost firm
would choose if he were a monopoly, and the – higher – price that the high cost firm
would choose if he were a monopoly. Now consider a merger between the two firms
in such a fully collusive industry. A merger would simply rationalise production by
transferring all production from the high cost to the low cost plant. As a consequence,
the price drops to the lower monopoly price. Therefore, under full collusion a merger
to monopoly reduces price even without any efficiencies other than output
rationalisation.
One may argue that this reasoning assumes that firms can enforce a perfect cartel
before the merger. Yet Verboven (1995) shows that this result generalises to industries
1.3.1.3 Conclusion
The above discussion shows that there are many different modes of (price)
competition, and accordingly, many different models of (price) competition. Actually,
any description of competition, including ours, is bound to be stylised. Just to mention
one example, the mode of competition is much affected by the exact information that
the firms have about themselves, about their competitors, and about their market. The
effect of merger on price has been studied for some such cases, but not for all of them.
Another, but related, problem is that some aspects of competition have hardly been
begun to be analysed in economic theory. (That is, there are still too few models of
competition!) One example is that almost all economic models of competition are
based on the assumption that the firms’ managers are fully rational, while they in reality
only are boundedly so.
This multiplicity of modes and models of competition creates a problem for
policy design. Should policy be based on only one of the models? If so, which one? Or,
are there ways in which policy can be made more flexible? The ideal choice, of course,
would be if competition authorities, in each individual merger case, apply that model
which conforms best to the actual mode of competition. In our policy discussion in
Part 5, we will describe some methods that can be used that are based on explicit
models of competition. We also describe simulation analysis, a flexible technique that
may be tailored to fit the situation at hand even better.
However, these methods presume that the competition authority is rather well informed about the mode of competition in the market that they investigate. If that is not the case, we suggest that
the authority may use an approach that is based on a worst-case scenario. Hereby, the
authority may combine flexibility with limited information. We also suggest that if the
firms, that presumably are more informed about the details of the competitive situation,
are not satisfied with such an analysis, and can provide a more well-tailored (and
verifiable) analysis, then the agency may use that instead.
1.3.2 Total surplus
Most economists take the view that competition policy should not be designed solely
to protect the interests of consumers. A common policy goal formulated by economists
is the maximisation of “total surplus”, the sum of consumer surplus and producer
surplus. Under this policy goal, transfers from consumers to producers due to price
increases are not considered as a problem. However, increased prices yield an
allocative inefficiency, i.e. a dead-weight loss due to sub-optimal production and
consumption.
1.3.2.1 Williamsonian merger analysis
Perhaps the most influential contribution which advocated the total welfare approach in
merger analysis is by Williamson (1968). Williamson proposed to compare the deadweight losses due to price increases after merger with the internal efficiencies that are
generated. Williamson concluded that cost savings need not be very high to
compensate for dead-weight losses induced by price increases.
We now illustrate Williamson’s analysis in Figure 1. We will show that the “trade-off” framework is
useful, though the conclusion that only small cost savings are required is not general,
since this depends on how intense one assumes competition is before and after the
merger. This conclusion is similar in spirit to the one obtained when discussing the
price effects of mergers. The degree of competition, both before and after merger, are
essential in examining the price effects of mergers.
Consider a homogeneous goods industry where unit costs are constant and equal to
AC1 before the merger, and drop to AC2 after the merger, as illustrated in Figure 1.
Consider three alternative scenario’s on how behaviour changes due to the merger:
(1) From perfect competition to monopoly.
This is the most simple case to analyse. Total welfare before the merger
coincides with total consumer surplus at the competitive price P1, indicated by the
triangle ABC. At the competitive price, there is no producer surplus. After the merger,
unit costs drop and the monopoly price P2 is charged. Total welfare is now the sum of
consumer surplus, ADE, and producer surplus, DEFG.
As a result, the change in welfare due to merger is the difference between the rectangle, BHFG, and the triangle, EHC. Intuitively, the cost savings evaluated at post-merger production (BHFG) need
to be traded off against the net losses from reduced consumption (EHC). The specific
example on Figure 1, which considers a unit cost reduction by 25%, shows that the
internal cost savings outweigh the losses from reduced consumption. But note that if
we had considered a unit cost reduction by only 12.5%, the reverse would have been
true.
(2) From perfect competition to “partial monopoly”
This is the original scenario, considered by Williamson’s article of 1968.
Williamson thus assumes that a merger in an initially competitive industry “extends
market power”, but not to the monopoly extreme. Consider for example a price rise to
P3 instead of P2. The result is a higher output after the merger, implying higher internal
cost savings than in the first scenario, amounting to BIFJ. These need to balanced
against the losses from reduced consumption, which are now only the triangle KIC.
The unit cost reduction by 25% generates internal cost savings which by far exceed the
losses from reduced consumption on Figure 1. In fact, under the assumed price
increase of this example, total welfare would increase for any reduction in unit cost by
more than 2%. This confirms the claim made by Williamson.
The difference between scenario 1 and scenario 2 shows that it is important to
know the extent of the price increase after the merger. Scenario 1 had been considered
by DePrano and Nugent (1969) in a critical review of Williamson’s analysis.
Williamson (1969) replied that the drastic price rise to monopoly considered by
DePrano and Nugent (1969) is unrealistic, since a merged firm needs to take into
account the risk of entry when raising its prices. Entry can almost never be blockaded,
argues Williamson, and therefore one may expect that a merged firm will follow a
“limit pricing strategy”, taking into account potential competitors as well as actual
competitors.
The debate between Williamson and Deprano and Nugent illustrates the
importance of predicting the extent of the price increase after a merger. In other
words, it is necessary to properly assess the nature of post-merger competitive
interaction. The third scenario stresses that one should also properly assess pre-merger
conduct.
(3) From “partial monopoly” to monopoly
To illustrate the importance of considering pre-merger market power in the same
Figure 1, consider a merger in which the initial price is P3, which exceeds marginal cost
AC1. After the merger the marginal cost is AC2, and the monopoly price P2 is charged.
In this scenario, total welfare changes from the area AKBI before the merger to
the area AEFG afterwards. Therefore, the trade-off is between internal cost savings of
BHFG and losses from reduced output of LKHI. In the example of Figure 1, which
assumed a 25% reduction in unit costs, the internal cost savings would be sufficient. In
fact, in the example any reduction in unit costs by 9% would suffice for welfare to
increase.
A general intuition on the trade-off between internal cost efficiencies and the
losses due to output reduction can be by considering a merger in industry which is
initially non-competitive, and which causes a “small” reduction in total industry output,
accompanied by a “small” internal cost efficiency. Generally speaking, the gain because
of internal cost savings is proportional to the industry output, the loss because of
reduced output is proportional to the price-cost margin. To assess the welfare tradeoff, one therefore needs to have a good estimate of both pre-merger output and of premerger price-cost markups. Pre-merger output times the expected cost reduction
would be approximately equal to the expected welfare gains. Pre-merger markups
times the expected output reduction would be approximately equal to the expected
welfare losses. Of course, the main empirical difficulty in comparing the expected gains
to the expected losses is in the assessment of the expected cost reduction and the
expected output reduction.
