Transcript 200665111657175

The economics of
vertical restraints
Patrick Rey (IDEI, Toulouse)
Cargese
May 7, 2004
Outline
• Introduction
• Inter-firm agreements
– Intrabrand coordination
– Interbrand competition
• Market power: foreclosure
1
Introduction
• What are vertical restraints?
– Revenue sharing: other than a simple “linear price”
• Non linear tariffs (two-part tariffs, rebates, menus of tariffs, … )
• Royalties (on sales, total sales, profit)
• Commissions
– Obligations (restrictions on behaviour)
•
•
•
•
•
Service specification (upstream/downstream, national/local advertising, …)
Quantity forcing, quotas
Exclusivity: exclusive dealing, exclusive territories (active/passive sales)
Resale price maintenance – RPM (price ceiling/floor, recommended prices, …)
Bundling, mixed bundling, full-line forcing
2
Introduction
• Questions
– Positive economics: Motivation?
– Normative economics: Competition policy (allow, encourage, forbid, …)?
• Two aspects
– Inter-firm agreements
• Vertical coordination (intrabrand)
• Strategic effects (interbrand, « short-term »)
– Abuse of market power: foreclosure (entry/exit, « long-term »)
3
Intrabrand vertical coordination
• Prices Spengler (1950)
– double marginalisation
w > c => p > pM
– solutions
• RPM (ceiling), sales quotas (min.)
• two-part tariffs
• intrabrand competition
– private and social objectives coincide
what is good for the firm
is good for consumers
unit cost c
P
wholesale price w
R
retail price p
Consumers
q = D(p)
4
Intrabrand vertical coordination
• other dimensions
– risk-sharing Rey-Tirole (1986)
• local shocks on retail cost/demand
• bias towards ET / retail competition
unit cost c
P
wholesale price w
– retail services
• excessive price, insufficient effort
• solutions:
– RPM (ceiling) + sales quotas (min.)
– two-part tariffs
– not intrabrand competition
R
retail price p
retail services e
• Private and social interests may diverge
– marginal / infra-marginal consumers
Spence (1975)
Comanor (1985), Caillaud-Rey (1987)
– more congruence if free-riding
Mathewson-Winter (1984)
Consumers
q = D(p,e)
5
Interbrand competition
• “competition-dampening”
– Exclusive territories: strategic delegation
Bonanno-Vickers (1988), Rey-Stiglitz (1985, 1988)
• ET reduce intrabrand competition
• retail prices respond to increases in
rival wholesale prices
• higher wholesale (and retail) prices
P1
P2
R1
R2
Consumers
→ clear conflict between private/social interests
– More generally: strategic commitment
Caillaud-Jullien-Picard (1990), Caillaud-Rey (1994)
6
Interbrand competition
• RPM and interlocking relationships Rey-Vergé (2003)
Manufacturer A
A-1
Manufacturer B
B-1
A-2
Retailer 1
B-2
Retailer 2
Consumers
Note: consumer goods
7
Interbrand competition
– Competition in two-part tariffs: p < pM
• Manufacturer i recovers retail profits
→ sensitive to retail margins on rival brand
max (pi – c – γ)Di + (pj – wj – γ)Dj
• w > c to maintain « high » retail prices in spite of retail competition
→ i does not take into account the upstream margin on rival brand: p < pM
– Two-part tariffs + RPM: p = pM
• No need for w > c to maintain high retail prices
• But if wj = c, manufacturer i becomes sensitive to the full margin on rival
brand → p < pM
– Note: the Galland bill
8
Interbrand competition
• Collusion: improving market transparency
Example: RPM
Jullien-Rey (2003)
allows for more uniform prices in spite of variations in local
conditions of supply and demand
• Facilitates the detection of deviations
• But inefficient rigidity
→ can/will be used when enhances collusion substantially
→ negative impact on consumers and efficiency
9
Foreclosure
• Supposes an “essential facility”
– Essential
– Controlled by a dominant firm
– No “objective” reason to deny access
• Traditional concern (leveraging market power)
Dominant firm
– denies /limits access to some potential users,
– to extend its market power from the monopolized segment to the
complementary segment
10
Foreclosure
• Vertical / horizontal foreclosure
monopoly market A, competitive market B
upstream/downstream
complementary
A
B
M
C
A
C
M
B
C
C
Terminal RR 1912, Commercial Solvents 1973
IBM CPU - peripherals
Infrastructure: stadiums, ports, airports, tunnels, RR,
United shoe, Chicken Delight, Tetrapak
electricity grid, local loop, Computer reservations syst.
Server OS / PC OS / applications
Patents
Aftermarkets (Volvo, Renault, Kodak)
11
Practices
• Vertical foreclosure
– vertical integration +
•
•
•
•
refusal to deal
incompatibility
high wholesale prices
tie-ins, ...
– no vertical integration but
• exclusive dealing
• price discrimination, ...
• Horizontal foreclosure
– tying,
– access,
– incompatibility
M
M
C
M
C
C
M
C
12
Remedies
• Structural
– common ownership of bottleneck (Terminal RR)
– break up + line of business restrictions (AT&T)
• Regulation of access price
–
–
–
–
no discrimination among external clients (CAB 1984, Sabena / Saphir)
no discrimination between external and internal clients:
transparency, accounting separation (“chinese walls”)
price linkage: ECPR
access charge = final price - (marginal cost on competitive segment)
example: local loop resale
• Open access (common carrier)
• Regulation of wholesale quantities (Eurotunnel).
13
Vertical foreclosure

