Thursday, 26 February 2015

Competition Policy

Why intervene in markets

Market outcomes can be inefficient because of anti-competitive behaviour or due to
externalities. Information is an externality of growing importance, Economics would
regard unfairness as justification for alternative social policies. financial and
informational capability to secure superior outcome and addressing this problem is a
priority and will not lead to the neglect of other more urgent priorities or have negative
indirect effects.

Competition policy and fairness are closely connected in reality although economists
differ in theory. Neoclassicists assume no informational failures and argue that all
markets are always in perfect equilibrium, either a monopoly or competitive equilibrium.
Chicago economists have an economic and political argument against intervention
because they bevies that monopolies are always temporary, markets will break them in
time and fear both regulatory capture and the inefficiency of the state more than the
imperfections of the market. For Chicago transaction costs are real and cause friction,
with incentives more powerful at higher unit prices. A clever variant of Chicago theory is
contestability theory, even with only two firms in a market the freedom of entry an exit
can create market pressure. Austrians see monopoly as the incentive for capitalistic
entrepreneurialism and economies are constantly out of equilibrium. Information is never
perfect Behavioural evolutionary economists accept much of the economics of the
Austrians but allow for much greater room for error by consumers and that individuals
make consistent mistakes and can do so on the whole economy scale. For behaviourists
like Stieglitz, very small informational problems lead to enormous knock on effects and
natural selection may fail. For behaviouralists even if people have the correct information
they still make the wrong choices.

There are non economic goals in competition policy. Anti-trusts policy emerged in the US
when as part of the progressive era trusts were broken in order to discipline powerful
players and to reduce the concentration of economic and political power. US regulatory
policies are seen s aiming at fairness and not efficiencies until the Chicago schools.
German ordo-ilberalism saw the need to support the Mittelstand, this can be desirable to
break down concentrations of social and political power, somewhat justified by
behaviouralism where increased numbers of decision makers make it harder to make
bad choices.

Regardless of their politics, people agree that markets are not perfect but it is difficult to
agree how to correct market failures.

1
 Thursday, 26 February 2015

Introduction
Competition policy attempts to protect society from the effects of restriction competition -
but detrimental to whom, over what timescale and how? For example in the
pharmaceutical industry, developers pay generic firms to delay production, is this anti-
innovation?

Why do we need competition policy? Market failures obviously, firms anti-competietive

profit maximisation,

Where there are natural monopoly features which can lead to the abuse of monopoly
power. Where dominant positions persist due to sunk-costs, lock-in effects (microsoft),
network effects (Beta vs VHS).

Companies collude on prices, quantities etc. This is by definition secret, we only have
statistics about cartels that have been detected. Punishments can lead to bankruptcy
which further damages the markets and is paid by consumers. Some companies engage
in predatory pricing to eliminate competitors. There are also a range of other
exclusionary behaviours by companies.

In the EU competition policy is an important part of ensuring the completion of the

internal market.

There are two main forces behind competition - entry and rivalry. New entrants (or
simply the threat of them) challenge incumbents. Rivalry in price, quality, innovation also
drives competition.

From an economic point of view, we intent to maximise consumer surplus. Competition

policy seeks to drive prices toward cost because this ensures an efficient allocation of
resources. This is also accompanies by productive (incentives to reduce costs) and
dynamic (incentives to develop new products) efficiency.

Competition policy benefits consumers, businesses and the economy.

Market performs better for individual consumers (better quality price-ratio, more choice
and variety, more innovation over time, etc…). For businesses competition policy gives
commercial freedom. Although criticised by lower wages and more unemployment on
occasion, competition should benefit the economy as a whole: once-off effects of
increased competition (lower prices and higher output across range of markets), on-
going benefits from more rivalry (productivity, innovation, flexibility), overall policy import
(competition in one sector drives competition in others, etc...)

2
 Thursday, 26 February 2015
Regulation is justified by market failures but differs from competition policy. A telecoms
regulation may direct companies to invest in 3G or 4G but a competition policy authority
cannot. Regulators may dictate the market structure, for example dictating the number of
mobile companies or supermarkets. Competition policy is also ex-post (wiht the
exception of mergers which are normally ex ante) and occasional whilst regulation is
generally given ex-ante and is continuous. Regulation is fundamentally linked to
industries and sectors whilst competition policy looks across the whole economy.
Despite all this the competences of competition and regulatory authorise overlap in
many cases.

Why to study competition policy? Companies need to know what is lawful and unlawful
and need to avoid both bad press and criminal sanctions. The success of mergers and
exclusive contracts is key, a failure to achieve permission may leave the company
fragile. Even public statemnts must be filtered through this (Arcelor Mittal - ‘dominant in
all markets’). Compliance is also important witha culture of competition embedded in
europe.

Areas of intervention are many and varied.

Anti-trust includes both abuse of dominant position (note of course, not the dominant
position) and cartel policing. Dominant position in C21 is often linked to the technological
network effect where utility increases for users with the number of users, see Windows
and the winner takes all de facto monopoly and standardisation. In electronic industries
marginal cost are also so close to zero following development that standard cost policing
doesn't apply. Cartels, despite what anybody says, are always unlawful.

Mergers may only be unlawful is they are a significant impediment to effective

cooperation, but precisely what that means is unclear.

State aids can be crisis related or not, horizontal or sectoral and are prohibited with
many exceptions. Crisis state aid forms a significant portion of GDP, 3.4% of GDP in
recapitalisation and 1.6% in asset relief.

The pillars of competition policy began in the Treaty of Rome and via Amsterdam and
Lisbon are now embodied in Article 101 and 102 TFEU. Merger control was only
separately from the treaty articles in 1989, this regulation was itself replaced in 2004.
State aids depend on 107-109 TFEU. Broadly, although excluding state aids, the US has
similar, although much older legislation. American state aids are invisible and
unlegislated.

3
 Thursday, 26 February 2015
The rules in practice are difficult to establish and the entrepreneur is always suspect.
Higher prices mean abuse of position, lower prices mean dumping or predatory pricing
and the same price is collusion or cartel.

The CJEU has talked about different forms of competition in its judgements, normal
behaviour is difficult to understand.

The Harvard school argued that price must approach the marginal cost of production
and we need to promote market structures which maximise total welfare (consumer +
producer surplus). This is the origin of the Structure - Conduct - Performance model.
where market structure determines behaviour and hence performance, generally only
more firms in an atomistic market leads to efficiency. This is a deterministic model in
which competition authorities should increase the number of firms.

Structure - number of firms | size of firms | entry conditions

Conduct - pricing | trust
Performance - profits | market share

For the Chicago school, competition is a process of selection, concentration allows the
most efficient firms to survive. This model is less deterministic, allowing for feedback
effects of performance to conduct and conduct to structure. Competition may lead to the
disappearance of competition but this is transitory because monopoly prices attract
entry. Market power is a necessary counterpart to innovation, see patents and the entire
pharmaceutical industry. Consumers may even benefit from concentration because of
economies of scale. The Chicago school only supports intervention if there are evident
negative effects and only on the economic functions of competition. Competition
authorities should concentrate on efficiency.

Historically competition policy has been based on a legal approach, but is increasingly
economics led. Economics requires modelling based on certain assumptions, but law
deos not make clear behavioural assumptions. Law examines motive and intention but
economics considered solely outcomes and effects. Economics has only one criterion -
efficiency, but law has others.

Economics offers methods to measure or calculator the situation (market concentation,

demand elasticity, etc) and also provides plausible stories to explain market behaviour
(tacit or explicit collusion, or rational competition). Although economics is always an
abstraction is does also produce testable hypothesis.

How should policy makers use economics to formulate legal rules? They should promote
efficiency , be easy to apply and not costly or enforcers or firms. Good policy addresses
typical problems not every problem | Just as per se illegal practices are not always
grateful to consumers so per se legal practices are not always beneficial to consumers.

4
 Thursday, 26 February 2015
Pay for delay in pharmaceuticals, this raises costs but encourages investment and
innovation.