Which types of cost savings are valid to apply the framework of Williamson?
Note that Williamson’s formula does not depend on the actual type of efficiency that is
realised; it is sufficient that the efficiency implies a reduction in average costs. There
for, all types of efficiency in principle qualify for a Williamsonian type of defence,
provided, of course, that they involve sufficiently large average cost reductions.
Nevertheless, a case may be made that efficiencies that involve a marginal cost
reduction are superior to those that do not. This is because in this case, it may
generally be expected that the price increase will be lower than when there is no
reduction in marginal cost.
1.3.2.2 Externality analysis
Williamsonian analysis evaluates the efficiency gains from unit cost savings over the
total industry output. This approach is justified under the assumption that all firms in
the industry participate in the merger. Of course, in practice, most mergers only take
place among part of the firms in the industry. In this case, one should evaluate the
internal cost savings at the – smaller – output of the merging firms. Another difficulty
with Williamsonian analysis is that one needs to be able to measure the actual cost
reduction that will be generated by the merger. It is, of course, in the interest of the
merging firm to claim as high efficiencies as possible.
Farrell and Shapiro (1990) propose a methodology for evaluating mergers
among some of the firms in the industry without a need for relying on internal
efficiency claims. In a general Cournot setting, they propose to evaluate the externality
Farrell’s and Shapiro’s approach are twofold:
(1) they point out a potentially important effect that has traditionally been ignored in assessing
merger: a reallocation of output to competitors;
(2) they demonstrate that due to this effect mergers may be beneficial even when there are no internal efficiencies (or when internal efficiencies cannot be proven). Their analysis applies to Cournot competition.
1.3.2.3 Other studies
There has also been some literature that tries to draw inferences on the welfare from
the mere fact that merger is proposed and thus is profitable.
Levin uses a model with homogeneous goods and quantity-setting firms to show
that output-reducing mergers are never profitable if market share is less than 50%
(assuming no internal efficiencies other than output rationalisation). Furthermore, any
proposed (thus profitable) merger with a market share less than 50% increases welfare,
whether it increases or decreases output.
Werden and Froeb (1998) consider the entry-inducing effects of mergers. Entry
reduces the possibilities to raise price and thus the profitability. Hence, if we observe a
(profitable) merger, it seems more likely that it will reduce price (and thereby deter
entry). The entry possibility thus allows us to infer price reducing efficiencies. As
stated by Werden and Froeb: “If firms are rational and informed, they merge only if
they expect significant efficiency gains generated from merger, or they perceive
substantial entry obstacles such as sunk costs. Consequently, the entry issue can be
collapsed into efficiency considerations, and in the absence of strong evidence that an
otherwise anti-competitive merger generates significant efficiency gains, there is a
sound basis for presuming that entry obstacles will prevent entry in response. (Thus the
best way for courts to treat entry in many mergers may be not to consider it at all.)”
2. EMPIRICAL EVIDENCE
This part of the chapter reviews the empirical evidence on mergers. Most of this
literature has not attempted to directly identify efficiency gains from mergers. The
available empirical evidence may nevertheless help us to shed some indirect light on the
existence and the magnitude of efficiency gains.
Many studies have considered the effect of mergers on company performance, as measured by the accounting profits or the share prices of the firms. If company performance is found to increase, there is evidence that the merger created either market power or efficiency gains, or a
combination of both. Unfortunately, such evidence is not sufficient to discriminate
between the two explanations. Additional information may be gathered to obtain more
conclusive results. For example, evidence on consumer prices and market shares are
also potentially useful in disentangling the market power and efficiency effects.
There are also some papers that study the effect of mergers on the competing firms’ share
prices. If these also rise, this suggests that market power is more important than
efficiency gains, since increased market power tends to benefit the competitors,
whereas reduced production costs tend to harm competitors.
The evidence on share prices, profitability and consumer prices may not only tell
us something on the relative importance of market power and efficiency motives for
merger. It may also indicate how the gains or losses from mergers are distributed
between different groups in the economy. Since competition policy in the E.U. and
other jurisdictions are partly motivated by distributional concerns, such knowledge is
crucial for the proper design of merger control.
Section 2.1 considers the evidence on company performance, measured by
accounting profits and stock prices. Section 2.2 discusses the literature that has
attempted to disentangle the two motives market power and cost reductions. Finally, in
section 2.3 we provide evidence from a limited number of studies that have attempted
to directly quantify the importance of efficiency gains.
The main impression is that the available evidence is very limited in many
respects. First, there are surprisingly few studies of the effect of merger on
productivity, price, and market shares. Second, the evidence on company performance
is contradictory, and troubled by methodological problems. Third, most studies
concern mergers in manufacturing during the 1960s and 1970s, with a clear bias
toward mergers in the U.S. This evidence is not sufficient to draw any conclusions
about the likely effects of mergers in different sectors.
Bearing in mind the above reservations, our main conclusions are the following.
First, the empirical literature does provide some support for the fear that horizontal
mergers increase market power. Second, there seems to be no support for a general
presumption that mergers create efficiency gains. Third, in particular cases, however,
mergers do create efficiencies. Moreover, empirical evidence indicates that costs
savings are passed-on to consumers in the form of a downward push on price. It
appears that there may be substantial variance in the efficiency gains from mergers.
(Unfortunately, however, there does not exist any formal statistical testing that clearly
indicates that the variance is high.)
In sum, the empirical evidence suggests that controlling mergers is important, and that the presence and magnitude of efficiency gains may need to be examined on a case-by-case basis.
To complement this picture we believe that it would be valuable with a follow-up
study. Such a study should, in our view, focus on empirical estimates of returns to
scale. Such studies can be used at least as indirect evidence for the efficiency gains
from mergers. It would also be desirable to deepen the survey into pass-on. In both
cases, it should be possible to obtain results for specific industries.
2.1 Overall Company Performance
2.1.1 Effects of mergers on profits
The industrial economics approach studies merger performance by measuring the
(accounting) profits of the merging parties before and after the integration. From a
theoretical point of view, a merger may increase or decrease profitability. On the one
hand, oligopoly theory predicts that horizontal mergers increase market power, i.e. the
firms’ ability to set a price above marginal cost. On the other hand, mergers tend to
lower the merging parties’ market shares, since the rival firms may have an incentive to
expand their production as a result of the merger. The net effect of an increased markup but decreased sales is ambiguous. In addition, the effects of mergers on the internal
efficiency of the firms are also ambiguous. A merger may lead to rationalisation or
scale efficiencies, but it also reduces competition in the product market, and may hence
increase or create slacks in the organisation. Hence, the effect of a merger on the
merging firms’ profits is an empirical issue.
2.1.1.1 The European experience
A large-scale project consisting of studies from several European countries is reported
in Mueller (1980a). The study by Mueller (1980b), discussed below, on U.S. mergers,
is also part of this international comparison. An important advantage of this project is
that all the studies use the same methodology. The data cover firms undertaking
mergers during the period 1962-1972, for the five years preceding the merger and the
five years afterwards.
All studies use several different tests for the effect of mergers on
profitability. Two measures of profitability are adopted, namely the rate of profit on
equity, and the rate of profit on total assets. Two alternative methods are used to
control for external shocks. First, the merging firms’ profits are compared with a
control group consisting of a pair of firms that are similar to the merged ones. Second,
they are compared to the industry average.
