This presentation by Matthew Bennet, Vice president, Charles Rivers Associates, was made during the discussion “Hub-and-spoke arrangements” held at the 132nd meeting of the OECD Competition Committee on 4 December 2019. More papers and presentations on the topic can be found at oe.cd/hsa.

1. The economics of Hub and Spoke
Agreements
Matthew Bennett
Vice President, CRA
1

2. Matthew Bennett
December 2019
Structure
Overview
Economics of hub and spoke
Applying economics to law?
Conclusions
2

3. Matthew Bennett
December 2019
Why hub and spoke enforcement?
Want to prevent firms from colluding.
• Therefore rule out agreements between firms to set prices.
But what happens if firms just talk about intentions without explicitly agreeing to collude?
• Don’t want firms to be able to bypass law.
• Thus may want to take hard line on some exchanges, e.g. future pricing intentions between
competitors.
What happens if firms have same discussion through third parties?
• Again – don’t want firms to bypass law.
• Thus may be concerned by some discussions of prices between independent parties if believe they
are intended to be passed on to rivals.
3

4. 1. The economics of “hub and spoke” exchanges
4

5. Matthew Bennett
December 2019
Two broad cases of hub and spoke
Retailer
Supplier A Supplier B
Price
info
Price
info
Supplier to Retailer to Supplier Retailer to Supplier to Retailer
Supplier
Retailer A Retailer B
Price
info
Price
info
5

6. Matthew Bennett
December 2019
Supplier to Retailer to Supplier
Theory of harm
Flow of information may allow
coordinated outcomes between
suppliers.
Note however that exchanging
information through a third party is
generally more difficult than direct:
• Coordination often requires trust
built up through direct
communication
• Third parties may have differing
incentives particularly in vertical
relationships
6
Retailer
Supplier A Supplier B
A’s
price
info
A’s
price
info
B’s price
info
B’s price
info

7. Matthew Bennett
December 2019
Supplier to Retailer to Supplier
Retailer incentives to participate?
In simple model may not expect retailers to have an incentive to pass through
information that facilitates higher supplier prices:
• It directly increases their costs and hence reduces their profits (even if all their competitors
follow).
However retailers may still act as a conduit for exchanges:
• If suppliers have market power (must have) retailers may have little choice.
• Non-linear prices - Suppliers may provide retailers with discounts or rebates (e.g. side
payments) to provide retailers with incentives (i.e. as long as total profits go up, it can
reallocated).
• Retailers may be focused on ensuring parity between their rivals rather than lower input prices
for themselves.
7
Retailer
Supplier A Supplier B
Price
info
Price
info

8. Matthew Bennett
December 2019
Supplier to Retailer to Supplier
Pro-competitive effects
Pro-competitive effects
Competitive S2R2S exchanges may be very common, even more so than direct exchanges.
• Retailers will regularly go back and forth to suppliers to get the best possible price from a supplier.
• Retailers may quote prices from a supplier’s competitor in order to negotiate a lower price from the
supplier.
• Ruling out such behaviour may rule out retailers’ ability to bargain and increase consumer prices.
Balancing
Over all suggests one should be much more cautious in pursuing an indirect exchange of
information via a retailer than a direct exchange.
• Coordinating through third parties is generally harder than direct coordination.
• There are reduced incentives for retailers to facilitate them (relative to direct exchanges).
• There are much more frequent and stronger benefits in vertical set-up.
However this does not mean cases do not exist!
8

9. Matthew Bennett
December 2019
Retailer to Supplier to Retailer
Theory of harm
Flow of information may allow
coordinated outcomes between
retailers.
Supplier
Retailer A Retailer B
A’s price
info
B’s
price
info
B’s
price
info
A’s price
info
9