Policy makers must have rules to minimise decision errors, both false positives where
legitimate practices are condemned and vice vers ain false negatives. Policy makers
must also find the optimal sanction - companies must be genuinely deterred but the bill
mustn't be passed onto consumers. The level of repeat offending in cartels
demonstrates that the sanctions are not at the right level.

EU competition law is not clear about its goals, whether the aim is to protect competitors,
consumers or the general interest. Article 101 theoretically focuses on consumer welfare
but in practice case law leads us to less clear cut conclusions.

Welfare theories focus on the virtues of competition. A perfectly competitive market

corresponds to a Pareto equilibrium ︎(i.e. a situation were the economic welfare of one
agent cannot increase without deteriorating the welfare of another one) and that each
Pareto equilibrium can be achieved via a perfectly competitive market. The Pareto
criterion where an increase in welfare is only allowable if actual compensation takes
place. However the Kaldor-Hicks criterion (i.e. cost-benefit analysis) where the welfare
gain for an agent must be superior to the loss that the gain entails to another agent
whatever the latter is compensated or not.︎ Moreover, allocative efficiency is only part of
the picture; Productive efficiency (or technical efficiency) that is whether firms use their
resources in the most effective manner is also important to consider ︎ Idem for dynamic
efficiency.

In practice the rationale of competition policy depends on the country, priorities, and
period. In China foreign firms are targeted by the competition authorities in order to get
data which can be used to promote Chinese firms. To what extent should the
environment, employment and other non-economic considerations be taken into account
by policy makers. Competition policy also interacts with other economic and industrial
policy, for example, to what extent is R&D cooperation permissible? Trade policy can be
affected, anti-dumping measures are good for firms but bad for consumers. Competition
policy is affected by the wider economic environment such as the financial crisis, a
bailout is a competitive advantage.

All competition policy in Europe has a double objective of integrating markets and
protecting competition. All geographic discrimination and action to reduce cross order
exchanges are viewed withs suspicion or prohibited even where they might be welfare
improving. Competition policy is also meant to remove distortions from national policies.

5
 Thursday, 26 February 2015
In Europe Cion is both the investigator and the judge for competition cases, unlike in
America where criminal sanctions are possible. Investigation is conducted by officials of
DG Comp who draft the statement of objection (SO). The SO is then sent to Cioner and
an oral hearing is chaired by an official who does not belong to DG Comp. ︎ The decision
is adopted by majority vote by Cion as a whole.

The Commission has powers over:

Decisions (Art. 7 to 10)

-Acting on a complaint, on its own initiative or following a leniency application from one
of the participant to a cartel.
-Imposing behavioural or structural remedies.
-Ordering interim measures.
-Reopening proceedings

︎Powers of investigation (Art. 17 to 21)

-Requesting all necessary information.

-Inquiring into a particular sector (e.g. energy).
-Conducting inspections in any offices, and in homes but with the authorisation of a
judge.

Penalties (Art. 23 and 24)

-Up to 1% of the total turnover if information supplied is incorrect or misleading.

-Up to 10% of the total turnover in case of infringement. In practice this tends to be 2% /
3%.

National Competition Authorities have some powers to apply EC competition law and
obliged to do it where trade may be affected. There is a complex system of cooperation.
This was intended to increase the EU anti-trust capacity not to supplant Cion.

EU law + national law or EU law only.

No stricter national decisions in case of Art. 101.
Stricter national laws may be applied to sanction unilateral conduct (Art. 102)

A right to defines and to have an oral hearing exits and there are counter powers n the
CJEU, but in practice waiting 2 or 3 years for a decision is commercially unviable and in
practice one must comply with the judgement of Cion. Some cases are appealed
however, in the Lundbeck case took more than 10 years and now the firm found guilty of
pay-for-delay is appealing however as Cion was not clear in its criteria for making a
judgement.

The procedure and climate for competition policy actions have changed significantly and
are still evolving. Currently on the agenda:

6
 Thursday, 26 February 2015
-Encouraging private actions (are predominant in the U.S., 80-90% of total actions
brought against firms)
-State aids : how national responses to the crisis can comply with state aids rules ?,
State aids and climate change, State aids and public service.
-Sector inquiries (pharmaceuticals, energy, Internet, ...)

Merger Control
In Europe merger control is prospective and preventative. Given the cots of undoing
mergers it is though to be better to act beforehand, in Australia and New Zealand there
is an ex-post approach. The goal is to detect, a priori and within a fixed period the risks
of a dominant position.

Pre-notification to DG Comp is always wise but the process forty beings with a
notification and seeking clearance from Cion. Phase I lasts a month (+ 10 days if
remedies are necessary) and gives a clearance decision or launches a in depth
investigation in Phase II. This gives four months (+15 days if rememedies and 20 days
for stop-the-clock) for statements of objection etc and investigation. The Commission
may give Clearence (with or without remedies) or a Prohibition (many companies facing
prohibition withdraw their proposed mergers).

DG Comp’s actions are based on the expected effect on consumers and the market,
using their economic toolbox as a crystal ball. On average DG Comp deals with ~300
merger notifications each year. Out of all 5726 cases in (1989-2014) 5035 were cleared
in Phase I with no commitments. Only 24 prohibition decisions have ever been issued
(with 5 in 2001) and some of these have been overturned. Given that so much effort is
spent on cases where nothing is done, some have suggested moving to an export
system.

DG Comps’ control is selective as in theory the network of NCAs cooperates across

Europe, but there are problems. In the Pan Fish / Maritime Harvest merger there were
four NCAs involved, UK, Spain and Norway allowed without remedies whilst France
required obligations. In Eurotunnel / My Ferry Link the French concluded it was a
concentration whilst the UK did not. There are more than 30 cases where different NCAs
have concluded differently in cross border mergers.

DG Comp looks at cases with a sufficient size and European scope. DG Comp has
turnover thresholds but there is a strategic element of authority shopping, as in some

7
 Thursday, 26 February 2015
cases firms prefer DG Comp or a NCA. In Endesa / Gas Natural, a hostile buyout, Gas
Natural asked DG Comp to review the case to escape Spanish governmental pressure
to merge and build a Spanish gas leader. Although this resembled the EDP / GDP case
in Portugal, DG Comp refused to take the case in the end, even after appeal.

In Gencor / Lonhro the merger concerned South African firms whose turnover was
mostly outside Europe and not subject to the thresholds but DG Comp decided to review
it on the basis of the affect on the Internal Market. In General Electric / Honeywell Cion
reviewed it because of the Internal Market affect on Airbus. EU competition policy can
have a significant affect outside Europe, in the above cases South Africa and the USA
both complained. The referral doctrine allows referral up to Cion where the Internal
Market is affected or down from Cion where the NCA of the only affected country is more
appropriate.

The European merger control rules prevent the acquisition of control, even on the open
market with gun jumping regulations.

For example on 10 June 2009 the Commission fined Electrabel 20 million for premature
implementation of its acquisition of control over Compagnie Nationale du Rhône (‘CNR‘)︎
Electrabel had notified its acquisition of CNR on 26 March 2008, and the Commission
had approved this on 29 April 2008, but without deciding on the exact date on which
Electrabel had acquired control over CNR.

In June 2009 the Commission found that Electrabel had acquired de facto sole control
over CNR in December 2003 whilst it had acquired (slightly) less than 50 per cent of
CNR's shares, the wide dispersion of the remaining shares together with past
attendance rates at CNR's shareholders’ meetings were such that Electrabel in fact
enjoyed a stable majority. That finding was reinforced by the point that Electrabel was
the only industrial shareholder. ︎

In setting the amount of the fine, the Commission took into account : the seriousness of
the infringement, the fact that the standstill obligation is a cornerstone of EC merger
control and; the duration of the infringement.

The Commission also pointed out that Electrabel is a large company with experience of
EC merger procedure and should have known that the transaction would result in an
acquisition of control subject to the ECMR.

︎Electrabel has challenged the decision claiming, inter alia, that the Commission
confused failure to notify the transaction with premature implementation. The appeal is
currently pending.

8
 Thursday, 26 February 2015
_

There are classically four steps to merger analysis:

First step : The delineation of the relevant (product + geographic) market.