Kumps and Wtterwulghe (1980) study 21 mergers in the manufacturing and
retail sectors in Belgium during the period 1962 to 1974. They show that the merging
firms performed better than a control group of non-merged companies. Their profit
rates declined by less between the 5-year pre-merger period and the 5-year postmerger period, than those of the control group firms. Cable, Palfrey, and Runge (1980)
study around 50 mergers in Germany during the period 1964 to 1974. They find that
the merging firms did better than the control firms, although not in a statistically
significant way. Jenny and Weber (1980) study 20 – 40 (depending on the test)
One may argue that firms would not merge unless a merger is profitable. However, there are
circumstances under which firms may merge, even if that lower their profits
2.1.1.2 The US and Japanese experience
conclusion: around 60 percent of the merging firms increased profitability,
2.1.2 Effects of Mergers on Share Prices
A second strand of literature has assessed the performance of mergers by studying the
effects they have on the stock prices of the participating firms at the time of the merger
announcement. The empirical methodology is called event study analysis: to estimate whether there is a significant difference in stock prices a few weeks before and after an “event” such as a merger announcement. The change in stock prices is measured net of market-wide price movements.
An important potential advantage of event-studies is that stock prices reflect the
present value of expected future profits created by firms, under the assumption that the
stock market is efficient. This approach differs fundamentally from the profitability
studies based on accounting data. Accounting profits only refer to current profits,
which may be subject to exceptional temporary gains or losses; in addition, the
accounting profits depend on the precise methods that have been used to classify
revenues and costs.
2.1.2.1 Returns to targets
The average target shareholder gain varies between 20 to 35 percent. The target gains more in tender offers than in mergers. The gain is higher if there is more than one bidder. The gains have varied over
time, due to changes in the legal and institutional environment.
the total return to the target consists of two parts: the mark-up (which is the increase
in the stock price beginning the day the first bid is announced), and the run-up (which
is the increase in the stock price in the period of 40 days before the first bid). The
mark-ups are unrelated to the size of the run-up. One interpretation of this
result is that the run-up reflects information held by other potential bidders, rather than
insider trading or information leakage. If this interpretation is correct, then previous
event studies may have exaggerated the gains from mergers since they normally include
the run-up as a part of the merger premium.
2.1.2.2 Returns to bidders
bidders on average break-even, or do slightly better.
2.1.2.3 Total returns
shareholders of target firms realise large positive abnormal
returns in take-overs. The evidence on the rewards to bidding firms is mixed, but
suggest that they break even. Since targets gain, and bidders do not appear to lose, then
take-overs create value to the participating firms.
However, as pointed out by Jensen and Ruback (1983) the bidding firms tend to be
larger than target firms. Hence, the sum of the returns to bidding and target firms does
not measure the gains to the merging firms.
we conclude that the total gains on average are positive…..However, the long-term effects may be different. Rau and Vermaelen (1998) show that bidders in mergers under-perform (while bidders in tender offers outperform) the market in the three years after the acquisition. The long-term underperformance of acquiring firms is not uniformly distributed across firms. It is
predominantly caused by the poor post-acquisition performance of low book-to-market
“glamour” acquirers. To explain their findings, Rau and Vermaelen propose a
performance extrapolation hypothesis. The management, as well as the market, the
board of directors, and the large shareholders over-extrapolate the past performance of
the bidders’ management when they assess the benefits of an acquisition. That is, the
market overestimates (underestimates) the ability of glamour bidders (value firms) to
manage other companies’ assets.
2.1.3 Summary and conclusions
Profit flow studies:
If a generalisation is to be drawn from the profit-flow studies, it would have to be that mergers have but modest average effects on the profitability of the merging firms. A robust finding, however, reappearing in many studies is that the profitability of mergers is not guaranteed: a large proportion of mergers reduces profitability.
This result can be interpreted in several ways. The worst case scenario is that in
many mergers market power is increased, but that the mergers create internal
inefficiencies in the firm that offset any profit increases. Another possibility is that
there were no effects on production costs on average, and that losses are due to
competitors expanding their production in response to mergers. This is a very
surprising result. It is not difficult to understand that a merger can be socially
undesirable. But why should firms engage in activities that are even bad for them? This
is an issue to which we will return below.
It is important to keep in mind that there is variance around the mean profit
effects. Some mergers are profitable, and some mergers are unprofitable. This result is
consistent with the view that some, but not all, mergers do create efficiencies. This is
important, since it suggests that policy should not count on general cost savings from
mergers, but that policy somehow should allow for cost savings in specific cases–
perhaps an efficiency defence.
the use of competing firms to control for exogenous shocks, creates its own problems. The
reason is that a merger normally confers an externality on the outsiders. When two
firms merge and reduce competition, the remaining competitors also gain. In fact, the
outsiders may gain more than the merging firms (so called strong positive
externalities). The reason is that the outsiders gain from the increase in the price,
without having to reduce their own production. Thus, what appears to be an
unprofitable merger, may simply be a profitable merger with a strong positive
externality. The reverse is also true. What appears to be a profitable merger may be an
unprofitable merger with strong negative externalities. These problems warn us that
one should interpret the results of the studies with care.
Event Studies evidence shows that the target firms’ shareholders benefit (20-35 percent), and the bidding firms’ shareholders generally break even (in the short run). Moreover, the combined gains are mostly positive (up to 30 Rau and Vermaelen interprete a low book-to-market value as an indication that the current 10 percent gains). The increased total gains seems to suggest that merging firms either increase market power or experience efficiency gains, such as cost savings.
Unfortunately, almost all event studies discussed above use data from the U.S. It
is not clear to what extent these studies are also representative for Europe.
Furthermore, recent results invite to some caution concerning the event study
methodology. First, Rau and Vermaelen (1998) show that the bidders’ long term
performance is often negative because the short-term price movements in the stock market are not very good at predicting merger performance. If this interpretation is correct, Rau and Vermaelen’s results should be perceived as evidence against the efficient market hypothesis underlying the (short-term) event study methodology.
changes in share-prices may reflect other information
than the profitability of the merger (see below).
From this evidence, our two main conclusions are:
1. Merger control should not be based on a general presumption that mergers
create cost savings. The presence and magnitude of efficiency gains may
need to be examined on a case-by-case basis.
2. The available evidence on profitability and share prices does not point at
any easily observable conditions under which to expect efficiency gains.
We have already argued that both studies of profitability and studies of share
prices are troubled by methodological problems. Furthermore, if one combines the
evidence from studies of profitability and studies of share prices the emerging picture is
mixed, and perhaps even confusing. The available evidence indicates that a large management is efficient in managing the firm’s assets.
Why do unprofitable mergers occur?
How share prices increase when profitability falls?
