산업분석의 틀과 응용 SSNIP Test for Two-sided Markets 가격상승압력 분석(UPP, Upward Pricing Pressure) 3-4 한성대학교 경제학과 이내찬

양면시장이란? 양면시장 간접네트효과 다른 두 객체를 연결함으로써 수익을 얻는 시장 네트워크 효과는 이용자 수가 많아지면 개개인 이용자의 효용이 증가하는 현상을 의미하는데 양면 시장에서는 이와 같은 효과가 다른 객체에 대해 그 수가 증가하면 발생하는데 이를 ‘간접네트워크효과(indirect network effect)’라 한다. 이 효과는 양 객체 간에 피드백을 통해 발휘되는데 이용자는 양질의 콘텐츠를 많이 보유한 포털을 선택하고 광고주는 광고수익을 올릴 수 있는 보다 많은 이용자가 이용하는 포털을 선택하게 되는 선순환이 발생하게 된다. 반면에 이용자는 쓸 만한 콘텐츠가 없는 포털은 이용하지 않으려 하고 광고주는 이용자 수가 작은 포털에 광고를 내려 하지 않게 되는 그래서 결국 이용자나 광고주 수가 줄어 들게 되는 악순환이 발생하게 되면 포털은 성장하지 못하게 된다.

2-sided markets: First key factor 6/17/2018 2-sided markets: First key factor In a typical 2-sided market: DA is function not only of pA but also of qB DB is function not only of pB but also of qA What does this mean for market definition? SSNIP test: Could a hypothetical monopolist profitably sustain a small but significant and non-transitory increase in price? If no, the relevant antitrust market must be wider If yes, the market is not any wider (but could be narrower) Key question for 2-sided markets: Which feedback loops should be taken into account?

피드백 효과 앞서 단면시장에서는 SSNIP test에 의해 관련시장을 획정한다고 했다. 그러나 이 논리를 양면시장의 단면에만 적용시킬 경우 우려는 양면의 객체 간 간접네트워크효과의 피드백을 반영하지 못하면 관련시장의 획정에 오류가 발생할 수도 있다는 것이다. 클럽의 예로 이를 살펴보자. 클럽의 SSNIP test에 의해 여성 클럽 입장료를 인상한 결과 이윤이 증대되었다고 하자. 그림에서 이를 사업자 a라고 하면 사업자 b는 예를 들어 노래방과 같이 기능과 효용이 상이한 서비스이고 따라서 사업자 a만이 관련시장을 형성한다. 그러나 이와 같은 결론을 내리기 전에 간과해서는 아니 될 것이 간접네트워크효과의 피드백이다. 즉, 여성의 클럽 입장료의 인상으로 인해 여성 수요가 줄면 이는 곧 남성 수요의 감소를 야기하여 남성 클럽 입장료 수익의 감소를 야기하며 여성도 남성에 대한 선호도가 없는 것이 아니므로 여성 수요도 감소하게 하여 여성 클럽 입장료 수익의 감소를 야기한다는 것이다. 이런 피드백이 부정적으로 작용되다 보면 경우에 따라서는 단면만 보았을 경우에 비해 이윤이 증대되는 정도는 작을 수도 있으며 경우에 따라서는 여성 클럽 입장료 인상에 따른 이윤은 증대가 아니라 감소가 될 수도 있다. 그렇다면 관련시장은 사업자 a뿐만이 아니라 사업자 b도 포함되어야 한다는 것이다. 그러나 이 경우라도 사업자 a, b이 자체적인 간접네트워크효과뿐만이 아니라 서로에게 영향을 주는 피드백까지 감안하여야 하므로 분석은 점차 복잡해질 수밖에 없다.

