A SSNIP Test for Two-Sided Markets: some theoretical considerations Lapo Filistrucchi Tilburg University, CentER & TILEC & Univeristy of Siena “The Economics.

A SSNIP Test for Two-Sided Markets: some theoretical

considerations

Lapo Filistrucchi

Tilburg University, CentER & TILEC

&Univeristy of Siena

“The Economics of Payment Systems”,

Paris, October 2007

Market definition Market definition is normally the first

stage in competition policy analysis

Definition of the relevant market is crucial for cases of abuse of dominance or mergers:

A wrong definition might lead to blocking a welfare enhancing merger or to allowing a welfare detrimental one.

A wrong definition might lead to sanctioning behaviours which should not be sanctioned or to allow abuses which should be sanctioned

A wrong definition of the relevant market is sufficient for the courts to rule in favour of the parties irrespective of any other argument brought up by the antitrust authority

Market definition

Economic models usually assume the existence and size of the market (e.g. “Consider a market with 2 goods…”)

Competition policy analysis has therefore developed the Small Significant Non-transitory Increase in Price test (SSNIP test) and the Critical Loss Analysis (CLA).

The SSNIP Test The SSNIP test in theory:

Suppose three potential goods in a market: 1,2 and 3

Could an hypothetical monopolist of product 1 profitably sustain a small but significant and non-transitory increase in price? (idea..)

If yes, the market includes only 1.

If no, the relevant antitrust market must include also 2 or 3. Which one? The one with the highest cross price elasticity with respect to 1… Suppose it is 2.

Could an hypothetical monopolist of product 1 and 2 profitably sustain a small but significant and non-transitory increase in the price of both?

and so on…

Critical Loss Analysis The CLA in theory:

Suppose three potential goods in a market: 1,2 and 3

Consider an hypothetical monopolist of product 1

Calculate for a given price increase the maximum quantity in sales that can be lost by the hypothetical monopolist before the price increase becomes unprofitable

Estimate loss in quantity for the hypothetical monopolist following the given price increase.

If estimated loss is lower than the critical loss, the price increase would not be unprofitable and the relevant market should include only product 1.

If estimated loss is higher than the critical loss, the price increase would be unprofitable, so that the relevant market should include also product 2 or 3…

and so on…

The SSNIP Test So, empirically:

Estimate own and cross price elasticities of demand for all potential products 1,2 and 3.

Simulate profits of hypothetical monopolist (costs!)

Key question:

Which starting value for increase in prices? (Cellophane fallacy- du Pont case)

The SSNIP Test in 2-Sided Markets Key questions for 2-sided markets:

Which price should the hypothetical monopolist be thought of as rising?

Which profit changes and feedbacks should be taken into account between the two-sides of the market?

Two types of two-sided market: the “media type” and the “payment card type”

See Armstrong (2006) and Rochet & Tirole(2006)

A Two-Sided Market: Media

Media Firm

reader/viewer/listener

advertiser

Newspapers, TV, Radio, Internet…

ad feesad slot

price for content

advertising message

media content

Note: noper-transaction fees

Note: noper-transaction fees

A Two-Sided Market: Payment Cards

payment card scheme (3-party)

buyer sellergood

card service

Also: auction house, operating systems

merchant feescard

service fixed fee+per-trasaction fees

fixed fee+per-transaction fee

card-holders fees

Market Definition in 2-Sided Markets: the Literature

Argentesi & Ivaldi (2005) – media, elasticities

Emch & Thomson (2006) - payment cards, raise price level, usage only

Evans & Noel (2005a) – critical loss analysis

Evans & Noel (2005b) - some firms on one side of the market only

Holland (2007) – two-sides of the market

Carlton (2007) – price problem

The SSNIP test: Price - 1

In a two-sided market one can distinguish

a) the price level (the sum of the two prices)

b) the price structure (the ratio of the two prices)

Should the hypothetical monopolist be thought of as raising:

- both prices together keeping fixed the price structure?

- first one of the two prices keeping the other fixed and then the other price keeping the first fixed?

- the price level adjusting optimally the price structure?

- first one of the two prices and then the other price each time adjusting optimally the price structure?


If the hypothetical monopolist raises both prices together keeping fixed the price structure:

- how do you interpret the test in terms of “enough substitution”?

Not much sense…


If the hypothetical monopolist raises first one of the two prices keeping the other price fixed and then the second price keeping the first price fixed:

- easy to interpret in terms of substitution

- how do you make sure the market definition on on-side is consistent with the one used to define the market on the other side?