Most theoretical explanations for why unprofitable
mergers may occur rely on the idea that owners of the firms lack the instruments to
discipline their managers, and that the managers consistently overestimate their abilities
(Roll, 1986), or that the managers are motivated by a desire to build a corporate
empire (Shleifer and Vishny, 1988). Neither the hubris nor the empire building
hypothesis explains why the share prices may increase at the same time as the profit
flows decrease. In an attempt to answer both puzzles, Fridolfsson and Stennek (1998)
propose a hypothesis called the pre-emptive merger motive. Firm A may merge with
firm B, even if the merger reduces their combined profit flow as compared with the
status quo. This would happen if the relevant alternative is that firm B merges with
firm C, and this alternative merger would reduce firm A’s profit flow even more.
Expressed differently, even if a merger reduces the profit flow compared to the initial
situation, it may increase the profit flow compared to the relevant alternative—another
merger. Furthermore, even though such a pre-emptive merger reduces the profit flow,
the aggregate value of the firms—the discounted sum of all expected future profits—
may actually increase.
The reason is that the pre-merger value of the firms accounts for
the risk that the firms may become outsiders. Under the hypothesis that the stock
market is efficient—in the sense that share prices reflect firm values—these results
demonstrate that the two strands of the literature may be consistent. In particular, the
event studies can be interpreted to show that there exists an industry-wide anticipation
of a merger, and that the relevant information content of the merger announcement is
which firms are insiders and which are outsiders.
To sum up, these theories give a rather pessimistic picture of mergers. They may be motivated by hubris, or empire building, or an attempt to pre-empt other mergers. According to these interpretations of the available evidence, both market power and cost savings may be secondary
motives for mergers. Nevertheless, even if the motives are different, the mergers still
affect both market power and costs.
price tends to rise as a result of merger, eg as they lead to
more efficient operations. Yet, the impact of efficiency gains on price is more than
offset by increased market power. thus, wealth gains to the stockholders of merging firms do
not arise through value creation alone. relaxation of antitrust policy may result in
nontrivial wealth transfers from consumers.
-to fully understand the price effects of mergers, we should take into account other conditions that may have changed after the merger, for example changes in factor prices. This problem has been tackled by studying how the prices of the merged firms’ products have changed in comparison
to other prices.
studies show that:
-during the merger announcement period—when changes are primarily due to
the market power effect—prices are increased (plus 11 percent), but during the
completion period—when changes are primarily due to efficiency—prices decrease
(minus 9 percent). These results concern mergers that do not include failing firms.
-a significant positive relation between fare changes and changes in concentration.
-mergers do not generate gains in quality that offset the price
increases. The number of customer complaints filed with a governing agency increased
almost threefold during the studied period.
the Cotterill study observes pre- and postmerger market concentration, and computes how the mergers change market concentration, as measured by the Herfindahl index. Then Cotterill [using a established empirical relationship between market concentration and the market price
level], predicts the impact of the mergers on the price level. Finally, he computes the
implied increase in the consumer food bill. Using this methodology Cotterill predicts price changes in the range from 1pc to 3pc.
This methodology could be used by competition authorities to predict the
likely effect of merger on price.
2.2.1.2 Indirect Studies
The number of empirical studies of the effects of mergers on prices is surprisingly
small. For this reason it is worth to consider also the indirect but complementary
evidence obtained in studies that compare prices between geographically separated
markets with different concentration levels. If high concentration is associated with
high prices, this is indirect evidence that horizontal mergers increase price, ie the
market power effect dominates the efficiency effect.
these indirect studies how a positive relation between concentration and the price level
the cross-section studies confirm the existence of a relationship between price and concentration.
studies show that critical concentration ratios may exist in certain industries, for example, a four firm concentration ratio about 50. In other industries, for example supermarkets, no critical level was found. They conclude that a single critical concentration ratio for all of manufacturing is likely to be
incorrect.
-studies show that if prices in nearby markets are low and the distances to them are short, prices tend to be lower. It is also shown that prices are increasing in firm concentration within a local market.
To enter into a market a firm must be able to cover its fixed costs. A monopolist can charge a high margin, i.e. it can set a high price in relation to its marginal cost. As a result, the monopolist may need a relatively small market to cover its fixed costs.
2.2.2 Effects of Mergers on Market Shares
If a merger is driven by market power, the merged firm will
increase its price and lower its output. This initial change will increase the residual
demand for the competitors. As a response to the demand increase, the competitors
will increase both their prices and their output.
the market share of the merged firm will drop if the merger is driven by market power.
In contrast, if the merger generates sufficient variable cost synergies, the merging firms may
increase their market share.
mergers that tend to increase the price level
also tend to reduce the merging firms’ market share, and mergers that tend to
decrease the price level tend to increase the merging firms’ market share.
the market shares of firms engaged in horizontal
mergers decline. The evidence on the effects on market shares from non-horizontal
mergers is mixed.
2.2.3 Effects of Mergers on Outsiders’ Share Prices
The results on the merging firms’ stock price performance do not disentangle the gains
attributable to increased market power and the gains attributable to, for example,
technological efficiencies.
Unfortunately, the available evidence [on the effects of mergers on outsiders share prices] this point is not conclusive. Stillman (1983) finds no statistically significant effect on outsiders’ share prices. Eckbo (1983) finds a small but statistically significant increase. However, where competition authorities announce an investigation of the merger, the outsiders’ share prices are not affected in a significant way.
2.2.4 Summary and conclusions
There is considerable dispute concerning the welfare effects of mergers. Are mergers
mainly motivated by market power or efficiency gains? …. studies show contradictory results.
2.3 Studies of Efficiency Gains and Pass-on
2.3.1 Effects of Mergers on Productivity
The most direct way of assessing the efficiency gains from mergers is by measuring
productivity gains and economies of scale realised following a merger.
2.3.1.1 Indirect evidence
-returns to scale estimates may be used to assess indirectly the opportunity that mergers realise cost savings related to returns of scale. A second advantage with this strategy, is that is can be used to say something about individual industries.
2.3.3 Pass-on
In order to understand the welfare effects of a merger, it is not sufficient to know
whether the merger creates any efficiency gains. It is also important to know if, for
example, cost savings are passed on to consumers in the form of a lower price. This is
important in order to assess the effect of a merger on the distribution of wealth in
society. Pass-on is also important since it affects the dead-weight loss of a merger.
there is incomplete pass-through [to consumers] of import tariffs, and of exchange rate fluctuations.
who bears the burden of vatsand excise taxes?
Verboven (1998) for the car market, found significant tax incidence stemming from imperfect competition.
in agriculture, there is incomplete transmission of raw goods prices into final consumers prices, see e.g. McCoriston, Morgan and Rayner (1998). Bettendorf and Verboven (1998).
2.3.4 Summary and conclusions
There does not exist clear evidence that mergers, as a general
rule, create efficiency gains. However, at least some mergers do create efficiencies.
Moreover, empirical evidence indicates that (between 30 and 70 percent of) costs
savings are passed on to price.
To complement this picture we should focus on empirical estimates of returns to
scale; and also on pass-on…. Such studies can be used as indirect evidence for the efficiency gains
from mergers.
2.4 Distribution
The discussion above has primarily focused on the issue if mergers create any
efficiency gains. However, we are also interested in how the social surplus (or losses)
created by the merger is divided between the firms’ different interest groups [suppliers, buyers, competitors, bondholders, employees and management].
studies of consumer prices (and the merging firms’ market shares) indicate
that horizontal mergers reduce consumers’ welfare. Hence, if mergers create efficiency
gains, consumers do not, at least as a general rule, receive a share of that surplus.