The SSNIP Test: Feedbacks 6/17/2018 Positive externalities pA1↑ qA1↓ πA1↑ πA1↓ πA1↑ Normal Market: Does total πA↑? If yes, market not wider

The SSNIP Test: Feedbacks 6/17/2018 Positive externalities For given p1B pA1↑ qA1↓ πA1↑ πA1↓ qB1↓ πA1↑ Normal Market: Does total πA↑? If yes, market not wider Feedback to market B πB↓

The SSNIP Test: Feedbacks 6/17/2018 Positive externalities For given p1B pA1↑ qA1↓ πA1↑ πA1↓ qB1↓ Should we take this into account? πA1↑ Normal Market: Does total πA↑? If yes, market not wider Feedback to market B πB↓

The SSNIP Test: Feedbacks 6/17/2018 Positive externalities For given p1B pA1↑ qA1↓ πA1↑ πA1↓ qB1↓ qA↓ πA1↑ Normal Market: Does total πA↑? If yes, market not wider Feedback to market B Feedback to market A πB↓ πA↓

The SSNIP Test: Feedbacks 6/17/2018 Positive externalities For given p1B pA1↑ qA1↓ Should we take this into account? πA1↑ πA1↓ qB1↓ qA↓ πA1↑ Normal Market: Does total πA↑? If yes, market not wider Feedback to market B Feedback to market A πB↓ πA↓

The SSNIP Test: Feedbacks 6/17/2018 Positive externalities For given p1B pA1↑ qA1↓ and so on πA1↑ πA1↓ qB1↓ qA↓ … πA1↑ Normal Market: Does total πA↑? If yes, market not wider Feedback to market B Feedback to market A πB↓ πA↓

The SSNIP Test: Feedbacks with Differentiated Products 6/17/2018 Positive externalities For given p2A , p1B ,p2B qA2↑ pA1↑ qA1↓ πA1↑ πA1↓ qB1↓ qA1↓ … πA1↑ Normal Market: Does total πA↑? If yes, market not wider Feedback to market B Feedback to market A πB↓ πA↓

The SSNIP Test: Feedbacks with Differentiated Products 6/17/2018 Positive externalities For given p2A , p1B ,p2B qB2 ↑ qA2↑ pA1↑ qA1↓ πA1↑ πA1↓ qB1↓ qA1↓ … πA1↑ Normal Market: Does total πA↑? If yes, market not wider Feedback to market B Feedback to market A πB↓ πA↓

The SSNIP Test: Feedbacks with Differentiated Products 6/17/2018 Positive externalities For given p2A , p1B ,p2B qB2 ↑ qA2↑ Should we take these into account? pA1↑ qA1↓ πA1↑ πA1↓ qB1↓ qA1↓ … πA1↑ Normal Market: Does total πA↑? If yes, market not wider Feedback to market B Feedback to market A πB↓ πA↓

The SSNIP Test: Feedbacks with Differentiated Products 6/17/2018 Positive externalities For given p2A , p1B ,p2B qB2 ↑ qA2 ↑ qA2↑ pA1↑ qA1↓ πA1↑ πA1↓ qB1↓ qA1↓ πA1↑ Normal Market: Does total πA↑? If yes, market not wider Feedback to market B Feedback to market A πB↓ πA↓

The SSNIP Test: Feedbacks with Differentiated Products 6/17/2018 Positive externalities For given p2A , p1B ,p2B qB2 ↑ qA2 ↑ qA2↑ Should we take these into account? pA1↑ qA1↓ πA1↑ πA1↓ qB1↓ qA1↓ πA1↑ Normal Market: Does total πA↑? If yes, market not wider Feedback to market B Feedback to market A πB↓ πA↓

The SSNIP Test: Feedbacks with Differentiated Products 6/17/2018 Positive externalities For given p2A , p1B ,p2B and so on qB2 ↑ qA2 ↑ qA2↑ … pA1↑ qA1↓ πA1↑ πA1↓ qB1↓ qA1↓ … πA1↑ Normal Market: Does total πA↑? If yes, market not wider Feedback to market B Feedback to market A πB↓ πA↓

The SSNIP Test: Feedbacks with Differentiated Products 6/17/2018 Positive externalities For given p2A , p1B ,p2B and so on qB3 ↑ … qA2↑ qB2 ↑ qA2 ↑ … pA1↑ qA1↓ πA1↑ πA1↓ qB1↓ qA1↓ … πA1↑ Normal Market: Does total πA↑? If yes, market not wider Feedback to market B Feedback to market A πB↓ πA↓