- why should the hypothetical monopolist not be allowed to adjust the price structure optimally?

- too wide market definition if all feedbacks considered (see below)

Maybe…


If the hypothetical monopolist raises the price level adjusting optimally the price structure:

- right exercise to check market power

- easy to do in a “payment card type” two-sided market

but

- how do you implement it in a “media type” two-sided market?

- how do you interpret the test in terms of “enough substitution”?

Maybe…


If the hypothetical monopolist raises first one of the two prices adjusting the second optimally and then the second adjusting the first optimally:

- easy to interpret in terms of substitution

- smaller market definition if all feedbacks considered (see below)

- how do you make sure the market definition on on-side is consistent with the one used to define the market on the other side?

Probably the best

),()(max,

sbsb

pp

ppTcppsb

T

p

b

bs

bp

cpp

1

T

p

s

bs

sp

cpp

1

T

p

T

p

bs

bs

bspp

cpp

1

Rochet & Tirole (2006)

The Hypothetical Monopolist: Payment Cards (Use)

Some mathematical intuition:the profit maximizing price level is determined by the total elasticity of transactionsprice structure is then determined by maximising number of transactions given

sb

sb

pp

ppp

ts

ppTsb

*

,

..

),(max

sb ppp *

*p

The Hypothetical Monopolist: Media or Payment Cards (Adoption)

),(max,

sbssbb

FF

NNCNFNFsb

),(

),(

),(

),(sbss

sbbb

sbss

sbbb

FFnN

FFnN

FNNN

NFNNIf

andF

nthen

b

b

...

Armstrong (2006)

)),(),,((),(),(max,

sbssbbsbsssbbb

FF

FFnFFnCFFnFFFnFsb

b

b

s

bb

b

n

F

bb

ssn

Fs

ss

n

F

b

bb

nF

nF

F

cF

F

cF

11

s

s

b

ss

s

n

F

ss

bbn

Fb

bb

n

F

s

ss

nF

nF

F

cF

F

cF

11

...

b

s

s

ss

b

bb

n

n

F

cF

F

cF ...b

sssbb

n

ncFcF

bbn

F

n

F

bbn

F

n

F

ssn

F

n

F

bbn

F

n

F

n

F

b

bb

nFnF

nFnF

F

cFb

s

s

b

b

b

s

s

b

b

s

b

b

b

s

s

b

b

1

ssn

F

n

F

ssn

F

n

F

bbn

F

n

F

ssn

F

n

F

n

F

s

ss

nFnF

nFnF

F

cFb

s

s

b

s

s

b

b

s

s

b

s

s

s

b

b

s

s

1

The SSNIP Test: Feedbacks

pA1↑ qA

1

↓

πA1↑ πA

1↓

πA1↑

Normal Market:Does total πA↑?

If yes, market not wider

Positive externalities


pA1↑ qA

1

↓

qB1↓

πA1↑ πA

1↓

πB↓

For given p1B

πA1↑



Feedback to market B



pA1↑ qA

1

↓

qB1↓

πA1↑ πA

1↓

πB↓

For given p1B

πA1↑




Should we take this

into account?



pA1↑ qA

1

↓

qB1↓

πA1↑ πA

1↓

πB↓

For given p1B

πA1↑



Feedback to market A



qA↓

πA↓


pA1↑ qA

1

↓

qB1↓

πA1↑ πA

1↓

πB↓

For given p1B

πA1↑





Should we take this

into account?


qA↓

πA↓


pA1↑ qA

1

↓

qB1↓

πA1↑ πA

1↓

πB↓

For given p1B

πA1↑






and so on

qA↓

πA↓

…

My answer is:

Yes, all of them!

Because feedback loops only relevant if original substitution effect is there.

Antitrust authorities worried that:

If positive externalities and loops considered, market much wider than single-sided market

But that is exactly the point: it is a two-sided market!