Second, the event study evidence shows that the target firms’ shareholders
benefit, and that the bidding firms’ shareholders generally break even.
The unequal distribution suggests that competition authorities may want to condition the approval
of an international merger on whether the national firm is buying or being bought. In
particular, the competition authorities may be more favourable towards a merger in
which the national firm is being bought.
-effects of mergers on employees is a complex issue. A purely anti-competitive
merger reduces output and thus employment. A merger that creates efficiency
gains does so either because the assets in the two firms (including employees) are
complementary, or because some duplicated functions can be eliminated…. Thus the anti-competitive effect and the efficiency effect goes in opposite directions. The total effect: if the merger reduces output and increases consumers’ prices, then employment is likely to be reduced.
3. A “CHECK-LIST” [ to take efficiencies into account.]
Any merger control system is, at least implicitly, based on some welfare considerations.
A merger affects the well-being of many individuals. In fact, any person that belongs to
at least one of the firms’ interest groups (consumers, share-holders, management,
workers, suppliers, creditors, and also competitors) is likely to be affected.
The purpose of an explicit welfare standard is to indicate which effects are to be
taken into account [by competition authorities]….if the policy maker/regulator is concerned with the
distribution of wealth in society, the effect on each individual should be given a weight
Moreover, one may argue that hostile takeovers are likely to be triggered in those cases where
the current management has negotiated wage contracts that are too generous. The take-over
gain would then consist in renegotiations of these contracts. In that case, employees are hurt by
the transaction.
however, competition authorities analyse the effects of a merger at the level of interest
groups, and not at the level of individuals. Moreover, they do not include all interest
groups in the analysis.
1. Welfare Standard.
All competition authorities analyse the effects of a merger
on consumers and share-holders.
four poss welfare standards have been used in different jurisdictions to weigh
consumers’ interests against share-holders’ interests:
i. Total surplus (or Williamson’s) standard. According to this standard a
merger should be allowed if it creates more wealth to producers than it
destroys for consumers. Hence, distribution does not matter. It should be
emphasised that according to this standard also a merger that raises the
price may be approved.
ii. Consumer’s surplus (or price) standard. According to this standard a
proposed merger should be allowed if and only if consumers gain. The
reason to adopt this standard is if one is concerned with distribution, and if
consumers in general are poorer than the owners of the firms.
iii. Hillsdown standard. According to this standard, efficiencies must exceed
the losses to consumers. It can be shown that this standard is stricter than
the total surplus standard, but more allowing than the price standard.
Expressed differently, this means that there are distributional concerns, but
not as strong as in the consumer’s surplus standard.
iv. Killer standard. According to this standard, a merger should be allowed
only if all efficiencies are passed on to consumers. This standard is built on
very strong distributional concerns.
2. International Competitiveness.
In international markets the welfare tradeoffs are different from the welfare trade-offs made in national markets.
There are two reasons for this. First, in a market where mainly national firms sell to
mainly foreign consumers, an anti-competitive merger enhances national welfare, even
if there are strong distributional concerns (within the country). Second, in a market
where national firms compete with foreign firms, a merger of national firms that leads
to substantial cost-savings, enhances the national firms’ competitiveness. In some
jurisdictions, the international competitiveness of the domestic firms is considered an
objective for the merger control.
The Canadian Competition Act stipulates that, in the determination of whether a
merger is likely to bring about gains in efficiency, account should be taken of whether
such gains will result in: (i) a significant increase in the real value of exports or (ii) a
significant substitution of domestic products for imported products.
3. Future Viability.
Cost savings may also be necessary for the merging firms to
survive in the long run. An example of such a situation is if there are important returns
to scale, and if the merging firms are small in comparison to their competitors. Such
considerations are made in Sweden. Taking the firms’ future viability into account, as a
part of an efficiency defence, is not necessary if the merger control includes a failing
firm defence.
4. Inefficiencies.
In some cases the competition authorities suspect that a merger
will produce net inefficiencies rather than net efficiencies. Our review of the empirical
and the theoretical literature has shown that such mergers do occur. The question is
then, should competition authorities condemn a merger based inter alia on the
argument that the merger creates inefficiencies. Expressed differently, should
competition policy be used as a means to eliminate mergers that reduce firms’ internal
efficiency. Merger control in the UK includes such a possibility
Types of efficiencies.
An efficiency defence must state what types of
efficiencies are included. If a complete cost-benefit analysis is to be made, any social
efficiency that the merger generates should be included. Efficiencies may come in the
form of cost savings, improved quality, or improved services. Cost savings, in turn, may stem from rationalisation, or scale economies, or technological progress, and so on.
Poss restrictions that competition authorities, may put on the types of efficiencies they consider.
1. Redistributive (or pecuniary) gains. It is important to distinguish between
true social efficiencies and redistributive gains. This is, for example, done
in a clear way in the Canadian merger guidelines58. Only the first category
should be included in an efficiency defence. For example, cost savings
should only be included if they represent a saving of resources. If firms
save on costs because the merger increases their bargaining power and
enables the merged entity to extract wage concessions, or discounts from
suppliers (not corresponding to cost savings), that is only a wealth transfer
and it should not be counted as a social efficiency. Another example is tax
gains.
2. Quantity reductions. Another issue concerns cost savings that stem from
an anti-competitive reduction in quantity. In a complete cost-benefit
analysis, such cost savings should be considered. Of course, they should be
balanced against the reduction of consumers’ surplus, and if these are the
only cost-savings they will never suffice to make the merger socially
desirable. But the point remains, they should be counted as a cost saving.
However, there are two different ways to do this. One way is to include
these cost savings in the “stage-one” analysis that attempts to compute the
dead-weight loss due to the merger. If that is done, these cost savings
should not be counted once again in the efficiency defence. The other way
is to exclude these cost savings from the first stage, and to include them
instead in the efficiency defence.
3. Fixed costs. If competition authorities use a consumers’ surplus standard,
only savings in variable costs are normally to be considered. This means
that some types of efficiencies, e.g. duplication of administrative routines,
usually do not need to be discussed, since they do not usually affect
variable costs.
4. Other markets. Another subtle issue is cost savings (or other efficiencies)
…In U.S. case Lucy Lee/Doctors Regional Medical Center, the court held that
pro-competitive effects in one market could not justify anti-competitive
effects in a different market. In Germany and the UK, on the other hand,
the competition authorities balance anti-competitive effects in one market
against pro-competitive effects in other markets.
5. Industry level efficiencies. Another issue concerns the distinction between
efficiencies at the level of the merging firms and efficiencies at the level of
the market. An example of the efficiencies at the level of the merging firms
could be cost savings as a result of specialisation between the plants that
are owned by the merged entity. An example of efficiencies at the level of
the industry is that the merger between two firms may affect the R&D
incentives for the competitors (see sub-section 1.1.3). It appears that U.S.
and Canadian efficiency considerations only refer to firm-level efficiencies,
while Swedish efficiency considerations, in principle but perhaps not in
practice, include both types.63 A possible reason why only firm-level
efficiencies are considered is that they are easier to verify. Even if both
types are considered, one may argue that market-level efficiencies should
be treated separately, at least in assigning the burden of proof. The reason
is that it is probably only concerning the firm-level efficiencies that the
firms are (much) better informed than the competition authorities.
