The GSK France Decision of March 2007 Outline of the presentation 2. Predation Test 3. Application of the test A brief history of the case
• July 2000: Flavelab refers the case to CC, asks for interim measures
• November 2000: CC does not grant interim measures • December 2001: Flavelab goes bankrupt • April 2002: Flavelab acquired by Panpharma
– 2003: Complaint withdrawn – CC proceeded ex officio
• July 2004: Statement of objections
– 2006: first hearing, supplementary SO, final hearing
• March 2007: Decision on the merits, fining GSK France
• Paris Court of Appeal's judgement expected early 2008 Products
• Anti-infective drugs (“J” in the ATC classification) • Sold (almost) only to hospitals (bidding markets)
– Community market not involved
• Market A: Injectable Aciclovir (ATC: J05A)
• Market B: Injectable cephalosporins 2nd generation
– Prevent infections during surgical operations– Cefuroxime and Cefamandole– Not disputed
Market A: Aciclovir Market B: Injectable cephalosporins Market size
≈10 m€ Market size in 1999 and 2000
≈ 2m€ in 99 and 00 Injectable Zovirax® GSK: Zinnat® since 1983 Lilly: Kéfandol® Antitrust definition of markets and position of the firms
• Market A: No substitute for Injectable Zovirax® according
to the regulator and hospitals' pharmacists
– Market share: 90% in 2000, 80% in 2001– Supposedly protected until 2002 – Merck's entry late 1999 (around 10% in 2000)
• Market B (Lilly and GSK)
– Definition is not disputed– No dominant position on this market
TIMELINE Market A: Zovirax® protection extended until Sept 2002 Sept 99: Entry Merck Sept 02 Entries: Arrow No patent litigation Dakota-Pharm, Ggam Flavelab Entry → Panpharma litigation Market B: Zinnat® protected until May 1999 Predation period: 1999-2000 Practices
• 1999-2000: below cost pricing on market B
– Zinnat® purchased from the Adechsa company– Retail prices < purchase prices (PP)– Purchase prices computed from invoices net of all
– 12 markets (hospitals, dosage) in 1999, 29 in 2000– Selective price cuts: prices below cost when faced to
• Bundled rebates: special prices conditioned to the joint purchase of Zinnat® and Zovirax®
– 4 instances in 1999
➢ Price > ATC: ok (but exceptional circumstances)
ATC > Price > AVC: burden of proof on CC
– Multiple indicators: hard evidence, selective price cuts,
supplementary practices, likely recoupment, fact-based
theory of harm (e.g. Vulnerable prey, etc.)
AVC > Price: burden of proof on defendant
– defendant can rebut presumption by proving
recoupment ex ante impossible (low entry barriers),
perishable goods, learning by doing, switching cost.
Standard of proof suggested by CC = same as Motta (2004) to be confirmed by the Court of Appeal Relevant cost benchmark
• GSK: Drug purchased from a unit of the same group, transfer
prices within a group, no economic meaning
• CC: There are only two possibilities
1. Subsidiary company (GSK France) is held liable
● Competitive benchmark = maximization of the short-
2. Parent company (UK group, 2nd-largest drug company
● Competitive benchmark = maximization of the short-
Liability and competitive benchmark
• CC considered GSK France liable
– Internal organization and decision process of the firm:
• “business unit” dedicated to hospital drugs, with a
– The subsidiary company claimed that “no other company in the group was concerned by the practices”
Once a firm is considered liable, one should use its costs.
Purchase price (PP) = lower bound of AVC AVC (>PP) also include distribution costs (marketing, etc.)
PP effectively paid (invoices), quantity-dependent
Ex ante / Ex post
• GSK: CC should dismiss instances with Bid = PP
• CC: Since purchase prices are falling, ex post test conservative Bid for year (t+1) ≤ PP (t+1) < PP (t)
– GSK did not explain how expectations are formed, so PP(t)
– PP(t+1) follows from retroactive rebates that cannot be
Price analysis at which level?