Chicago school critique
Posner (1976), Bork (1978), Posner-Easterbrook (1981)
“Only one profit”: how can bottleneck owner earn more than one profit?”
M
• M charges
– Wholesale price w
– Franchise fee F
• Downstream competition
– Retail price p(w) = pm
– Profits recovered through F
C
C
Demand
14
Vertical foreclosure
• Response to the critique
– Incentive: restore, rather than extend market power
Upstream monopolist cannot exercise monopoly power without
excluding
Hart-Tirole (1990), O’Brien-Shaffer (1992), McAfee-Schwartz (1994),
Rey-Vergé (2004), ..., Rey-Tirole (2003) « A Primer on foreclosure »
– Analogies
• Patent: multiplication of licenses
• Franchising: multiplication of franchises.
15
Vertical foreclosure
• Model
M
C1
unit cost cu
C2
Cn
1 unit of input required
for 1 unit of output
unit cost cd
Consumers
16
Vertical foreclosure
• Example (“Cournot” downstream competition)
– Framework
• M offers wholesale contracts (e.g., two-part tariffs: Ti(qi) = Fi + wiqi)
• each competitor Ci orders its quantity qi and pays accordingly Ti(qi)
• each competitor Ci set its price pi
→ downstream competition analogous to “Cournot”
Kreps-Scheinkman: pi = P(q1 + … + qn)
– If wholesale tariffs are “public”: monopoly outcome
• Monopolistic franchise contract w = cu, F = πm
• Oligopolistic franchise contract w: pC(w+cd) = pm, F = πC(w+cd)
17
Vertical foreclosure
• Secret contracting: Opportunism
Hart-Tirole (1990)
When dealing with a competitor C, M has an incentive to free-ride on the
sales of the other competitors
πM + πi ~ (P(qi +Σj≠iqj) - cu – cd)qi (+ Σj≠i (Tj - cuqj))
→ it is optimal for each competitor C to order and for M to supply a
quantity that is the “best reaction” to the others’ production levels
→ Cournot outcome (quantity competition)
→ as number of competitors increases, price goes down to cost
(competitive pricing, no market power)
18
Vertical foreclosure
• Variant: Bertrand downstream competition
O’Brien-Shaffer (1992)
When dealing with one competitor Ci, M still has an incentive to free-ride on the
downstream margins of the other competitors
πM + πi ~ ( pi - cu )Di + Σj≠i(wi – cu)Dj
→ It is optimal for M and each competitor C to agree on a price that is the “best
reaction” to the others’ prices
→ Bertrand outcome (price competition)
Issue: which conjecture?
Passive, wary beliefs McAfee-Schwartz (1994), Rey-Vergé (2004)
19
Vertical foreclosure
• Foreclosure: restoring market power
– reputation, transparency
reduce scope for opportunism
– vertical integration
no incentive to free-ride on its own subsidiary
– exclusive contracts
eliminates downstream competition
– nondiscrimination laws (!)
eliminates opportunism
– RPM, …
• Remarks
– Incentive for foreclosure stronger
• the more competitive the downstream industry,
• the less competitive the upstream industry
– Some competition upstream generates some access.
20
Vertical foreclosure
• Comment: who’s “upstream”?
Bottleneck is upstream
M
C2
C1
C2
C1
Bottleneck is downstream
Clients
Illustration : US gas reform (pipelines)
M
Clients
21
Vertical foreclosure
• Efficiency defenses
– Benefits from vertical integration
– Maintaining upstream reputation
– Cost of increasing capacity
– Investment and innovation
• Regulation of access = regulation of rate of return
• Where does the market power come from?
–
–
–
–
historical reasons / legal monopolies (port, airport, ...)
scale economies (Otter Tail, Hecht, ...)
network externalities (WorldComMCI, Open Network Provision directive, ...)
investment / innovation
22
Horizontal foreclosure
• Chicago school critique
Complement goods
→ competition in the B market
• makes A good more attractive
• raises profit for M
→ M has no incentive to
reduce competition in B market
A
M
B
C
C
Demand
q = D(pA+pB)
23
Horizontal foreclosure
• Response to the critique
Two lines of arguments:
– Non complements
a second source of monopoly power does not devalue M's monopolized product A
– Complements
entry in the adjacent market B may facilitate entry in the monopolized market A
24
Horizontal foreclosure

Independent goods: Whinston (1990)
A
B
M
M, C
• Assumption: commitment to sell A and B only as a bundle
→ M becomes a very aggressive competitor if C enters B market
• Conclusions
–
–
–
–
tie-in costly if C enters/stays in market,
tie-in tends to discourage entry (commitment to being aggressive)
tie-in may be profitable if deters entry,
(because of Chicago school reason) does not work if A and B complements.
25
Horizontal foreclosure
• Complements: protecting home market
– Choi-Stefanadis (2001)
• Initially, M monopolist in both A and B
• In each market, C can invest I
– enters with probability ½
– if enters, gets profit π
• Without tying, C enters (in both markets) if ½ π = π/2 ≥ I
• With tying, C must succeed in both markets:
→enters only if ¼ π = π/4 ≥ I
– Carlton-Waldman (2002)
• Sequential entry in B, then in A
• Scale economies
26
Horizontal foreclosure
• Efficiency reasons for tie-ins
– Prevention of inefficient input substitution
illustration: durable good maintenance
– Protection of reputation
Signal of quality for a durable good
(profit made on complementary good)
– Price discrimination / metering of demand
• IBM cards, Chicken Delight
• aftermarkets
27