6. Problem of proof. In practice some efficiencies are more difficult to verify.
Some competition authorities have chosen to state explicitly which types of
efficiencies are less likely to be considered due to such problems (see
below).
6. Net effects. An efficiency defence may take into account retooling and other
costs that must be incurred to achieve efficiency gains, and deduct them from the total
value of the efficiencies – computing the net efficiency. That is done in the USA,
Canada and Sweden. Note, however, that if these costs only affect fixed costs, they
will not be passed on to consumers. Hence, including these costs is not important if
competition agencies use a consumer’s surplus standard.
7. Measurement. In some jurisdictions it is explicitly indicated what firms
should prove, and what kind of documentation they should use.
In the US merger guidelines it is said that the merging firms must substantiate
efficiency claims
8. Merger specificity. In some jurisdictions, it is argued that an anti-competitive
merger should only be approved because of the efficiencies that it creates, if those
efficiencies would not be realised through less anti-competitive means.
It may be difficult for competition authorities to judge what alternatives should
be considered. For this reason it is explicitly stated in the Canadian merger guidelines65
that only if the other means is a common industry practice will it be considered.
Examples of alternatives are: internal growth; a merger with an identified third party; a
joint venture; a specialisation agreement; or a licensing, lease or other contractual
arrangement.
Note that the “merger-specificity” requirement entails that an efficiency defence
is invoked only if the anti-competitive concerns cannot be resolved through divestiture
or other remedies.
9. Discounting. In the Canadian merger guidelines66 it is explicitly stated that to
compare efficiencies and anti-competitive effects that occur at different points in time,
one needs to remove the effects of anticipated future inflation, and apply a standard
real discount rate.
10. Mode of competition. Already an analysis of the effect of a merger on
competition (assuming that there are no cost savings) requires that competition
authorities have information about the mode of competition in the market. The analysis
of the effect of cost savings brought about by the merger also depends on the mode of
competition.
There are several issues involved. The first issue concerns whether the firms on
the market compete or whether they collude. In the US merger guidelines the first
possibility is discussed under the heading “unilateral effects” and the second under the
heading “co-ordinated effects.” Second, if firms compete, analysis of oligopolistic
11. Efficiencies as an offence (anti-competitive effects). Cost savings are per
se desirable. However, cost savings can have negative side effects. In particular, if two
firms merge and thereby lower their variable costs, they become a tougher competitor.
Actually, if the cost reduction is big enough, the merger may imply that competitors
are driven out of the market, or that new entry is blocked. In this sense, cost savings
may be anti-competitive. In a complete cost-benefit analysis such anti-competitive
effects should be included in the analysis of a merger, and there is no inconsistency if
efficiencies are treated both as an offence and as a defence. Cost savings have been
treated as an offence both in the USA (Brown shoe) and in the E.U. (MSG Media
Service).
12. Pass-on (pro-competitive effects). Competition authorities do not only need
to assess the existence and magnitude of efficiencies. They also need to assess the
extent to which the cost savings are passed on to consumers.
In a non-collusive market (unilateral effects), four issues are important for
assessing pass-on. First, it is necessary to distinguish between variable and fixed costs.
Only reductions in variable costs are (directly) passed through to consumers. Second, it is necessary to estimate the degree of post merger competition. The more competition
there is ex post, the more the cost savings will be passed on to consumers. For
example, if competition is very intense, a reduction of marginal cost by €1 would lead
to a reduction of the price by €1. If the merger creates a monopoly (and if demand is
very price insensitive) the price may essentially be unaffected by the cost savings.
Third, pass-on also depends on the exact “shape” of the demand function. In
particular, if the price elasticity of demand is higher at higher price levels, then the
effect of a reduction in cost on price tends to be small (everything else equal). Fourth,
the slope of the marginal cost function also affects pass-on.
In a collusive market, all three aspects mentioned above are still crucial. In order
to estimate the degree of post merger competition, it is also necessary to estimate the
degree of post merger collusion. The more collusion there is, the less a given reduction
in cost will be passed on to consumers (much like the non-collusive case). It may be
more difficult to assess the effect of efficiencies on the price level in a collusive market.
The reason is that cost savings may reduce the likelihood that collusion is successful.68
However, economic theory is not well developed to analyse these issues.
In one US case (Long Island Jewish Medical/North Shore Health) pass-on was
considered likely since the merging parties where not-for-profit organisations.
13. General-presumptions versus case-by-case methodology. Essentially all
horizontal mergers increase market power and reduce allocative efficiency. If this was
the only effect of horizontal mergers, a general prohibition against horizontal mergers
would be the natural policy. However, mergers also lead to cost-savings and other efficiencies. There are two main methods by which efficiencies are balanced against the
reductions in allocative efficiency in merger control today:
i. The general-presumptions method: Not all horizontal mergers are
prohibited. Typically, only those mergers that significantly reduce
competition may be challenged. For mergers that only reduce competition
insignificantly (according to some indicator such as the Herfindahl measure
of concentration or the merging firms’ market shares), cost-savings are
presumed to be more important than the anti-competitive effects. The
global method is used in by the Bundeskartellamt in Germany. According
to some, including the E.U. Commission, the global method is used in the
E.U. merger control.
ii. The case-by-case method: Efficiencies may be balanced against anticompetitive effects on a case-by-case basis (efficiency defence). Elements of the case-by-case method is used in the USA, Canada, France, UK and Sweden.
iii. The sequential method: For mergers that only reduce competition
insignificantly, cost-savings are presumed to be more important than the
anti-competitive effects. Case-by-case considerations of efficiencies are
initiated as a second step, in case the analysis of a first step suggests that
the proposed merger causes non-negligible concern for anti-competitive
effects. Symmetrically, one may use general presumptions to define a level above which
anti-competitive effects are presumed to dominate (and case-by-case analysis is not
allowed). In case both a high and a low threshold are defined, case-by-case
considerations are allowed only in intermediate (or perhaps marginal) cases. For
example, according to the US merger guidelines efficiencies almost never justify a
merger to monopoly or near-monopoly. Similarly, according to Article 81(3) of the
Treaty of Rome, an anti-competitive agreement cannot be exempted if the agreement
gives the firms the possibility of eliminating competition in respect of a substantial part
of the products in question.
14. Burden of Proof.
the main problem with an efficiency defence is that, in many jurisdictions, it is the firms that have the burden of proving that a merger (already found to be anti-competitive) produces sufficient efficiencies not to be blocked. This is the case in the USA and Canada, and in the EU concerning
exemptions for agreements that restrict competition. A likely reason for this is that it is
the firms that have the best information.
15. Standard of Proof.
As it is difficult to predict efficiencies before a merger is consummated, one should rather talk
about a persuasion standard. But the question remains, for an efficiencies defence,
should one require that the cost savings be possible, probable, or virtually certain?
Some competition authorities clarify which types of efficiencies that are less likely to be considered…. That helps to clarify the standard of proof that is used
17. Prosecutorial discretion versus Litigation : Efficiencies may be considered either by the competition agencies, or by the courts, or by both.
