1

TWO SIDED MARKETS

Bruno Jullien

IDEI and GREMAQ, Toulouse

ESNIE - CARGESE

2

GETTING MULTIPLE SIDES ON BOARD

gamers videogame platform game developers

platform

buyers sellers

buyers

"eyeballs"

cardholders

suppliers

advertisersmerchantsdebit & credit cards

portals, newspapers, TV

B2B platform

Chicken and egg problem. Must get both sides on board/court each side while making money overall.

B2C website

3

1 INTRODUCTION

2 MONOPOLY

3 COMPETITION

4 USAGE FEES

5 COMPETITION POLICY

Organization of lecture

4

Platform enables or facilitates interaction between "buyers" and "sellers"

Buyer

Platform

Seller

usage charge usage charge

+ membership charge + membership charge

5

Some 2SPs:

Exchanges

B2B.

Employment agencies.

Dating services.

Exchanges/auctions (eBay, Amazon).

Real-estate agencies. Futures and securities exchanges

Internet backbone services.

But also...

Academic journals.

Shopping malls.

Telecoms.

Communications

6

What are two-sided markets?

Externality: Participants on one side care about the level of participation and usage of the other side

Differentiated treatment of each side The profit and the allocation depends on the structure of price not only on

the total price.

Not all platforms are 2SM

Example: electricity

Buyer GRID Producer

Bilateral contract

Only the total price charged on the two sides matters, as they negotiate how to share it: similar to tax neutrality

7

A « classical » industry may become a 2SMs

Example 1 : computers / video games

Example 2 : TV operators

Hardware producers

users Operating systemdevelopers

Content(cinema, sport…)

users operator

Advertisers

(vertical desintegration)

8

Illustration : Encoding vs. reading• Adobe Acrobat, Text Processors: free reader, charge or royalties for

encoding.

• Contrast: books.

Illustration : why did credit cards and debit cards adopt so markedly different business models?

• Credit (Visa, MasterCard, Amex): high merchant discount, low (negative) cardholder price.

• On-line debit: low merchant discount.

Often results in very skewed pricing pattern

Illustration : Videogame platforms.

• Sell console at or below cost, royalties on games

9

2.1 MONOPOLY WITH SUBSCRIPTION

For the moment no transaction fee/ cost

( )

( )B B B B S

S S S S B

N D p v N

N D p v N

demandfunctions

externalities

Network size

Platform sellersbuyersAccess cost

Registration RegistrationBp Sp

,B Sc c

10

MONOPOLY PRICES

( ) 1B B S S

B B

p c v N

p

Adjusted margin Demand elasticityfor fixed participation of the other side

Profit:

( ) ( )B B B B B S S S S S S Bp c D p v N p c D p v N

BN

( ) 1S S B B

S S

p c v N

p

Volume / margin trade-off

SN

11

Price will be low/zero/negative if

presence of buyer generates substantial revenue on seller side, buyer side reluctant to get on board (elastic demand).

Standard formula for profit maximization:

Elasticity = % variation in demand for 1% decrease in price.

Example: price to buyers.

Cost = opportunity cost, smaller than cost incurred in serving buyer

[attracting extra buyers generates revenue on seller side either through usage charges or by being able to increase sellers' membership fees.]

12

Comments :

The non adjusted margin is lower on the side where the elasticity is the highest and/or the externality created is larger.

In some cases prices may be negative (if possible, otherwise gifts, tying…) or null (free newspapers)

If one side is captive, the price is higher on this side and smaller on the other side (debit cards).

13

Other examples of skewed pricing patterns:

14

Mind the cross-group externalities

More complex story: within-side externality

Marquee buyers

Platform

Sellers

attracts

Other buyers

large fee (because marquee buyers)

good deal

Illustrations: Amex corporate card.

Killer application/game.

Key store in shopping mall.

15

Welfare: Optimal prices Unrestricted :

Price equal to the net opportunity cost → marginal cost net of the value created for

the members of the other side

B B S S

S S B B

p c v N

p c v N

16

Optimal prices

Budget constraint ( )

( )

B B S S

B B

S S B B

S S

p c v N

p

p c v N

p

Ramsey-Boiteux prices depend on elasticity and on externalities (λ is determined by the budget constraint or cost of public funds)

17

Budget balanced allocation

Ramsey prices with respect to the net opportunity cost

→ marginal cost net of the value created for the members of the other sides

Low or negative price if i) participation generates a relatively higher

externality on the other group, and

ii) the own price demand elasticity is high

18

Monopoly, summary

Competitive access (marginal cost pricing) is not efficient

One price should be below access cost (if no fixed cost), it may be negative.

Similar pattern of price skewness with unregulated monopoly and Ramsey pricing

Monopoly may be more efficient than competitive access

→ Optimal market structure?

19

3 COMPETITION

Variant 1 : single-homing bilateral

• price smaller on both sides• expectations of users play an important role (multiplicity of possible equilibria)• "divide and et conquer"

Platform 1

Platform 2

buyers sellers

20

Two identical platforms Participants register with only one

Competitive benchmark If usages can be fully taxed in a non-distortionary way and

negative registration prices are feasible, then in equilibrium Only one platform is active Zero profit

But conditions are very restrictive!