The SSNIP Test: Feedbacks with Differentiated Products 6/17/2018 Positive externalities For given p2A , p1B ,p2B and so on qB3 ↑ … qA2↑ qB2 ↑ qA2 ↑ … pA1↑ qA1↓ πA1↑ πA1↓ qB1↓ qA1↓ … πA1↑ Normal Market: Does total πA↑? If yes, market not wider Feedback to market B Feedback to market A πB↓ πA↓

Positive externalities The SSNIP Test: Feedbacks with Differentiated Products & Different Number of Potential Products 6/17/2018 Positive externalities For given p2A , p1B ,p2B and so on qB3 ↑ … qA2↑ qB2 ↑ qA2 ↑ … pA1↑ qA1↓ πA1↑ πA1↓ qB1↓ qA1↓ … πA1↑ Normal Market: Does total πA↑? If yes, market not wider Feedback to market B Feedback to market A πB↓ πA↓

Example: A merger in UK grocery sector Traditional method in retail Draw a circle to define the relevant market Calculate market shares and HHI for merging parties But some obvious problems Rather crude 0/1 definition of rivals (cf Sainsbury’s) Those stores differs in f.ex. product range Why not directly measure rivalry between Morrison and Somerfield? http://blogg.nhh.no/beccle/wp-content/uploads/2012/11/Mergerscreening-13112012.pdf

Diversion ratios – A direct measure of rivalry Are they fighting for the same consumers? Must check whether consumers regard those two stores as close substitutes (large overlap) Can detect this by measuring consumers’ second choice Would shoppers at Morrisson have Somerfield as their best alternative, and vice versa? Diversion ratios (DR) measures the overlap If 45 % of those leaving Morrisson goes to Somerfield, then the diversion ratio is 45 % This new approach a sound theoretical foundation

Acquisition of Morrisson in 2005 in the UK Competition Commission in UK used a survey among shoppers to estimate DR Shoppers outside Morrisson: Where would you have shopped if this store was closed? Anti-competitive concern if large diversion ratio to Somerfield Somerfield would pick up much sales diverted from Morrisson An upward pricing pressure on Morrisson store after merger But what is a ’large diversion ratio’? 15 % 45 % 20 % 10 % Others 5

The information needed for merger screening Price pressure upward/downward? Downward: Lower marginal costs Upward: Large value of diverted sale Large diversion ratio to other merging product High margin on recaptured units An upward pricing pressure (UPP) if: 𝑈𝑃𝑃 1 = 𝐷 12 𝑃 2 − 𝐶 2 − 𝐸 1 𝐶 1 >0 Symmetric Case  𝐷>𝐸 𝐶 𝑃−𝐶 =𝐸 1 −𝐿 𝐿 http://faculty.haas.berkeley.edu/shapiro/uppcritical.pdf

Derivation http://www.oecd.org/daf/competition/Marketdefinition2012.pdf

Old versus new approach Market shares no longer of importance Focus directly on overlap between merging parties products Diversion ratios Margins Other factors of importance in the final assessment Efficiencies Repositioning Entry barriers Others

Incorporated into merger guidelines US merger guidelines August 2010: ’The Agencies rely much more on the value of diverted sales than on the level of HHI for diagnosing unilateral price effects in markets with differentiated products’ ’Diversion ratios between [merging firms’ products] can be very informative for assessing unilateral price effects’ UK merger guidelines September 2010: ’The combination of diversion ratios and gross profit margins can give a strong indication of unilateral effects. These two factors together help quantify the change in the merged firm’s incentive to raise its prices or worsen its non-price offers.’

에실로아메라인베스트먼트의 경쟁제한적 기업결합 행위에 대한 건 에실로아메라인베스트먼트의 경쟁제한적 기업결합 행위에 대한 건  공정거래위원회(위원장 노대래)는 시력교정용 안경렌즈 세계 및 국내 1위 업체인 Essilor Amera Investment PTE. LTD*(이하 에실로)가 국내시장 안경렌즈 2위인 대명광학(주) (이하 대명광학)의 주식을 취득하는 건이 경쟁을 실질적으로 제한할 우려가 있다고 인정하여 이를 불허하기로 결정함. http://www.ftc.go.kr/news/ftc/reportView.jsp?report_data_no=5565