• GSK: CC should consider average prices across all hospitals • CC: exclusion occurs at the hospital level. Bids at each
invitation to tender relevant, because of selective price cuts
• GSK: contests selective price cuts • CC:
– This is only an indicator (makes predation less costly) – Large differences in the bids depending on competitive
pressure (can be anticipated depending on the wording of
Bids with comp. < PP(n) < PP(n+1) < Bids without comp. Link between markets A and B
• GSK : contests the link
– Same seller: GSK France (hospital business unit) – Same buyers (hospitals) buy Zinnat® et Zovirax® – Bundled rebates in 1999 – Rationale for predation in (small) market B: deterring
generic drug companies from following Merck into (large)
• Same as in Akzo: the incumbent prices below cost in the
small market to deter ECS from entering its core market
Theories of harm (1/2)
• Threats of entry in 1999 in spite of patent's extension
• In 1999, 7 (3) generic drug companies had regulatory approvals to enter market A (B)
• Merck: small-scale entry on market A in 1999, but 1m€
• GSK was unsure about the outcome of a patent
litigation, did not challenge Merck in court
• Build a reputation of aggressiveness [Kreps and Wilson, 1982]
– Two types of incumbent: tough/weak, 2 markets, 2 periods
– After entry on market B at T=1, I has incentives to prey to avoid entry
on market A (deter entry or discipline an entrant)
TIMELINE Market A: Zovirax® protection extended until Sept 2002 Sept 99: Entry Merck Sept 02 Entries: Arrow No patent litigation Dakota-Pharm, Ggam Flavelab Recoupment Entry → Panpharma litigation Market B: Zinnat® protected until May 1999 Predation period: 1999-2000 Theories of harm (2/2)
• CC also mentions a possible financial predation scenario
• Common story: I sacrifices short-run profit to manipulate E’s
profit expectations, distort entry decisions, and recoup later
– Sacrifice is impossible to check look for short-run losses– Conservative test
• Reminder: Since P<AVC, CC has not the burden of establishing such a theory; the defendant has to prove it is manifestly impossible and to give an objective justification Average Price Zinnat® 1.5g (€) Flavelab Flavelab Panpharma Interim measure decision Nov 2000 Possibility of recoupment
• GSK: No entry barrier after patent's expiration, recoupment
– No need to prove recoupment ex post– BotEC just to reply to the defendant's point– Losses on market B: 75 000 € in 1999-2000 (very little!)– Gains on market B in 2001-2002 = extra profit earned by
pricing above the competitive price (Proxy: price of
– Recoupment even easier on market A, where GSK has
more market power (this is just an example)
Other arguments
• GSK: raised prices for fear of being found guilty of predation,
following CC 2000 decision, not for recoupment
• CC: The motivation of the price rise does not matter. This is
just to show that recoupment is not impossible.
• GSK: Flavelab went bankrupt because inefficient • CC: Flavelab good at selling Cefamandol (in market B) and
Flavelab affected by predation (concerned drug accounted for
• CC: GSK did not provide any justification for its behavior Meeting competition defense
• GSK: price cuts necessary to align on Flavelab
• CC: GSK could have aligned without violating the GSK Bids below costs for n+1 < PP(n+1) < Flavelab < PP(n)
• GSK: Price cuts necessary to align on Kefandol® (Lilly)
• CC: Not true: Kefandol® price > predatory prices of GSK Evolution of markets
• Market B:
– 3 generic drug companies have approval, only one
– 6 years after patent's expiration (2005), GSK market
• Market A
– 3 years after patent's expiration (2005), GSK market
– Out of 13 firms having regulatory approval, only 5 enter– Generic drug companies active on market B observe the
– Panpharma did not try to enter market A
Sanctions
• Infringement of Art. 82 (multinational firms, imports, etc.) • Seriousness
– Exclusionary practices– 40 markets involved, generics' entry delayed– GSK: first supplier of hospitals
• Fine: 10 m€
– Legal ceiling at the time: 5 % of GSK France turnover in
– Fine < 1 % of the ceiling– Fine ≈ Size of market A
• Publication of the decision in professional journals Conclusion
• Sends a simple message to markets:
“If you are an autonomous firm and if you are in dominant position, don’t sell below your average variable cost, unless you have an objective justification for it”
Central Union High School District Health and Welfare Benefits 2010-2011 All Benefit Eligible Classified Employees (and COBRA Enrollees) Carol Moreno, Director of Human Resources HEALTH AND WELFARE BENEFITS EFFECTIVE OCTOBER 1, 2010 The annual open enrollment for making medical, dental, vision and life insurance plan changes is from Tuesday, August 10, 2010 through Friday, Se
The GSK France Decision of March 2007