-To identify and define the boundaries of competition between firms, ie. to identify in a
systematic way the competitive constraints that the undertakings involved face.

Second step : The assessment of the importance of the dominant position.

-Assessing market share and concentration

Third step : The assessment of the impact of the merger on competition.

-Anti-competitive effects vs. efficiencies gains

Fourth step : The implementation of remedies.

-Structural and/or behavioural remedies

Step 1 - Delineating the product and the market.

This is a cornerstone of EU competition policy. Distinguishing relevant product markets

is difficult, what is really an equivalent product. Geographical markets are relatively easy
to distinguish.

The three pillars of relevant market definition are:

Demand side substitution

-Substitutability of products from the viewpoint of consumers

︎Supply side substitution

-Ability of producers to switch production to the good considered, away from other goods

︎Potential competition
-Taking into account the possible entry of companies which currently are not producing
the relevant good and/or which are not operating in the same geographic area

Demand side substitution can be examined by consumer preferences and behaviour.

Companies all do this and there are huge data reserves. NCAs also conduct market
tests to understand how products are used (bananas and year round fruits). Other
concerns are technical and physical aspects of products (chemicals, long distance
cement). Legal attributes (different types of oranges for juice and consumption).
Distribution channels matter. Also consumer perception (supermarket vs boutique
perfume with same composition).

9
 Thursday, 26 February 2015
The use and abuse of market power can be constrained by demand substitutability
(consumers substituting for cheer products) or supply substitutability (producers, absent
from the market dominated by the monopolist, may use their assets to produce the
demanded product if its price has increased (the new supply must be effective and
immediate)

In America they use the hypothetical monopoly test -︎ An iterative method to delineate the
market which stops when the increase in price is sustainable.

Product market definition can also depend on price elasticity of demand. How much
quantity of demand changes for each 1% change in price. Formula =

(DELTA quantity/AVERAGE quantity) / (DELTA price/AVERAGE price)

e >1 is elastic, <1 inelastic

Price elasticity of demand can be used to measure cross-substitution by consumers.

if e<0 - X and Y are complementary

i.e. if the price of TV sets increases, then the demand of video recorder should decrease

︎if e=0 - X and Y are independent

︎i.e. a variation of the price of X does not affect the demand of Y

︎if e > 0 - X and Y are substitutes

if e > 1 : X and Y are close substitutes (on the same market)

Examples can be context dependent, trains and airplanes are for example thought to be
substitutes of short distances (hardly ever the case I think actually). The tool isn’t
perfect, we can only use it as we work on our assumptions.

We also consider the bundle of indices include the nature of the need satisfied and
product differentiation. Distribution channels for example the difference between buying
Harry Potter in a bookshop and via France Loisir are satisfying the same need but are
not the same for competition purposes as they belong to different markets. Conversely
coffee and tea belong to the same market as they satisfy the same need even though
they are distinct. There are only two lipstick factories in Europe but consumers do not
believe that they are substitutes. The fight in many competition cases is about the size of
the relevant market, eg Parker / Reynolds pens vs gifts, settled eventually on pens for
gifts.

In Varta / Bosch the overlap product was car batteries and the competition authorities
distinguished between the markets for manufacturers and the retail market. Cion felt that

10
 Thursday, 26 February 2015
prices increases in the manufacturing market would not be sustainable with collective
bargaining power, however separately for individuals it would be possible to pass on
price increases. Therefore Cion saw two markets, separated by consumer and ability to
pass on producers.

In Unilever / Ortiz Miko the product in question was ice cream and the distinction was
between place of consumption, home consumption, restaurant/café consumption and
impulsive consumption.

To understand relevant product markets and to know where to deploy analysis requires
intuition and imagination.

In the absence of elasticities we can use price correlation analysis; if products are
substitutes changes in price should correlate over time. A correlation of 0 indicates no
statistical relationship and the relationship can vary from -1 to 1.

In Nestlé / Perrier there was a high degree of price correlation between still waters,
between sparkling waters and between sparkling waters and still waters. Between soft
drinks there was a very low price correlation, and between soft drinks and still and
sparkling waters. The market under concern therefore was the market for still and
sparkling waters - there was an exception for Italy and Germany however where much
more sparkling water is consumed.

Critical loss analysis demonstrates where a price increase becomes unprofitable. Its
seldom used by CAs because the data is hard to come by, especially in industries where
there are joint costs.

Price response analysis shows how markets have responded in terms of price in the
past. In the Proctor & Gamble / VP-Schickedanz Cion looked at the historical effect of
the introduction of Always into the German market which demonstrated inflation
expected price increases for tampons over 3 years but real terms decreases for sanitary
towels.

Diversion ratio analysis is based on survey data attempting to measure substitutability. It

loos to measure the diversion from product A-B and vice versa based on increases in the
prices of either of the two products. As with critical loss analysis finding the data is very
hard.

11
 Thursday, 26 February 2015
Stationary / non-stationary tests are a more sophistcated the of price correlation analysis
looking at whether the relative prices tend to return to a stable level of time and how
quickly the prices revert in the long run. For this you need daily or at least weekly prices.

Supply side substitution.

Do right and left shoes belong to the same relevant market, even if they are not a
substitute from the consumer standpoint? When demand doesn’t measure or define
relevant markets we can look at the supply side.

For Cion the key measure is the ability to switch production of a good or set of goods to
the product under consideration, quickly and without incurring significant costs (nor
being exposed to significant risk). In practice, this is seldom accepted: there are often
costs to switch production (such as buying new machines or adapting them) and the
process can be fairly long (establishing new distribution channels or buildings)

Potential competitors

The threat of entry by ‘potential’ (as opposed to actual) competitors may limit the ability
and/or incentives of one or several firms active on a market to exercise market power.

-A long lasting price increase would attract newcomers, who are not currently active or
simply operate on an adjacent market.
-This threat of entry is not captured in existing market data - a potential competitor has
by its very nature a market share equal to 0.

The extent of potential competition is highly disputed in many cases, such as the
Lundbeck appeal.

In geographic markets distance may change substitutability between products.

Transportation costs -︎ Hitachi/NEC for semiconductors:︎ EU market (same preferences,

low transportation costs).

Specificities linked to the distance - cement cannot be transported >50km

Legal barriers, cultural barriers, etc. - Promodes/Carrefour ︎national market (advertising,

supply)

Qualitative evidence also supports geographic area delineation

-Customer catchment areas

-Specificity of consumers (location)

12
 Thursday, 26 February 2015
-Technical constraints
-Country-specific regulation
-Effective reach of product supply
-Transportation costs
-Linguistic barriers and cultural specificities

Defining markets is the achilles heel of competition policy. In a dynamic setting the USA
attempts to define innovation markets, potential future changes to relevant markets. DG
Comp accepts that this might be possible in pharmaceuticals but doesn’t engage in this
controversial analysis.

Step 2 - Assessing competition

Does a merger change competition conditions? They can potentially eliminate

competitors, reinforce merged entities or potentially encourage anti-competetive
behaviour changes.

Competition authorities examine market share based on the idea that the larger the
number of independent firms operating after the merger, the less likely it is that the
merger will be detrimental to consumers.

So once the relevant market has been delineated, need to measure market shares (level
+ trend), i.e. to measure the concentration before and after the merger + market
conditions (i.e. growth rate).

If the combined market share is > 50%︎may in itself be evidence of a dominant position.

If the market share is > 25% ︎ enquiry.

If the market share is < 25% ︎ clearance (de minimis rule : market power is not
substantial/no dominance) but exceptions always possible

︎Productive capacities (i.e. if rivals can increase production after the merger, it is likely
that the price will raise). But the Commission and the CFI have recognised for a long
time that a dominant position cannot be only inferred from market share. Also requires
an analysis of :

Entry: Necessity to evaluate existing barriers to entry in the industry (i.e. sunk costs, ...)