10. Matthew Bennett
December 2019
Retailer to Supplier to Retailer
Supplier Incentives?
Suppliers may not want to pass information that facilitates higher retailer prices:
• It increases retailer price, reduces retail sales and hence demand and thus reduces supplier
profits, even if all rivals follow.
But suppliers may still facilitate exchanges:
• Bargaining power - if suppliers need access to retailers (must have) then suppliers may have
little choice.
• If competing suppliers face similar agreements, then there may be less incentive to fight back
against retailers.
• May allow suppliers to reduce retailers bargaining power thus increasing their share of the pie
(even if overall pie is reduced), see next slide.
10
Supplier
Retailer A Retailer B
Price
info
Price
info

11. Matthew Bennett
December 2019
Retailer to Supplier to Retailer
Reduction in retailer bargaining?
Supplier bargaining power story behind supplier incentives?
• By reducing competition between retailers, a supplier may soften retailers’ bargaining
power/incentives vis a vis itself.
• If all retailers know they have the same price, they have less incentive to negotiate lower input
prices from supplier.
• If supplier ensures a common retail price it means no retailer can translate a wholesale cost
advantage into a retail price differential and thus less incentive to bargain on wholesale prices
(assuming there is a cost to bargaining).
• Therefore a common retail price may increase the share of pie that the supplier gets.
In a bargaining game it may result in the supplier having an increased share of the pie,
even if its volumes fall.
11
Supplier
Retailer A Retailer B
Price
info
Price
info

12. Matthew Bennett
December 2019
Retailer to Supplier to Retailer
Pro-competitive effects?
Pro-competitive effects:
Pro-competitive R2S2R facilitated exchanges may be more common than direct
exchanges, but less common than S2R2S.
• When suppliers have limited output which they need to allocate between retailers they may
negotiate on the basis of the retail price – with the lowest price retailer allocated the most
volumes.
• When suppliers have a high value brand, they may have an incentive to have a high retail price
in order to protect that brand – therefore prevent discounting.
Balancing?
Over all suggests one may still be more cautious about pursuing these cases than
direct exchanges, but worry more about these than S2R2S cases given less obvious
pro-competitive rationale.
12
Supplier
Retailer A Retailer B
Price
info
Price
info

13. 2. Applying economics to law?
13

14. Matthew Bennett
December 2019
What does economics say about designing a law?
Want to differentiate between pro-competitive and anti-competitive:
Pro-competitive
• Use of information in order to bargain and extract lower prices which can be passed on
downstream.
Anti-competitive
• One party giving information to another knowing that (or expecting that), it will be passed on to
its competitor in order to coordinate.
Differentiator regarding to ‘knowing’ or ‘expectation’?
• In pro-competitive case there is no expectation from firm that its information will be passed onto
its rival.
• In anti-competitive case there is the expectation it is passed to rival.
Thus perhaps use evidence of expectations of being passed on?
14

15. Matthew Bennett
December 2019
The UK test: Three limb test
Limb 1:
A to B
A discloses future
pricing information
to B in
circumstances
such that A can be
assumed to
intend, or does
intend, that B will
pass onto C
15

16. Matthew Bennett
December 2019
The UK test: Three limb test
Limb 1:
A to B
Limb 2:
B to C
A discloses future
pricing information
to B in
circumstances
such that A can be
assumed to
intend, or does
intend, that B will
pass onto C
B passes info to C
in circumstances
such that C may
be taken to know
context in which it
was passed on,
and/or information
is passed on with
As understanding.
16

17. Matthew Bennett
December 2019
The UK test: Three limb test
Limb 1:
A to B
Limb 3:
Use by C
A discloses future
pricing information
to B in
circumstances
such that A can be
assumed to
intend, or does
intend, that B will
pass onto C
C uses the
information in
determining its
own future pricing
intentions
Limb 2:
B to C
B passes info to C
in circumstances
such that C may
be taken to know
context in which it
was passed on,
and/or information
is passed on with
As understanding.
17