True that benchmark for “enough substitution” is in some sense changed (but still the “hypothetical monopolist”)

True that also time-dimension to the feedbacks story (but here “non-transitory”)


The SSNIP Test: Feedbacks with Differentiated Products

pA1↑ qA

1

↓

qB1↓

πA1↑ πA

1↓

πB↓

For given p2A , p1

B ,p2B

πA1↑






qA1

↓

πA↓

qA2

↑

…


pA1↑ qA

1

↓

qB1↓

πA1↑ πA

1↓

πB↓

For given p2A , p1

B ,p2B

πA1↑






qB2 ↑

qA1

↓

πA↓

qA2

↑

…


pA1↑ qA

1

↓

qB1↓

πA1↑ πA

1↓

πB↓

For given p2A , p1

B ,p2B

πA1↑





Should we take these

into account?


qB2 ↑

qA1

↓

πA↓

qA2

↑

…


pA1↑ qA

1

↓

qB1↓

πA1↑ πA

1↓

πB↓

For given p2A , p1

B ,p2B

πA1↑






qB2 ↑

qA1

↓

πA↓

qA2

↑qA

2 ↑


pA1↑ qA

1

↓

qB1↓

πA1↑ πA

1↓

πB↓

For given p2A , p1

B ,p2B

πA1↑





Should we take these

into account?


qB2 ↑

qA1

↓

πA↓

qA2

↑qA

2 ↑


pA1↑ qA

1

↓

qB1↓

πA1↑ πA

1↓

πB↓

For given p2A , p1

B ,p2B

πA1↑





Positive externalitiesand so on

qB2 ↑

qA1

↓

πA↓

qA2

↑qA

2 ↑ …

…

My answer is:

Yes, all of them!

Because feedback loops only relevant if original substitution effect is there.

Antitrust authorities worried that:

If positive externalities and loops considered, market much wider than single-sided market

But that is exactly the point: it is a two-sided market.

The SSNIP Test: Feedbacks with Differentiated products


pA1↑ qA

1

↓

qB1↓

πA1↑ πA

1↓

πB↓

For given p2A , p1

B ,p2B

πA1↑





Positive externalities and so on

qB2 ↑

qA1

↓

πA↓

qA2

↑ qA2 ↑ …

…

…


pA1↑ qA

1

↓

qB1↓

πA1↑ πA

1↓

πB↓

For given p2A , p1

B ,p2B

πA1↑






qB2 ↑

qA1

↓

πA↓

qA2

↑ qA2 ↑ …

…

qB3 ↑

…


pA1↑ qA

1

↓

qB1↓

πA1↑ πA

1↓

πB↓

For given p2A , p1

B ,p2B

πA1↑






qB2 ↑

qA1

↓

πA↓

qA2

↑ qA2 ↑ …

…

qB3 ↑

…

The SSNIP Test: Feedbacks with Differentiated Products & Different Number of Potential Products

pA1↑ qA

1

↓

qB1↓

πA1↑ πA

1↓

πB↓

For given p2A , p1

B ,p2B

πA1↑






qB2 ↑

qA1

↓

πA↓

qA2

↑ qA2 ↑ …

…

qB3 ↑

…

My answer is again:

Yes, all of them!

However, the feedbacks depend on the products on the other side

So that

either we take into account feedbacks from all potential products

or market definition on one side depends on market definition on the other side

The SSNIP Test: Feedbacks with Differentiated products

Conclusions In extending the SSNIP test to two-sided markets, there are 2 key questions:

Which price should the hypothetical monopolist be thought of as rising?

Which profit changes and feedbacks should be taken into account between the two-sides of the market?

One should remember that the logic behind the SSNIP test is identifying a threshold for substitution to be relevant enough to include two products in the same market.

Profits changes on both sides of the market and all feedbacks should be considered

First one of the two prices should be raised and then the other price each time adjusting optimally the price structure?

Future research 1Apply the proposed SSNIP test to the newspaper marketData

Readers’ side: monthly observations by day of the week on circulation, cover prices and content characteristics of 7 daily newspapers in Italy for 31 years (market level data).

Advertisers’ side: monthly observations by day of the week on advertising quantity, prices and readers characteristics on those same daily newspapers for 21 years (market level data).

The paper will test empirically the traditional classification of newspapers into general interest, financial and sport originally devised by the Italian Federation of Newspapers Publishers (F.I.E.G.) and used by the Italian Antitrust Authority in a complaint investigated some years ago (Decision 3354/95 "Ballarino vs Grandi Quotidiani").

Future research 2 Apply the proposed SSNIP test to flower auction house market and/or the payment cards market Data

being collected

The papers will test empirically for the geographic market definition (auction houses) and for the product market definition (payment cards)

A SSNIP Test for Two-Sided Markets: some theoretical considerations Lapo Filistrucchi Tilburg University, CentER & TILEC & Univeristy of Siena “The Economics.

Documents