The firms’ ability to raise price is limited by the existence of potential entrants

13
 Thursday, 26 February 2015
Whether entry can achieve significant market impact within a timely period
If entry would be profitable (i.e. a likely response to a merger)
If entry would be sufficient to return market price to their pre-merger levels

Difficult assessment: need to evaluate if firms are likely to enter and when (uncertainty)

︎Other important factors:

Demand variables (e.g. price elasticity)
Buyer power (can go in both directions) i.e. competitive pressures by customers in
switching to other suppliers, in retaliating the supplier on other markets and in
sponsoring a potential entrant (! often over-estimated)
Failing firm defence

Related to the Harvard / Chicago distinction, there are two views of the relation between
concentration and profit - collusion and differential efficiency.

The collusion hypothesis (based on oligopoly theory)

-The lower the number of firms, the lower the intensity of competition, hence the higher
the profit.
-Arguments : Cournot competition; the risk of cartel increases when the number of firms
decrease.
-Policy recommendation : concentrated industries must be broken up...

The differential efficiency hypothesis (H.Demsetz)

-Firms within an industry are heterogeneous and some are superior to others because of
better products or lower costs. These superior firms tend to dominate the market. A
strong correlation is observed between market share and a firms’ profit.
-Policy recommendation - such a break up will penalize superior firms and provide
disincentive to produce better products or at lower cost

CR4 is used to measure the weight of a leader on the market ad is a simple, low data
efficient tool. This tool doesn't five much information about size distribution however.

The Hirschman-Herfindhal index gives more information than the concentration ratio.

HHI=(100s )2 +(100s )2 +...+(100s )2

Competition auhorities use the HHI as an indicator of the safe harbours for mergers, the
concentration on markets and the increase in concentration. What is acceptable can be
highly sector specific - in the soft drink market there are only two players but there is
substantial competition.

14
 Thursday, 26 February 2015
_

Step 3 - Assessing the effects of the merger

What effect will mergers have?

Hroztonal - overlapping in a product or geographic markets, with risks of collusion

Vertical mergers -

Conglomerate -

Unilateral effects - a merger may reduce competition even if it does not lead to single
firm dominance or collective dominance (i.e. it does not cause coordinated effects).

-tacit collusion (each firm raises prices in the expectation that others will also).

-uni/multilateral effects (each firm sets price as best response to others in the absence of
coordination).

Possible price increase by the merged entity, with respect to prices that would be
charged by the parties absent the merger. Current price levels and future price
expectations (“counterfactual”)

Indices

-Change in market concentration

-Elimination of the most direct competitor?
-Elimination of competitive interactions between the parties

More analytical approaches

-Switching analysis – Shock analysis

-Merger simulation

Coordinated effects - Risk that in the new industry configuration, the merged entity could
coordinate with its “competitors” (to a larger extent than pre-merger).

Test : the “Airtours” criteria :

-Coordination between firms easier/more likely.

-Triptyque: Monitoring (transparency) / Retaliation/ Lack of countervailing power

Transparency

-Ability of a firm to monitor the behavior of the other firms (prices, quantities sold,
installed capacity, R&D)– at least that of the main firms

15
 Thursday, 26 February 2015
-Ability to retaliate.
-Ability of the participants to the coordination to “punish” a firm that would deviate from
the terms of the coordination (e.g. through targeted price war)

-Lack of countervailing factors

-Ability of the external firms to compromise the co-ordination – Buyer power

Non-horizontal mergers are concerned by the foreclosure effect. CAs must look at the
ability, the incentives and the effect s on consumers and competitors of foreclosure.

There are two main scenarios for foreclosure, Input and customer.

Input - Raising barriers to entry vis-à-vis potential entrants

-downstream rival(s) of the merged entity cannot source inputs.
-If entry to take place both upstream and downstream, « double hurdle »

Input - Raising downstream rivals’ costs

-Material effect if excluded firm plays an important role in downstream competitive
process, e.g. large market share, aggressive strategy, closest competitor to merged
entity

Customer foreclosure - Raising rivals’ costs

-Increase in the cost of downstream rivals that purchase from U2, given this latter is
excluded or marginalized at upstream level.
-Longer term impact if there are negative effects over investments, productivity gains or
R&D at upstream level.
-Negative effects if significant part of (upstream) input production is concerned by
foreclosure effect.
-Negative effects if rivals are capacity constrained or if products are differentiated

Customer foreclosure - Increase in barriers to entry

-Upstream rival(s) of the merged entity cannot access key customers.
-Hurdle to supply to new entrants.
-If entry has to take place at both upstream and downstream levels (« double hurdle »
theory)

There are countervailing effects for both of these.

Vertical mergers and coordinated effects.

Ability to coordinate effectively

-Decrease in the number of suppliers if vertical merger results in exclusion.

16
 Thursday, 26 February 2015
-Increase in the degree of symmetry
-Elimination of a « maverick »

Transparency and monitoring

-Access to sensitive information about competitors
-Reduction in the number of rivals.

Retaliations
-Merged entity can retaliate more easily/effectively when rivals are key consumers and/
or suppliers.

Firms merge to improve the efficiency of production (allocative, productive or dynamic)

via business synergies.- rationalising and specialising production, introducing economies
of scaled scope, enhancing innovation capabilities, increasing bargaining power.

Assessing these in reality Cion uses three cumulative criteria.

Synergies must be quantifiable and verifiable - the burden of proof lies on the firms
themselves.

Synergies must be specific to the merger - only available through merging rather than
achievable by other means.

Synergies must also benefit consumers - the efficiency gains must offer direct
improvements for consumers.

Merger efficiencies may lead to simultaneous price decreases and market share
increases. This is the trade off between long term benefits and decreases to competition.

CAs have to consider that costs they are really considering, variable costs or fixed, as
only variable costs are really likely to lead to savings.

Efficiency gains may lead to the elimination of double marginalisation (two monopolies,
one upstream and one downstream) - this leads to coordination between suppliers and
distributers and eventually to cost reduction.

CAs also look at the protection of specific investments - brands and intellectual property
rights.

Mergers can also simply be ways to avoid hold-up or other factors.

17
 Thursday, 26 February 2015
The ‘Williamson trade off’ describes the balance between efficiency gains and anti-
competitive effects.

Efficiency gains reduce average costs - an efficiency effect, but concentration leads to
an increase in prices - a counterpoised market power effect.

These create a dead-weight loss for consumers and an increase in consumer surplus.

CAs have to balance welfare against efficiency, where prices increase but efficiency
gains grow, mergers are often allowed, but with remedies required.

EXAMPLE - Aerospatiale-Alenia / de Havilland

first - community dimension - yes on turnover

second - determine relevant market - firms - 2 based on regulation or second steward

over 50 seats | Cion - 3 with an intermediate market 40 - 59 seats, looked at direct costs
and regulatory constraints

Third - market share (taking Cion definition) - here they would have a dominant positon
in the middle market and be the only firm present in all three markets.

Fourth - effects of the merger - economies of scope for producers, cost economies for
airlines.

Fifth - assess market - a cyclical market, for next 4 years 200 aircraft pa, after ’95 sell 70
pa.

Sixth - potential competitors - not the large aircraft producers but the producers of small
jets - R&D costs are very high and a change of motorisation takes about 5 years but this
would mean entering the market at the trough of the market - Cion concludes that jet
producers are not real potential competitors and therefore no pressure on price.

Seventh - determine demand - demand comes from incumbent airlines and companies
in the process of entering the market - Companies entering the market will not be
financially equipped and will lease, this will give them wider choice. The most popular
leasing aircraft are made by Aerospatiale and de Havilland has a stake in this. Whilst
there is a priori choice, there is not a real choice - Companies already on the market
almost always buy the aircraft they already have or training and parts interoperability and
to diminish switching costs.

After losing their battle on the definition of the market the parties decided to
acknowledge all Cion’s points.