18. Matthew Bennett
December 2019
Conclusion on A2B2C cases
All A2B2C cases can have the same anti-competitive effects as direct exchange cases -
coordination.
• But direct contact is often a key agreement to coordinating (requirement of trust).
• Furthermore conduits may have incentives to resist these anti-competitive effects.
• In addition there may be pro-competitive stories regarding bargaining – particularly with regards to
R2S2R cases.
Need for legal test that differentiates.
• Legal test requirement of knowing/expectation that it will be passed on, attempts to do this.
• Minimises cases of pro-competitive exchanges being wrongly prosecuted.
Should conduit’s incentives play a role?
18

19. 4. Platform/Retail MFNs and Hub and Spoke
19

20. Matthew Bennett
December 2019
Platform or retail MFNs (i)
Platform A Platform B
Retailer (1) /
Hotel (1)
20
Consumers
Price on platform B (P1
B)Price on platform A (P1
A)
Charge for Retailer
being on platform
(or commission)
I don’t mind what price you
charge on my platform,
as long as it is no different
from what you are charging
on Platform B (i.e. P1
A = P1
B)
(P1
A) (P1
B)
P1
A = P1
B

21. Matthew Bennett
December 2019
Platform or retail MFNs (ii) – Potential Harm
Platform A Platform B
Retailer
/Hotel 1
21
Consumers
Price on platform B (P1
B)Price on platform A (P1
A)
Charge for Retailer
being on platform
(or commission)
I don’t mind what price you
charge on my platform,
as long as it is no different
from what you are charging
on Platform B (i.e. P1
A = P1
B)
CBCA
In normal competition there is a
cost of increasing wholesale
cost because it gets passed
through to retail prices and thus
results in switching to rival
brands and lower volumes.
MFN breaks or makes the link
‘stickier’ between wholesale
price and retailer price. If
Commission A goes up, then
Retailer can’t increase price on
platform A (P1
A) without
increasing price on B as well.
Incentive for Platform A to
increase price increases
because it knows it won’t lose
any volumes to B as relative
prices won’t change.

22. Matthew Bennett
December 2019
Potential benefits from platform/retail MFNs
Main concern is potential for platforms to face free-
riding by sellers or other lower quality platforms.
• Platforms undertake investments to attract
buyers onto the platform.
• However seller has an incentive to find buyers
on the platform but then encourage them to
transact off of the platform to avoid
commission.
22

23. Matthew Bennett
December 2019
Similar to RPM the upstream
firm is constraining the
downstream retailer in what it
can charge to consumers.
• But means of effect is not
really same, in RPM direct
effect is to restrict
competition between two
retailers, whilst indirect it
softens upstream
competition, whilst
Platform MFN directly
softens upstream i.e.
doesn’t directly restrict
downstream competition.
Not so clear there are
similarities to hub and spoke?
• Coordination theory is
between Platform A and B,
but information flow is on
the retail price, not the
upstream commission.
• Thus not clear any
information flows facilitate a
cartel beyond the softening
of competition theory?
Similarities may arise if there
are additional clauses around
requiring information
disclosures on wholesale price.
Similarities to RPM and information exchange?
Platform A Platform B
Retailer 1
23
Consumers
Price on platform B (P1
B)Price on platform B (P1
B)
Charge for Retailer
being on platform
(or commission)

24. 24
London
Tel +44 (0)20 7664 3700
8 Finsbury Circus
London EC2M 7EA
United Kingdom
Brussels
Tel +32 (0)2 627 1400
143 Avenue Louise
B-1050 Brussels
Belgium
Munich
Tel +49 89 20 18 36 36 0
Palais Leopold,
Leopoldstraße 8-12
80802 München
Germany
Paris
Tel +33 (0)1 70 38 52 78
27 Avenue de l’Opéra
75001 Paris
France
www.crai.com/ecp
Lyon
Tel +33 1 70 38 52 78
3 place Giovanni Da Verrazzano
69009 Lyon
France
Sydney
Tel +61 (0)406 820 214
Level 22, Tower 2
101 Grafton Street
Bondi Junction NSW 2022
Australia