1

The economics of non-horizontal merger guidelines

Nikos Vettas

Professor, Athens University of Economics and Business, Greece andMember, European Advisory Group in Competition Policy (Chief Economist

Group, DG-Competition)Research Fellow, Centre for Economic Policy Research, London

Athens, June 2 2007

2

The issue: non-horizontal merger guidelines

Non-horizontal mergers - vertical - conglomerate

EU guidelines?

Compare to Horizontal merger guidelines

Compare to US non-horizontal guidelines

3

European Commission public consultation

Draft merger guidelines: circulated February 2007

Public consultation- started February 13, 2007- ended May 12, 2007

Commissioner Neelie Kroes: “…providing clear and predictable guidance for businesses…”

DG - Enterprise and Industry: “The Efficiency-Enhancing Effects of Non-Horizontal Mergers” (RBB Economics report)

4

EAGCP preliminary paper (advisory note)

◊ EAGCP (Economic Advisory Group in Competition Policy)

◊ Mergers subgroup Marc Ivaldi (University of Toulouse)Bruce Lyons (University of East Anglia)Monika Schnitzer (University of Munich)John Van Reenen (London School of Economics)Frank Verboven (University of Leuven)Nikos Vettas (Athens University of Economics and Business)Xavier Vives (IESE Business School, Barcelona)

◊ First set of (ten) general principles – Spring 2006First, on key economic differences between horizontal mergers and NHM. Second, on what the guidelines should achieve.

◊ Personal views

5

The presumption in a horizontal merger is that it directly leads to increased market power and an ability to raise prices

If the merging parties have market power, a horizontal merger will typically remove the competitive pricing restraints parties place on each other.

This is not the case for a non-horizontal merger. Horizontal mergers are among “substitutes”, non-horizontal mergers are among

“complements”. Enhancing efficiency is expected. Principle backed both by economic theory and some empirical evidence. The competitive effects of non-horizontal mergers will depend on characteristics of the situation

examined.

Important: NHM parties often have previous contractual relationships, in which case the competitive change following a merger will be muted. Parties to a horizontal merger typically do not have prior contracts.

1. The competitive impact of NHM is fundamentally different from that of HM.

6

A horizontal merger can harm competition by its direct impact on the incentive to raise price.

A non-horizontal merger has no direct impact on pricing. It may still cause competitive harm through a change in supply or procurement policy, or the way in

which a product range is offered to consumers. Then, cost or demand of rivals is affected and their pricing changes.

The merged firm sometimes has the incentive to soften price competition in the downstream market in order to protect its upstream profits. This situation may be difficult to track, since the downstream firm has no incentives to complain.

The effect of non-horizontal mergers on competitive conditions tends to be limited, since NHM parties often have previous contractual relationship, e.g. that limits buyer or supplier switching, whereas HM parties usually do not have prior contract.

Indirect effects of non-horizontal mergers can potentially impede effective competition, but they require a particularly careful analysis in order to justify a likelihood of harm.

2. NHM have indirect impact on competition.

7

There are “canonical models” for horizontal merger (Bertrand with differentiated products or Cournot with homogenous products and capacity constraints). Much more difficult to analyze non-horizontal mergers, there are many possible effects.

The Chicago School (“no harm”) models for NHM: specify conditions under which there is no loss of competition due to a non-horizontal merger.

Monopoly profits “can be taken only once” along a vertically linked chain Vertical integration can reduce distortions by eliminating “double marginalization”. Vertical integration often results to production and organization efficiencies. Note: the same reasoning applies to mergers between complements.

The Chicago approach relies on particular assumptions. In practice, the appropriate theory of competitive harm must be carefully “tuned” to the merger in question.

3. Non-horizontal mergers take many forms and produce a variety of effects.

8

This well-established fact provides an essential filter for screening out spurious concerts.

Market power is a necessary, but not sufficient condition for foreclosure.

When there is market power at both stages of the vertical chain, vertical integration helps to avoid double marginalization (that often exists in absence of sophisticated contracts) and thus results to efficiency gains.

Even in the presence of market power, merging firms need not have the incentive to take actions that would reduce the market available to rivals; and even if this is a likely consequence, it need not result in consumer harm.

4. Market power in an existing market is a pre-requisite for competitive harm from foreclosure.

9

The main vertical externality: double marginalization

U

D

consumers

10

Remarks

Double marginalization: with linear pricing, we obtain final prices that are too high for the final consumers and also for the sellers.Worse than VI monopoly.Distortion increasing with the number of stagesThis problem can be resolved with- vertical integration- nonlinear pricing: a two-part tariffWho sets the prices? Bargaining?Only problem if market power at both stages.