18
 Thursday, 26 February 2015
-barriers to entry are high (high sunk costs especially)
-demand is captive (difficult for airlines to change)
-weak actual substitutability

The parties were using the contestable market theory to demonstrate that the market
doesn't have a sustainable configuration. Most of the manufacturers in the aircraft
industry receive subsidies, they are not profitable in and of themselves. The market
doesn’t play its true role in delivering selection. We have a problem with the market. The
parties were arguing that aircraft are an intermediate market and that demand is
derivative of the final market in transportation services. Consumers are argued to have
power because of the strong ability to delay purchases and a high bankruptcy rate
leading to a strong lemon market. The parties argued that a reduction in the number of
suppliers was necessary to make the parties profitable. The same theory gave
dissentingly different conclusions.

Step 4 - Remedies

Remedies are intended both to restore the status quo ante and to eliminate concerns
surrounding competition. This is a twofold reading in the CA’s crystal ball.

Remedies are proposed by the parties and the burden of proof rests with them.

Remedies come in two, normally mixed forms.

-Structural (divestitures) : subsidiaries, production assets, brands, shareholdings.
-Behavioural: conditions of access to essential production factors, license agreements.

19
 Thursday, 26 February 2015
Cion assesses whether or not the remedies will restore competition, including whether
the implementation is likely to be successful, full and timely. Some delay can be possible
through trustees but divesting assets is difficult.

Structural remedies

Types of divestitures:
-Capacities or assets : Total Fina/Elf, Carrefour/Promodes.
-Brands : Unilever/Bestfoods.

Problems with divestitures:

-Assymmetric information.
-Possibility for the merging parties to divest unprofitable assets or decrease the value of
the divest assets.
-Identifying the right buyer.
-Risks of collusion (more symmetric situations, multi-market contacts...)

Behavioural remedies

Types of remedies:
-Licensing agreements: Astra/Zenaca.
-Non-discriminatory access to « essential facilities » : Vivendi/Canal+/Seagram
-Giving up important exclusivity agreements : Lufthansa/SAS
-Chinesewalls

Problems with behavioural remedies:

-Require on-going regulation or constant monitoring.
-Easier to « evade » (asymmetric information between firms and the authorities).
-Might require a (transitory) period of collaboration between the merging parties and the
entrant/licensee (risk of collusion, incentives not to collaborate effectively,...)

Cion and other CAs have a natural preference for structural remedies, which is
embedded and explicit in their operating procedure. Compared to behavioural remedies,
divestiture is more effective because it suppresses the cause of the pb. and is less costly
to implement because merging firms are incentivised to adopt socially desirable
behaviour. Unfortunately they don't always work and divestitures are not always
possible, particularly where there are intangible, joint or complementary assets.

Increasingly Cion works with other CAs, particularly USA and Japan to coordinate
remedies for global countries, this has obvious challenges given different procedures
and timetables.

20
 Thursday, 26 February 2015
Damages

The commission is not infallible, there are many examples of where Cion has failed in
analysis or due process. In Cion vs Schneider Legrand (2007) the CFI offered damaged
to Schneider for the first time. The CJEU however set aside the damages, requiring Cion
only to pay Schineider’s costs.

In My Travel Group vs Cion, the CFI annulled the prohibition decision but still didn’t find
the errors ‘manifest and grave’ with in the meaning of the damages test.

Mergers and the Crisis

SInce 2008 there has been reduction in the number of cases nut not in their complexity.
Mergers are more defensive and less driven by financial investment but rather are
horizontal concentration working for industrial consolidation.

Do CAs have to take into account the crisis, how should they consider rescue mergers
and the filing firms defence.

Mergers conclusion

From a historical point of view merger control in Europe has improved, nevertheless the
need to be careful remains, and the need to expand the number of tests.

‘A lawyer who has not studied economics is very apt to become a public enemy’ - Justice
Brandeis but also vice versa.

Cartel deterrence and punishment

« Our competitors are our friends; our customers are our enemies » OECD

This quote from a n OECD survey sums up the philosophy of cartels.

Cartes are based on agreements but most agreements in the business world are kawfyk,
they share risk, raise economies of scale, share know-how and allow products to be
launched more rapidly.

Unlawful agreements are focused on raising prices and increasing profits. Collusion in
these cases is tacit or explicit. Explicit collusion involves actual direct communication
whilst tacit collusion more often involves market based communication.

Was Adam Smith right - ‘People of the same trade seldom meet together, even for
merriment and diversion, but the conversation ends in conspiracy against the public’?

21
 Thursday, 26 February 2015
How in a world of more and more agreements may CAs distinguish between lawful and
unlawful agreements.

Profile of Cartels

Forming a cartel allows a firms to reach collective monopoly prices and thus monopoly
profits.

Why do firms chose not to form cartels if it is economically rational?

-most cartels form in chemistry, industrial construction, maritime transportation, never

detected in the car industry
-can we derive that various conditions promote the formation of cartels.

Bearing in mind that a statistical weight doe s not constitute necessary or sufficient
conditions for cauterisation we an design a probabilistic average cartel. The average
cartel allows resources to be concentrated.

Conditions likely to lead to Cartels

Collusion is likely in markets with a low number of suppliers - simply the ease of
reaching agreements is much higher. In some cartels however a professional
association is the investigator or conductor of collusion.

Low price elasticity, with quantity not greatly affected by price changes makes price
raising collusion easier.

Intermediary products (added vitamins for example) are easier to collude on as the
source of price increases is harder to identify.

Cyclical markets, particularly with high fixed costs, also encourage collusion as it allows
firms to smooth out price fluctuations.

Barriers to entry encourage collusion as it prevents new firms entering the market and
breaking one’s cartel.

Homogeneity of firms makes price collusion easier as similarity of production cost and
product similarity facilitates convergence. This is more likely in mature industries rather
than new, highly innovative and unstable industries.

How to organise a cartel

There are many ways to organise a cartel in practice:

22
 Thursday, 26 February 2015
-Common price determination
-In Brazil the government privatised an electricity company, valued between USD 1.78bn
and USD 3bn. The government set the reserve price at USD 1.78bn. After a beauty
contest, on 15 April 1998, three companies (ENROM, a Consortium and UBC (who didn’t
really have the means to bid, but were present politically to represent Brazil)) were to bid
in the auction. On 14 April the bidders were staying in the same hotel, the Consortium
proposed to ENRON that they would not bid and would get a sub-contract instead. To
avoid corruption allegations the Brazilians closed the stock market and broadcast the
auction. The Consortium bid the reserve price. - The outcome was a loss for the
Brazilian government, a net gain for ENRON who weren’t; really interested but only
wanted to bid up the price. - What went wrong? Brazil was trying to be transparent after
years of dictatorship and too much transparency allowed the Consortium to see how
ENRON was bidding on the day. - Reduced uncertainty enabled collusion. - This
came to light 5 years and 1 month later, sadly 1 month after the end of the competition
policy deadline. In 1999 however there was another privatisation with the same
companies which the Consortium also won, if this was the same behaviour both
privatisations would be subject to penalties as continuing collusion. - The moral of the
tale is NEVER SPEAK UP.
-Price tables can be set up, for example lawyers in America set up price tables for
divorces. The CA said that prices had to be based on costs.
- Alternatively companies may use a bass points system, for example where
transportation costs differ, even with a collusive base price the final price of the product
will differ. This would make it very difficult for the remainder of the cartel to check on the
compliance with the agreement by their partners, it is therefore likely that they would
also calculate a centralised basis point system of transportation costs to compare for
power collusion here too.

-Geographic sharing of the market

-Geographic sharing, a typical phenomenon in the cement market gives each company
their own territory. This was also found in the Scandinavian air route industry.

-Customer division
-Some cartels have divided the customer base between them. An example occurred in
Luxembourg with the HORECA cartel where each supplier agreed only to supply HOtels,
REstaurants, or CAfés - HO RE CA. This has been seen in beer i the UK too.

-Boycott
-In Lyon, Optical Centre gave discounts to several chain opticians in the city. The
independent opticians decided to boycott Optical Centre to encourage the to change
their business strategy - however boycotts are not allowed under competition law.

-Production quotas
-A production quota is equivalent to a production monopoly. Quotas will be divided based

23
 Thursday, 26 February 2015
on production function; in real life different companies have different efficiencies and
therefore higher production functions and lower marginal costs. Rationally only the most
efficient firms should be allowed to produce and the company with the highest marginal
cost shouldn’t produce at all and should in fact be paid not to produce by the more
efficient firms. In reality this can happen with lateral payments.