In general a positive profit equilibrium is possible, unless there is enough homogeneity within sides and coordination between sides

Single-homing and competition

21

Divide and Conquer

Divide and conquer strategies Divide: subsidies one side Conquer: charge participation of the other side

Competition generates “cross-subsidies” From the high externality group to the low externality

There is some scope for positive profit, but much less than in the case of standard network goods uniformly priced

Raise dynamic contestability by limiting the ‘first-mover advantage”

22

Divide and Conquer: example

1 buyer and 1 seller: νB = νS= ν

Platform 1 charges pB and pS>0

Platform 2 charges: pB - ν to buyer and ν to seller Profit pB- cB -cS

Eq. prices if small cost (total cost less than ν) pB= pS= cB +cS (if less than ν) Profit = cB +cS

23

Variant 2 : competitive bottleneck

Platform 1

Platform 2

buyers(single-homing)

sellers(multi-homing)

lower prices for buyers higher prices for sellers

24

Charge « monopoly prices » in multi-homing market High profits on the multi-homing side but dissipation

of these profits through to the single-homing side

Illustration: advertisers multi-home. Eyeballs don't (and even if they do, rehearsal effect). Subsidy eyeballs

Multi-homing and competition

Endogeneous MH: Easy to divide but difficult to conquer

Limits tipping by facilitating coexistence

25

4 USAGE FEES

Fees per transaction / interaction One-sided : only one sided is taxed or tax neutrality Two-sided : Non-commercial transaction, restrictive rules (payment

cards)

Usage fee affects : the probability of “trade”; the net benefits from “trade” (νB, νS ); the platform revenue

Balancing fees: set transaction fees to maximize total surplus from trade, use registration fees for coordination / revenue Same limits as for two-part tariffs: heterogeneity, risk aversion, incentive Mature platforms rely more on registration fees

Two-sided (no registration fees) : same analysis adjusting for the opportunity cost

26

(illustration: no surcharge for payments with card)

The platform as a price regulator.

The platform as a competition authority.

The platform as a licensing/certification authority

(illustrations: exchanges: solvency requirements, prohibition of front-running; dating clubs; Nintendo's mid 80s decision to control quality of third-party games)

2SP performs balancing act through other instruments than membership and usage fees:

(illustrations: auto auctions arbitration processes, eBay’s feedback forum)

The platform as a supplier of information and enforcement.

Regulation of interactions between end-users

27

5 COMPETITION POLICY

The issue is the lack of clear benchmark

Efficiency is not achieved at price equal marginal cost (or TLIC)

Efficiency may require cross-subsidies, or direct subsidyTwo violations of anti-trust: “dumping” on one side,

excessive price on the other side

28

Market definition

Changing the tariff on one side affects the demand and the profit generated on the other side: SNIP test? Estimation of demand elasticity must account for the presence of the

other side : due to feedback effects, the elasticity at fixed participation of the other side is not equal to the apparent elasticity

One or two markets ? Change the evaluation under dominance criterion Yellow pages , medias : two markets, readers and advertising M2M termination charges: two markets (origination, termination) +

regulation of termination (one market should lead to no regulation under EC rules)

Credit cards: one market with 2 sides

29

Price abuse

High price-cost margins do not imply market power even if they are low-fixed costs.

Competitive cross-subsidy Competition leads to more cross-subsidy

Competition leads to more price-discrimination

Another efficiency defence for price below costs

Predation tests: accounting for both sides→ Measure of “total price”→ Switch to effect based approach?

30

Tying as coordination device

Divide and Conquer strategy Subsidy one side

Negative prices may be not feasible Targeted offers Tie a good with registration so that registration has a

value even with no participation of the other side.

31

Indirect network effect

Possibility of coordination failure and multiple equilibria:

Solving the problem may require negative prices and price skewness

( ( ))

( , )B B B B S S S B

B B S

N D p v D p v N

N p p

32

Coordination failure: positive price

N

N

33

Solving coordination failure: one negative price

N

N

34

COMPETITION POLICY

Should we regulate?

No clear distortion

No clear guidelines for regulation

No rational for cost based regulated price

Large informational requirement

The regulatory response may be worse than the (imperfect) market response

Partial regulation (platform neutrality, reciprocal termination charge, …) ?

35

Some ReferencesNon-technical :

David Evans (2003) "Some Empirical Aspects of Multi-Sided Platform Industries," Review of Network Economics, 2: 191-209.

Rochet, J.C. et J. Tirole (2005). "Competition Policy in 2 SMs", mimeo IDEI,forthcoming "Advances in the Economics of Competition Policy".

Jullien, B (2005): “Pricing and other Business Strategies for e-Procurement Platforms”, IDEI working paper, forthcoming “Handbook of Procurement”

D. Evans, D. et R. Schmalensee (2005) “The Industrial Organisation of Markets with Two-Sided Paltforms”, NBER working paper.

36

Some References

Technical :

Rochet, J.C. et J. Tirole (2006) "Two-Sided Markets: A Progress Report", forthcoming, Rand J. Ec.

Armstrong, M. (2006) "Competition in Two-Sided Markets,“ forthcoming, Rand J. Ec.

Caillaud B. et Jullien B. (2003) “Chicken and Egg: Competition between Intermediation Service Provider“, Rand J. Ec., 34.

Jullien, B. (2005) "Two-Sided Markets and Electronic Intermediaries," CESifo Economic Studies, 51.

Rochet, J.C. et J. Tirole (2003) “Paltform Competition in Two-Sided Markets”, Journal of the European Economic Association