11

Adding horizontal strategic competition

consumers

U

D1 D2

12

Multiple downstream

consumers

U

D2 D3D1 Dn...

13

Common Agency / bottlenecks

consumers

U1 U2

D

14

Other structures

consumers

U1 U2

D1 D2

15

Modern theory of firm and empirical evidence inform us that firm as an institution exists due to incomplete contracts and transaction costs, which make it more efficient to carry out certain activities coordinated within a firm as opposed to through markets.

If complete contracts could be written (at low cost), the need to organize economic activities within firms would be limited. However, contracts that specify a precise outcome for each possible future contingency are too expensive.

The efficient range of activities carried out within a firm may change over time. A non-horizontal merger may be efficiency-enhancing.

The downside of a decision error that prohibits a merger is likely to be much greater for NHM than for horizontal mergers.

5. Stronger efficiency arguments for non-horizontal mergers than for horizontal mergers.

16

Example: specific investments

Assurance of quantity, quality and time of supply

Specific production

Specific assets: danger of opportunistic behavior (hold-up) implies below optimal investment incentives (Coase, vertical boundaries of the firm; Williamson; Grossman-Hart)

Solution? VI? Long-term contracts? Short-term contracts?

Single sourcing (exclusivity) vs. multiple sources.

17

Major difficulties inherent in writing detailed guidelines. Difficulties based on the very nature of non-horizontal mergers and

the multiplicity of forms they may take. Difficulties are reflected on the current state of the academic

literature.

Having a suitable set of principles to deal with NHM is important. They will contribute towards predictability of decisions and

consistency. Predictability is important, especially to businesses and their

advisors. Case handlers need guidance. Good guidance will be useful for authorities beyond DG-

Competition.

6. NHM Guidelines could enhance the accuracy and predictability of decisions.

18

Efficiencies that lead to reduction in prices or to more attractive product offering may lead to reduction of competitors’ market share, but are to be welcomed as consumer benefit.

Elimination of double-marginalization can create an incentive to reduce customer prices.

Not always a double margin to be eliminated, even in the presence of imperfect competition, because parties have an incentive to contract around this source of inefficiency.

Pre-merger contracts are very relevant and important in this context. Depending on the particular type of market structure under

investigation studies, double marginalization does not necessarily reduce the profits of the firms involved.

7. Guidelines should focus on competitive effects resulting in benefit or harm to consumers (and not to competitors).

19

Particularly important to recognise the benefits of anticipated customer price reductions because competitors anticipating lost market share have a strong incentive to complain loudly to the Commission.

An expectation of consumer harm needs very careful support.

7. …… (and not to competitors).

20

For actions harmful to consumers, both the ability and the incentive of merged firms to take such actions should be established.

Each particular scenario of harm should rely on a package of pre-requisites, e.g. market share, product differentiation, feasibility of bundling, margins.

Other useful information to be considered: existing contracts, sales practices, procurement methods, natural experiments.

The assessment of the effects of a vertical merger must not presume consumer harm and then look for countervailing efficiencies. In other words, the model used to analyze a vertical merger must allow for potential efficiencies from the start and not for competitive harm only.

8. Guidelines should indicate the methodology of analysis and the relevant evidence.

21

Qualitative guidance may be provided on the relative risks of different types of harm (e.g. there is usually far greater concern if foreclosure results in exit, as opposed to changes in market share).

‘Safe harbours’ might be indicated, at least in relation to the pre-requisite of market power.

If none of the firms involved in a merger has significant market power in any of the markets, then there is almost always no need for a competition authority to scrutinize this particular merger.

If after a vertical merger the upstream market is supplied by vertically integrated firms only, there is a risk that upstream firms could sustain higher prices for the upstream products, forcing (not vertically integrated) downstream firms to raise their final prices. Hence the number of firms acting in the different markets plays a crucial role.

Overall, the scale of likelihood should tie appropriately with the required standard of proof for the Commission.

9. Guidelines should distinguish “more likely” from “less likely” competitive harm whenever possible.

22

The horizontal merger guidelines can be interpreted to define existing market power in the context of a non-horizontal merger case (even though those guidelines deal directly with changes to market power, not the existence of market power).

There might be some important links to Article 82, as firms may find ways to establish vertical relationships among themselves (e.g. long-term specific contracts) that possibly replicate part of what would be achieved via a merger.

There should also be consistency with notices on vertical restraints.

10. Non-horizontal Guidelines should be consistent with other Guidelines / Notices / Green Papers.

23

On the role of guidelines.

On the role of competition policy (in the general mix of economic policy).

Competition, free market! Rules to be clear and application of policy should be

transparent, consistent, efficient and predictable. Regulatory cost as an entry cost. Entry! Innovation!

Concluding remarks