-Most-favoured customer clauses

-Agreements like the John Lewis price guarantee encourage collusion because it uses
the consumer to check up on whether the collusive agreement is respected or not. Akaï
and Fnac were found to be doing this and it is also common in information technology.

Stability of cartels

Cartels are by nature secret, so our statistics only include those which are detected.
However, from these we know that many disappear naturally because of cheating within
the cartel. On average half last loner than 5 years, and 12% - 17% last more than 10
years. The Variance of life span is huge, some are ‘born dead’ lasting weeks whist
others lasted for more than 30 years

External factors leading tot he collapse of a cartel include shocks, new firm entering the
market and antitrust units. External shocks can include new technology and other
developments, new firm can break into a market and undercut a cartel, antitrust work
has particularly effective since the introduction of immunity for cartel breakers changed
incentives.

Internal factors argue that once firms decide to collude, the best strategy for each of
them is to cheat and to secretly decrease its price (Stigler, 1968). If Stigler is true, then
there is no use to fight cartels as they destroy themselves.

Game theory and the prisoners dilemma show the conflict between collective interest
and individual gain. This tells us that with a finite number of periods of interaction it is
always better to cheat than to respect the agreement. Where there are an infinite
number of periods trigger strategies encourage maintaining the collusive price (Pm) until
one party cheats and establishes a deviation price (Pd).

T T+1 T+n

Firm A Pd Pd Pc

Firm B Pm Pc Pc

The instant profit of deviation is higher than the instant profit of collusion, πd > πm/c

The decision to obey or not is based on the discount factor which accounts for the cost
of time. If the future is depreciating this incentivises cheating now, whereas if the future

24
 Thursday, 26 February 2015
will improve companies are likely to obey their cartel agreements. Choosing partners for
cartels carefully is important

Factos favourable to stability include:

-low number of firms - cheating is easier to detect and trust easier to build
-high frequency of transaction - difficult to punish people when nuclear power plants are
only bought every five years.
-market transparency - transparency, like the in the Brazilian auction system, allows
people to detect cheating.
-market with high growth - collusion never occurs in R&D because uncertainty is too
high. Mature, stable markets have lower depreciation of future value and encourage
stable collusion. In unstable markets the incentives are to make s much profit as
possible now before the market collapses.
-cost symmetry and production capacity symmetry - this allows each member of the
cartel to punish the others.
-multi-market contacts - multi market contacts allows for punishment on all markets,
even where there was no deviation or collusion. The ‘mtlatiory effect’.
-information exchange - without information exchange there is no trust and no way to
check if people are respecting the collusion agreement.
-financial stability - companies close to bankruptcy will always cheat, whereas stable
firms are likely to profit from the future.
_

Why prohibit cartels?

Cartels are genuinely very costly - a citric acid cartel in the US cost consumers >USD
100m, price increases of 90% are known. Merger investigations look at a 5% price
increase but cartels normally start at 20% and go up.

Cartels also involve a welfare loss.

-m firms
-MC = AC
-f. demand P=500-Q
-MC = 100 therefore P = 100, therefore Q = 400

-consumer surplus = (A-p*) x Q*/2 = (500 - 100) x 400 / 2 = 80,000

-profit = 0 (MC = AC), therefore producer surplus is 0 and total welfare = 80,000 + 0

In a cartel:
-Pm = 300
-Qm = 200

-consumer surplus = (A-p*)x Q*/2 = (500-300)x 200 / 2 = 20,000

25
 Thursday, 26 February 2015
-profits from collusion or producer surplus = (Pm - MC) x Qm = (300 - 100) x 200 =
40,000

-total welfare = 20,000 + 40,000 = 60,000

Cartels involve we see a transfer of welfare from consumers to producers.

Fighting Cartels

A priority for DG Comp but a recent move, in 1998 DG Comp established its first unit and
today there are 5 + a dedicated settlement team. In Europe there were many legal
cartels for years and culturally the fight against the cartels is relatively new in
comparison to the USA. Only in the UK and Ireland, with Common Law, can cartels
result in incarceration and criminal sanctions.

Fines are rarely upheld by the courts, serving as a weak deterrent. Cion never takes its
full 10% of turnover and in 60% of cases the fine was below 1% of cases, more than 9%
has only been fined in 24 cases. These will be even lower once adjusted for court
judgements. Many firms are repeat offenders, St Gobain for example have been caught
at least 4 times. The trade off between deterrent fines and the need protect consumers
from carrying the burden of fines or the risk of forcing firms into bankruptcy is unclear. In
roman law countries the culture is different and people face no criminal sanction or
banning from corporate positions. Cion is however pushing to implement class actions
and criminal sanctions.

Currently there are incentives to continue with collusion given that fines are affordable.
Some event studies analysis has suggested that fines have knock-on effect on share
prices

Cion generally falls on the problem of giving proof beyond reasonable doubt. Prices are
never stable within the EU and equal prices do not necessarily demonstrate collusion
rather than good competition particularly that where the possibility of pro-competitive
interpretations exits they must be accepted. Authorities have to infer collusion from hard
evidence which is can be difficult to reach the criminal standard of proof.

Fines, when given, are set based on the gravity of the infringement taking into account
its nature the size of the market, the actual impact and the duration. Aggravating
circumstances include a refusal to cooperate, previous collusions etc. whilst attenuating
circumstances include a passive role.

Private enforcement

26
 Thursday, 26 February 2015
increasingly private anti-trust actions are both possible and pursued. These are not only
against collusion but in a recent case in the Brussels commercial court, against the
judgements of Cion where their evidence was insufficient.

Summary

Whilst competition policy is broadly in line with economic thinking now there is still more
to be done on enforcement. It is unclear that fine are large enough to act as deterrents,
and quantifying harm is still difficult. We need a coherent european approach to
collective redress

Abuse of dominant position

Cartels are much more likely to be prosecuted than abuse of dominant position, not
because it is less serious but because it is much harder to prove. Bringing cases which
they lose damages the political support for competition agencies and weakens the
deterrence effect.

When is a firm a monopoly and when is it a dominant position? Total control, >50% of a
market? Classically monopolies have inelastic demand. However demand in reality is
always elastic as inefficient substitutes will emerge at some level. Competition
authorities no longer look at high prices and few firm but at the practices monopolies use
to preserve their position, particularly erecting barriers to entry.

The European and American authorities differ in their approaches - in America

dominance and the acquisition of it is illegal whilst in Europe it is only the abuse of
dominance.

Methodology

Necessity to assess the dominant position before inferring a possible abuse ︎ Same
analysis as for mergers but the approach is retrospective and not prospective. The idea
is to reconstitute the past and see what it would have been without the possible abuse.︎

1 - Does a dominant position. Within a market︎ this step requires to define the relevant
market.

2 - Whether the firm which enjoys a dominant position is abusing that dominant position
through: Sacrifice,︎ Intended or effective exclusion, or︎ Recoupment.

27
 Thursday, 26 February 2015
Handling these issues has moved under Article 102 TFEU from a form based approach
to an effect based approach. Efficiency and the social interest are key to deciding if
practices actually damage consumers interest.

The Chicago school argues that If there is enough profit entry will always be possible
and new challengers to the market will always exist.

Competition authorise have the hard challenge of always producing counterfactuals

which they can demonstrate are superior and that firms have actively worked to prevent
them. Per se rules lure one away from the challenges of this, but are incomplete, not
covering all potentially damaging practices and not assessing welfare.

The abuse of position is mainly about exclusionary practice (ie the aim of deterring entry
or forcing exit of rivals). If this were not the case other would be able to join the industry
and join in the excessive profits. This is a more Austrian perspective, a high profit margin
in not a sing of wrongdoing in and of itself.