18. Rebuttal versus Defence. T
here are at least two different ways that one may
frame an efficiency justification for a merger, namely as rebuttal or as defence. The
difference originates from the fact that the term “competition” can be used in two different meanings:
i. According to one definition, the degree of competition is measured by the
price level (or perhaps consumers’ satisfaction). Using this definition,
horizontal mergers can be both pro-competitive (those that reduce price),
and anti-competitive (those that increase price).
ii. According to another definition, the term “competition” refers to the
degree of market power. Market power may be measured by the price to
cost mark-up that firms charge. With this definition, all horizontal mergers
reduce competition (i.e. increase mark-ups) at least slightly, and some of
them significantly reduce competition.
Merger control forbids mergers that reduce competition (significantly). The exact
meaning of this prohibition depends on which of the two meanings that one gives the
term “competition.” Consider a merger that increases mark-ups significantly, but
reduces price (due to large counteracting cost savings). Since mark-ups are increased,
such a merger would be blocked using the second interpretation of the term
“competition.” Since price is reduced, such a merger would be permitted using the first
interpretation of the term “competition.”
Furthermore, depending on the interpretation of the term “competition,” an
efficiency justification of a merger must be framed differently. If competition refers to
the price level, a merger should be blocked if, and only if, it raises the price
(significantly). With this definition one may justify a horizontal merger by rebutting a
claim that it reduces competition.
i. Rebuttal: Because of the efficiencies, the merger does not lessen
competition (it does not increase price), and should be allowed.
In contrast, if competition refers to the degree of market power (mark-up),
almost all horizontal mergers reduce competition, and are consequently hit by the
ii. Defence: The merger lessens competition (it increases mark-ups).
However, due to the efficiencies, it is nevertheless desirable, and should be
allowed.
4. A FRAMEWORK FOR MERGER ANALYSIS
This part aims to bring the threads together and propose a framework that may be used in
incorporating efficiencies into merger analysis. Broadly speaking, three approaches
may be distinguished.
-The case-by-case approach explicitly analyses the magnitude and
effects of efficiencies in every merger case. This approach has the potential problem of high information costs in measuring efficiencies and their effects.
-The general presumption approach makes use of general structural indicators (such as market shares) with an implicit recognition on the existence of average efficiencies in mergers. This approach has the potential problem that there is a lot of aggregate uncertainty concerning efficiencies from mergers, in which case the structural indicators are not perfect predictors of the net benefits from mergers.
-the “sequential” approach, which aims to combine the
relative advantages of both extremes by minimising on both information costs and
errors that may arise from the unreliability of structural indicators.
4.1.1 Merger decision-making and types of errors
By its very nature, merger policy offers only a limited set of choices to influence
competition. The competition authority either accepts the merger, rejects it, or, as a
compromise, accepts the merger conditional on certain requirements such as
divestiture. The limited, discrete set of alternatives in merger policy is in stark contrast
with regulation, where a larger and more flexible set of policy instruments is available
to influence competition. The regulation of prices, for example, allows for an infinite
number of alternatives.
Whatever the goals of merger policy are (to protect consumer interests,
international competitiveness, or total surplus), the merger authority faces the
tremendous task of computing the net effects of a merger, by comparing the effects on
market power with the effects on various types of efficiencies. The empirical evidence
in Part 3 indicates that even for firms it may not be an obvious task to compute their
net private benefits from merger. The task for the merger authority is even larger, since
they have considerably less knowledge regarding the markets in which the firms
operate.
Given the difficulties in computing the net effects of a merger, and given the
limited set of alternatives available to the competition authority, one may distinguish
between two types of error:
· Type 1 error: Accept a merger that has net harmful effects.
· Type 2 error: Reject a merger that has net beneficial effects.
The challenge is to design and apply a policy that minimises the consequences of
the two types of errors. This does not mean that it is necessary to minimise the number
of errors per se. Rather, it means that a good policy should minimise type 1 errors in
those cases where significant net harmful effects are likely; and minimise type 2 errors
in those cases where significant net beneficial effects are likely. At the same time, it is
necessary to economise on information costs.
4.1.2.1 The case-by-case approach
A first way to analyse the net effects of mergers, including both market power and
efficiencies, is to simultaneously consider all factors for each individual merger that is
proposed. Such a case-by-case approach explicitly recognises the possible presence of
efficiencies in every single case. One cannot speak, however, of an efficiency defence,
since efficiencies are considered in the merger analysis in a fully integrated way. A
case-by-case approach would, in principle, allow one to keep type 1 and type 2 errors
at a minimum, at least to the extent that the investigation is successful without any
unforeseen contingencies.
In practice, however, a case-by-case approach entails tremendous information
costs. Broadly speaking, there are two types of information gathering activities:
information gathering regarding market power effects, and information gathering
regarding efficiencies.
4.1.2.2 The general presumptions approach
An alternative approach to evaluate mergers and incorporate efficiencies is by relying
on general presumptions about their likely effects. An extreme example would be to
forbid all horizontal mergers (net effects are presumed to be negative) or to allow all
mergers (net effects are presumed to be positive). A better variant, however, is to
make the approval contingent on some easily observable indicators that contain some
(but imperfect) information about the likely net effects of the merger. In particular,
based on past experience regarding the types, the magnitude and the effects of
efficiencies associated with mergers, one may construct structural indicators, such as
market shares or concentration indices, that measure the average net benefits from
mergers. The general presumptions approach is therefore based on the implicit
recognition of efficiencies.
The general presumptions approach obviously eliminates the high information
costs involved in assessing mergers on a case-by-case basis. Instead, there is now a
need for a set of structural indicators that measure the average net expected benefits
from mergers. The quality of these structural indicators as a predictor of the net
benefits from a specific merger depends on the degree of aggregate uncertainty
regarding merger effects.
4.1.2.3 The sequential approach
The information costs associated with the case-by-case approach and the aggregate
uncertainty problems associated with the general presumptions approach raise the
question whether there exists an intermediate approach that combines the advantages
of both extremes. The sequential decision-making approach aims to combine aspects of
both approaches. Broadly speaking, a sequential decision approach starts by assessing
mergers based on general structural criteria. If the merger meets the criteria, then it is
accepted. If not, a more detailed analysis is allowed into some, though not necessary
all, aspects of the merger.
Most countries do, in fact, adopt some version of the sequential approach in
evaluating mergers. For example, a first stage of the procedure consists of some
routine tasks to check whether, for example, the total turnover of the merging firms
does not exceed a critical level. Only if the firms do not meet these criteria, a further
investigation is followed. The reason for this procedure is that information costs are
too high to justify a detailed investigation of all (even “minor”) cases, yet aggregate
uncertainty about the merger’s effect is too high to limit the analysis to a routine test
on all (in particular borderline) cases.
Once a merger fails the routine tests in the first stage, a more detailed
investigation starts. Again the question arises whether the merger authorities should
simultaneously consider all factors to assess the net benefits of the merger, or whether,
in contrast, a more limited investigation based on general presumptions is preferable. In
principle, the various dimensions that may be investigated either simultaneously or in
sequential steps include: the definition of the relevant market and a market share test,
an entry barriers test, a more detailed test on expected market power effects (e.g. a
collusion test), a failing firm test.