Price cutting for example is designed to make entry into the market unprofitable. In
reality however this is not sustainable. The dixit model describes predatory behaviour
where no firm can credibly deter other firms by over producing and cutting prices but
installing large capacity can deter other firms by not elevating the marginal cost but
creating a fixed cost for competitors to overcome. The Chicago view is that entry and
exit are naturally easy but the contestability model introduced thinking about what
portion of the fixed costs of entry were sunk costs which cannot be recovered. So long
as there was an incentive to leave they cannot keep other firms out of the market but if
sunk costs are high firms cannot retreat from a market.

Unnecessary similar products can be produced to flood the market (cornflakes) or

patents can be raced for to control the development of replacement technologies (OPEC
and biofuels). Predatory pricing takes many forms and may not may bot be economically
inefficient (Microsoft high priced site licences and marginal cost software).

Only two case of predatory pricing have ever been brought by Cion, Tetrapak and Hilti In
Tetrapak the high margins in one area allowed cross subsidising in another area. In
practice and on studies of American cases, predatory pricing is almost impossible to
prove. Given the challenge of proving is Chicago would say its unwise to interfere and
Austrians that competition is good and always exists, so what if somebody cuts prices
below marginal cost.

Vertical restraints
Vertical restraints are restrictive contracts between suppliers and consumer at
successive stages in the chain of production. Typically they fix the conditions at which
the product may be sold on. Cion has taken a strong line that nether vertical restraints

28
 Thursday, 26 February 2015
nor IPR can be used to fragment the European market. There has been a presumption
in economics that vertical agreements are not harmful to efficiency, particularly in the
Chicago school. The EU has built a strict system of rules to combat vertical restraints
based on principles not advanced economics and which are aimed at integrating the
common marketing date all the way back to the 1950s. The Americans, influenced by the
Chicago gang, believe that if there is competition at any horizontal level vertical
restraints cannot restrain efficiency as new and less onerous suppliers will emerge to
replace vertical retainers (in the absence of monopoly power and barriers to entry).

The Americans complain that the EU rules make it impossible to have exclusive
dealership even when this would be economically efficient. Cion was not motivated in
the 60s by economics however but by a porto-single market policy. Exclusive
arrangements were thought to pose a risk of dumping where one firm could sell more
cheaply in one market than another. In the 1960s the CoJ was so opposed to anti-
dumping that Cion was supported in fighting vertical restraint (under-dumping legislation
any form of price discrimination is illegal, even with no competitive effect). Cion also
acted to prevent IPR being used to segment the internal market.

Only in the car industry were all these regulations ignored. With political pressure to
protect national car industries and the industry claiming vertical restraints provide a
benefit to consumer service no harmonised industrial policy was needed. The car
industry structured itself with sole and exclusive marketing, this guaranteed minimum
retail prices and localised dealerships and voided guarantees if consumers disobeyed.
Firms claimed that this guaranteed good service quality and covered marketing and
advertising costs. Competition policy is to just the application of the law, there are always
other interests at play.

At the time of the eastern enlargement Cion argued that multinationals should not be
allowed to charge higher or lower prices in new member states to cover marketing costs
or to boots market share without allowing parallel imports. In the ‘Silhouette’ case
spectacle frames were much chirp in Bulgaria than Europe, under pre-accession
agreements Bulgaria had to apply competition law but not IPR rules.

Vertical restraints can be economically efficient however - the non selling on of site
licences issued by Microsoft is a vertical restraint but one which allows Microsoft to
license copies of software at zero marginal cost.

Cion is attempting to move toward having a less per se rule based system of regulating
price discrimination and toward a more economically reasoned assessment of vertical
restraints. Part of this was the existential crisis for Cion in 1992 with the completion of
the single market and the move from creating the Single Market to acting at the
competition authority.

29
 Thursday, 26 February 2015

IPR, R&D, Patents

Patents are designed to give a temporary monopoly offsetting the anti-competitive
effects of this by promoting innovation. More and more industrial economics has
suggested that patents can be as much a restrictor of innovation as a promotor of it.
Sequential innovation means that new innovators do not take out patents on new
products but tiny incremental improvements to complex products.

In the US an obsession with patentability led to extreme and anti-innovation IPR from the
1980s. The EU did not go so far down the same rabbit hole which slightly reflects the
difference where IPR in the US is constitutional (useful arts) and anti-trust is statutory,
whilst competition law has been in the Treaties since Rome whilst IPR is subordinated to
it. R&D is a fixed cost and in competitive markets the incentives may not exist to spend
on innovation. Cion tries to encourage firs to cooperate to develop new products but not
to form cartels to market them (eg CD formats).

The linear model of innovation is unreliable tis is incremental by process rather than a
big bang expense upstream development. The extreme example is the Japanese kaizen
demand led system with most innovation taking place on the shop floor. Innovation is
mostly thought to be undertaken by firms in the pursuit of maintaining their market share.
When inefficient firms are subjected to competition they are likely to give up, more
efficient firms have more to lose and respond well to market pressure.

Patents, Trade Marks and Copyright are all distinct legal concepts with different aims
and effects. Patents protect the monopoly on the idea of doing something whilst
copyright is a monopoly on the form of expression of an idea.

Patents are designed to protect moral rights and to encourage the sharing of the
production of new products. The assumption is that patents allow you to set prices
above marginal cots and therefore to recoup the costs of R&D. They discourage
wasteful duplicate innovation and can help raise venture capital. They can however
create monopoly power, slow diffusion of new products and ideas, promote litigation and
can act as an obstacle to cumulative sequential innovation. The Patent race in particular
can be an anti-competitive hindrance to innovation.

In the 1980s it becomes clear that government promoted innovation was not working
and governments sought to make innovation more profitable by extending the scope and
domain of patenting.

For a long time the US FTC was sympathetic to all attempts by drug companies to use
their IPR. In 2012 the FTC and Cion both moved to prosecute firms (the Actavis case) to
forbid them from ‘pay for delay’ contracts.

Increasingly in the US Patent Thickets (the jumble of calmative patents produced by

incremental innovation) are putting US biotechnology at risk. One solution is patent

30
 Thursday, 26 February 2015
pooling where a group of firms cross licence all of each others patents; whilst these free
up innovation preventing anti-competitive hold up they act as significant barriers to entry.

In the EU the patent regime has led to less monopolistic abuses as Cion has fought
market segmentation. IPR regimes are still national and so far do not have a weaker
level than any other product rights.

Balancing competition law and patents is a challenges at all levels, but must to an extent
be ex post where it emerges that IPR has given a special or disproportionate treatment
with a negative outcome to IPR holders. When the value of patents is revealed in the
market the degree of the monopoly may need to be revisited.

Copyright infringement is complex, the first CD playing computers were argued to be

making illegal copies by moving data from CDs to computers. Copyright is generally how
software is protected in the EU rather than by patents. It is not possible to violate
copyright which you were not aware of but it is possible to violate a patent of which you
are not aware.

The US allowed very poor patents but had a very weak opposition system with
compulsory licensing for social benefits. The parenting of things like the peanut butter
and jelly sandwich demonstrates the absurdities that strong patents which the European
Patent Office have resisted.

International competition
For Cion, competition policy was always trade policy, working to encourage a single free
market within the EU.

Despite how markets are modelled markets are not perfectly competitive, under
monopolies and the forms of restricted competition supply curves do not move upwards.
One of the most important roles of international trade it to make markets more
competitive. The single market in the EU gave the opportunity to have both more
competition and large firms. Despite this in the US there is a deeply held view that
merely opening markets by removing trade barriers will create competition and a deeply
embedded opposition to multilateral competition rules.

International competition policy is important, although mergers can be dealt with by

different jurisdictions (the most stringent wins) business is generally supportive of a
single authority. There are strongly imperfect interactions between national completion
an international trade policy. For example the interaction between anti-dumping (about
price differences between home and foreign markets) and predatory pricing (about profit
restrictions). Voluntary Export Restraints are in reality a cartel and a violation of
competition law, competition authorities do not address VERs however as they see them
as a form of trade policy. Real international cartels, vitamins, soya, potash etc are all

31
 Thursday, 26 February 2015
problems with no international regulation, American law allows US companies to form
cartels overseas and Germany has a long history of legal cartels. The EU has found that
simply opening broders is not enough tot ensure free competition, many cartels survived.

Fighting cartels as a major priority for 1940s international trade policy, the GATT was the
trade in goods chapter of the Havanna charter of the International Trade Organisation (a
proto-WTO nixed by the US Congress refusing to sign), in Germany and Japan cartels
were seen as being responsible for allowing the emergence of Hitler and the
Imperialists. Fighting cartels didn’t remerge until the 1980s when the US became
concerned about cartels ripping off americans.

In the anglo-saxon tradition unfair competition is simply a slogan used to undermine

competition by those who oppose free competition. The continental tradition is more
inclusive of a role for fairness in competition. Dumping for example has no relation to
monopolisation or predation, even if making a profit in both countries an exporter can be
accused of unfair damage to competitors by having lower prices. Anti-dumping
legislation is intended to be protectionist and to be inefficient and ineffective. Anti-
dumping is not however much more inefficient than the forms of protectionism as it is not
intended to address consumer welfare or efficiency. Anti-dumping has seemed to attract
investment into Europe however.

Adam Smith devoted 50 pages of the Wealth of Nations to railing against the East India
Company and monopolies. There was the some concern about monopoly buying power
depressing income but also the raising of prices in Britain. Lenin echoed this in his attack
on imperialism. The Havanna charter supported action to prevent restrictive business
practices distorting trade although this did not make it into the GATT. The EU did adopt
these principles in the Rome Treaty, the OECD did to an extent in UNCTAD. The WTO
appears to have snuck some in the GATS but the americans remain a road block to
international competition agreements. The EU continues to push for international
competition authority, it is mandatory between EU states, there are voluntary
cooperation provisions in FTAs and some member states have strong bilateral
agreements such as the UK-US Mutual Legal Assistance Treaty which, despite there
being no requirement for competition authorities to cooperate, can result in extradition in
competition cases.

The problems of cooperation in international competition include the effects doctrine.

The US rejects the doctrine of jurisdiction where outside action has internal effects, the
EU has a much more catholic interpretation of this. The British-Japanese car cartel was
conducted in Singapore with only one British partner to avoid UK law but failed to
remember that EU law applied as there were intra-EU effects. There are very few formal
international cooperation agreements, with the exception of a few weak EU-US
agreements which introduce ‘positive comity’ - this is the request to tackle actions illegal
in one jurisdiction which are having and effect in the other. Again and again international

32
 Thursday, 26 February 2015
competition cooperation falls down on authorities refusal to exchange confidential
information. British American Tobacco for example relies on Cion not to reveal its
actives, legal in the EU, which are illegal in the US. Only the US - Australia treaty
provides for assistance when each other laws are violated. The developing countries
avoid signing treaties because they gain no useful information to fight malpractice and
are forced to adopt competition policies which open their markets to european firms in
particular.

Big international cartels last on average 7 years (although Phoebus lasted 50), US used
unilateral extra-territoriality to reject the principle of anti-trust rulings to be reviewed by
the WTO. India supported moves for inter national competition rules but found EU
wanted other WTO member to introduce competition law, but unwilling to agree
obligations on case by case assistance

VERs allow exporters agree not to sell more than set quota and tariffs not raised – so
they get rents. Under both the voluntary export restraints for clothing & cars firms try to
upgrade and get round then VER and cartels to exploit quota premiums. Uruguay Round
banned previous grey area measures including private market sharing rule affecting
trade. The UK - Japanese car VER violated EU law (no sovereign compulsion at either
end) and continued until the EU signed an anti-VER law.

Anti-dumping s about price discrimination below normal value, not about predation. The
normal value is a contentious point and equal prices might not reflect normal prices. Anti-
dumping is the last weapon in political economy with anti-dumping similar to
countervailing anti subsidy duties. Almost all FTAs provide for anti-dumping duties and
always will do. Anti-dumping can be used strategically, with anti-dumping actions brought
against themselves by potash companies to force companies to raise prices.

There have been only a tiny number of cases where a merger has been banned in one
market and not in another. The result is either divestment or failure to merge.
Cooperation does now occur on mergers as it is in everybody’s interests.

Whilst there are no actual competition rules internationally there are WTO rules affecting
competition policy. GATT III has a generalised anti-discrimination provision. GATT XXIII
ensure nonviolation. GATT XIX bans VERs. GATS VIII & IX require member states to
prevent firms in a service industry without liberalisation from exploiting their monopoly or
dominant position in one market in another. TRPS allows for breaking IPRs for generic
drugs in cases of public health emergencies.

At the Cancun round of WTO negotiations the EU proposed that worldwide acceptance
of some core principles of comp policy esp. ban on hard core cartels and a framework
for voluntary co-operation. Many developing countries were resistant although some
opposed some favoured it principle. Few thought EU plan would deliver much of value,
but failed to ask for more. G-90 opposition (not India) killed it at Cancun when the African

33
 Thursday, 26 February 2015
countries walked out despite the fact that competition is now in most FTAs. There has
been some talk at WTO of revival but no action as yet.

Only two WTO cases have been brought as yet, Kodak Fuji and Mexican Telecoms -
whilst the Mexican telecoms case did demonstrate that there was some competition law
in the WTO the Kodak case found nothing of any particular interest.

Whilst there are ballooned pros and cons to international competition cooperation it is
fundamentally because of the US that there is no world competition regime.

Regulation
The British were historically opposed the international competition policy until 1997 and
Claire Short started to push the EU agenda because of the development benefits.
Regulation is required when a market cannot be naturally rendered competitive.

Cases where regulation is expected include Natural monopoly, Excessive entry,

Asymmetric information, Bounded rationality (only for consumers?), Externalities
(spillover effects). Regulation may be seen as distinct from competition policy in that it
proscribes ex ante rules of conduct and competition policy lays down ex post
punishment of forbidden actions.

Price regulation often seen as distributional by non economists, the US Progressive Era
presented anti-trust as way to discipline powerful players and US regulatory policies
seen as aiming at fairness (lower prices) not efficiency until the Chicago School. German
Ordo-liberalism saw need for Mittelstand for socio-political reasons. Sharia law has
always embraced the idea of a “just price”.

The authority in which regulatory authority is vested can vary considerably from self
regulating state owned enterprises, directly from ministries, independent regulatory
agencies with well defined mandates, or via contacts for service provision. It has greenly
been accepted that the role of a regulator is not to keep prices low but there is still a role
to ensure that disadvantaged classes are not disproportionately affected. One of the
paradoxes of regulation is that regulators set the costs of consumer protection so high
that it becomes a barrier to entry.

The principle argument in the regulatory literature is that if you lay down rules in advance
you run the risk of regulatory capture most likely by the offer of highly remunerated jobs
outside the regulatory authority. Competition policy is as prone to regulatory capture as
any other area. DG Comp was in alliance with the US DoJ and not with DG Trade.
Sectoral regulators go to sectors, general regulators find jobs in law firms and
consultancies. The political economy of competition regulation should look at them as
captured entities.

34
 Thursday, 26 February 2015
Correct regulation can have net benefits so long as firms can cover costs, costs of
administering regulation must not exceed gains, regulators must also have adequate
access to information and not be captured.

If we accept that free competition will not produce an optimal outcome what strategies
for good and outcome improving regulation can regulators take. Consider asymmetric
information: how do regulators know true costs etc so set regulatory parameters right?
Policy may be too favourable to firms if info all from them. How do firms know true
intentions of regulators? Firms may refuse to invest without risk premium reflecting risk
of policy change, e.g. mobiles in Africa. How can regulators pre commit without losing
scope to change rules if new circumstances or information? Eg placing conditions on
mergers. Regulators may seek to exploit “lock-in of firms”, changing rules -or just
demanding bribes.

Natural monopolies emerge when economies of scale are such that costs of production
will always be lower when only one firm produces at any imaginable demand.
Economists now claim that technology has changed and few industries are truly natural
monopolies – or that if parts are (eg rail or phone(?) network) we can open up use of that
network to competing services. Regulation may make it possible for core of network
being open to all as an essential facility, “common carrier”

35