4.1.3 The efficiency defence
For simplicity, we limit our attention to two central dimensions in merger analysis:
market power (or “anti-competitive”) effects and efficiencies.
The sequential approach
then amounts to a two-step approach. In the first step simple structural indicators are
computed which implicitly aim to incorporate both market power effects and
efficiencies. During this step, there is no explicit analysis of efficiencies, but rather
some general presumptions. If the structural indicators meet certain thresholds, then
the merger is automatically accepted or rejected. In other cases an efficiency defence is
allowed in a second step. In this step market power and efficiency effects are balanced
in a detailed and explicit manner.
Two central questions arise in the implementation of the efficiency defence.
(i) How should the thresholds for the structural indicators be determined in
the first step?
(ii) How should the detailed analysis of market power and efficiency effects be
performed in the second step?
4.2 Calculating minimum required efficiencies (MREs)
If a merger has failed the test for acceptance or rejection in the first step, then a second
step with an efficiency defence may be started. This step requires a more thorough,
case-by-case investigation into the possible net beneficial effects arising from
efficiencies. It should be emphasised that this step does not simply require an adequate
measurement of the actual (expected) efficiencies from the merger, but also a good
calculation of the market power (or anti-competitive) effects. Both aspects should be
measured in comparable units so as to arrive at an overall assessment of the net
benefits. Conceptually, one may thus distinguish between:
(1) the measurement of actual efficiencies, and
(2) the calculation of minimum required efficiencies (MREs) in order to
compensate for market power effects.
4.2.2 Minimum required efficiencies in a worst case scenario
4.2.2.1 Expected price increase in a worst case scenario
The expected percentage price increase following a merger may be decomposed into
two separate components. First, there is the expected price increase arising from
increased (unilateral or collusive) market power, holding costs constant. Second, there
is the reduction in marginal cost multiplied by the expected pass-on to consumers.86
The problem with computing the first component is that one needs to have an
idea of the current and future expected intensity of competition. Are firms currently
competing according to the Cournot model or according to the Bertrand model? Are
they currently colluding or behaving in another, less well-defined way? How will firms
behave after the merger? Is the merger likely to increase the likelihood of collusion, or
are only unilateral market power effects to be expected?
To avoid these difficult questions, the merger authorities may opt for a “worst
case scenario”:
if the two merging firms would not be able to profitably increase their prices by 5-10%, then the
relevant antitrust market would be larger than these two firms, and one can conclude
that there is at least no merger to “monopoly.” The competition agency has to continue
to include additional competing products to the definition of the relevant market until a
sufficient number of products is reached so as to make the hypothetical joint price
increase profitable.
If the relevant antitrust market is defined according to this procedure, the worst
case scenario would be that all firms in the market would increase their price by 5-10%
(assuming no changes in efficiencies). This is because a larger price increase would by
the definition of the relevant antitrust market not be profitable to the firms.
4.2.2.3 Assessing pass-on in practice
-pass-on is incomplete.
Can efficiencies be created through alternative (non merger) means?
The following alternatives:
(1) internal growth,
(2) a joint venture,
(3) a specialisation agreement,
(4) a licensing, lease or other contractual
agreement,
(5) another merger.
A reduction in competition is normally believed to produce dead-weight losses (allocative inefficiencies). However, according to the theory of second-best, a reduction of competition in one
market may be beneficial for allocative efficiency, if competition is already low in markets for
close substitutes. If competition authorities would include efficiencies in other markets, one may argue
that they also should take into account second-best considerations.
4.3.3 How can efficiencies be verified?
The most difficult issue regarding the analysis of efficiencies is their verification.
Special care has to be taken so that the analysis of efficiencies can be undertaken with
a reasonable degree of confidence and without relying on too costly information
collection activities.
4.3.3.1 Burden of proof
It is acknowledged by most merger authorities that the merging firms have the burden
of proof regarding the type, likelihood and magnitude of efficiencies, as well as the
merger-specificity of the claimed efficiencies. The merging firms may be presumed to have more information about their businesses, especially about efficiencies if that was
an important factor in deciding upon the merger.
4.3.3.2 Standard of proof
Nevertheless, even if the merging firms have the burden of proof, informational
problems remain a key problem, because of the asymmetry of information between the
merging firms and the merger authorities.
4.3.3.3 The partial efficiency defence
Not all efficiencies are equally easy to verify. For this reason, it would be desirable to
distinguish between alternative types of efficiencies not only based on their expected
effects, but also based on their verifiability. For example, Areeda and Turner would
especially favour efficiencies relating to economies of scale, at least if demand is not
growing rapidly.
4.3.3.4 Verification and certification of information
To alleviate the asymmetric information problem between the merging firms and the
merger authorities, it may be desirable to put different weight on the amount on
efficiency claims, depending on the source that certifies the validity of the information.
The Canadian merger guidelines109 distinguish between the following sources of
information: (1) plant- and firm-level accounting statements, (2) internal studies, (3)
strategic plans, (4) capital appropriation requests, (5) management consulting studies
(where available), or (6) other available data. There may be an important advantage of
information that is certified by outsiders such as management consultants, since they
need to protect a reputation of future reliability.
It may also be argued that less weight should be assigned to studies on
efficiencies that have been prepared after the merger has been taken to a detailed
investigation by the competition agency. This is because this would reveal that they
were not the ostensible basis for the merger decision.
4.3.3.5 Post-merger review
Scherer (1991) has proposed to first approve mergers only on a temporary basis.
Mergers for which persuasive preliminary evidence on expected efficiencies can be
provided may be allowed for a temporary trial period of, for example, three years. At
the end of this period, an evaluation can be conducted to find out whether the
promised efficiencies have, in fact, been realised. If so, the merger could be definitely
accepted. Otherwise, the company may be required to divest again into the two
independent parties that existed before the merger.
4.3.3.6 Merger licence fees
Direct verification, certification, or post-merger review all have in common that they
require an direct evaluation (either ex ante or ex post) of the actual efficiencies that are
(expected to be) realised through the merger. A drastically different approach in
screening mergers would be to implement a revelation mechanism through the
institution of merger licence fees to be paid to the government. It is not clear how the licence fee for the merger to be accepted should be determined. This will for example depend on the objective the merger authority has in mind (maximisation of consumer surplus and total welfare).
4.4.1.2 Political considerations
1-First, there has been the Kali-Salz decision, which the Court for the first time
accepted the concept of joint, or oligopolistic, dominance, rather than single firm
dominance. This suggests that the merger control policy has become stricter. This
development may make it more natural to consider efficiencies more explicitly today.
2-Second, many merging firms try to make use of efficiency arguments in
convincing the merger authorities on the likely net benefits from the merger. At
present, the Commission is simply not positioned to consider these efficiency claims in
a formal way. It would potentially improve the transparency of merger review that
efficiency considerations can be dealt using a clear procedure.
3-Third, there is a political argument why it may not be desirable to allow
efficiency considerations. One has to be careful that introducing efficiency
considerations in a formal way does not open up the possibility of a more active
industrial policy.