Date post: | 10-Apr-2015 |
Category: |
Documents |
Upload: | alexia-gaudeul |
View: | 283 times |
Download: | 2 times |
Network effects and two sided marketsOFT lecture - 16th of January, 2007
!Alexia GaudeulUniversity of East Anglia and
ESRC Centre for Competition Policy
Plan of the presentation
! Two parts: Network effects and Two-sided markets
! In each part
Definition
Motivation / Application
Models with basic concepts and logic
Policy implications
Review of required reading
Method of presentation
! Theoretical models to understand the mechanisms ofcompetition in the specific setting under study.
! Application of theory to draw out policy implications
! Work required:
Work out models on your own to assimilate their logic.
Read the required readings!
Network effects
! Def: A change in the benefit, or surplus, that an agent derives from agood when the number of other agents consuming the same kind of goodchanges.
! Spreadsheets (Brynjolfsson and Kemerer 96):
1% increase in installed base means a 0.75% increase in price
Products that adhere to the standard get 45% higher price
! Network equipments (Forman 01 and Forman and Chen 03)
Firms are able to exploit cross-product and cross-firm switching costs toinduce strong legacy effects and lock-in.
Compatibility is a strategic variable in firms’ product design
! DVD vs DIVX war (Dranove and Gandal 03)
Rational expectations and multipleequilibria: Katz and Shapiro (85)
! Duopoly with two incompatible products.
! Consumer of type z gets utility v(qi)+z from network i=1,2 of size qi.
! Net utility: u(z)=v(qi)+z-pi
! Net price: Pi=pi-v(E(qi))
! All consumers choose the network which net price is the lowest. For bothfirms to survive, I must thus have P1=P2=P
! The choice thus depends on the expectations of the consumer.
! If z is uniform over [0,1], then
z>P buy one of the goods and
Total demand is q=1-P= q1+q2=1-pi-v(E(qi)), i=1,2
! Profit is then qi(pi-c). One obtains reaction functions and a Nashequilibrium, with the requirement that E(qi)=qi in equilibrium (rationalexpectations).
Rational expectations and multipleequilibria (suite)
! Competing for the market, or competing in the market(standardization)?
! What drives expectations?
Chicken and egg problem: low price to encourage adoption, highprice to exploit adoption...
! Pricing for progressive adoption (subsidy then exploitation ofthe user base)
Do consumers anticipate this potential lock-in problem?
Coordination and lock-in(Farrell and Klemperer, 2006)
! Consider two users who consider two technologies, old or new.
The old technology provides utility u(q) if q=0,1,2 join, while the newtechnology provides utility v(q) if q=0,1,2 join.
! Assume u(2)>u(1) and v(2)>v(1): there are positive network externalities.
! Assume also that u(2)>v(1) and v(2)>v(1) (it is preferable that both userscoordinate on the same network.
! There will be excess inertia if v(2)>u(2) and both users stick to the oldnetwork for fear the other may not follow.
! There will be excess momentum if u(2)>v(2) and both users switch to thenew network for fear not to have followed.
Coordination and lock-in (suite)
! Path dependence:
Initial conditions in the market determine its future.
Initial conditions are greatly influenced by random factors
! Patterns of adoption depending on strength of preferencesfor one or the other network
Early adopters and followers (bandwagon)
Compatibility problems on thesupply side: Besen and Farrell (94)
! Two firms have the choice to make their product compatibleor incompatible.
If products are incompatible, then each firm has an equalprobability to win the whole market and make monopoly profit M.
If firms choose compatibility, then they will both make duopolyprofit D.
Typically, M>2D , as competition reallocates some social surplusfrom the firms to the consumers.
Compatibility problems on thesupply side (suite)
! Consider first a compatibility game that occurs over two periods.
In the first period, firms choose compatibility or no compatibility.
In the second, profits are made.
Since M>2D , firms prefer no standard and probability _ to win in thesecond period.
! Consider now a modified game whereby it is the firm that spends themost in the first period that wins the standards war.
Then, if incompatibility is chosen in the first period, one firm preemptsthe other and spends M so the other spends 0 and the expected profit is 0for both.
Therefore, firms will choose compatibility.
Compatibility problems on thesupply side (suite)
! Shapiro and Varian (1999) “The Art of Standards Wars”
"Evolution" strategy (new technology compatible with old, egColor TV)
"Revolution" strategy (incompatibility, eg DTV)
"Rival Evolutions" (compatible with old, incompatible with othernew, eg DVD vs DIVX)
! See also Besen and Farrell (1994), “Strategies and tactics instandardization”.
Policy issues
! Fragmented vs. dominant network
Ex-post and ex-ante efficiency
When to intervene? (DTV, GSM)
! Network as essential facility
The problem of bottlenecks (eBay, Microsoft)
How far can one prevent access to one’s network?
! Network tipping
Abuse of dominance, eg Microsoft
Problem if consumers are myopic
Required reading: summary
! Katz and Shapiro (1994): “Systems competition and networkeffects”
! Divergence from social optimum is possible (monopolies,externalities)
! However:Markets may self regulate.
Government may have wrong incentives, biased in favor ofexisting users.
Lack of information to impact the outcome favorably.
Two sided markets
! “(M)arkets in which one or several platforms enableinteractions between end-users, and try to get the two (ormultiple) sides “on board” by appropriately charging eachside” (Rochet and Tirole, 2004)
Two sided markets
Auctions: eBay" Buyers and sellers benefit from variety and competition on the otherside.
Payment systems: credit card networks" If merchants multi-home, credit card holders need not do so.
Shopping malls" Ability to select participants on one’s own side (upscale, downscale)
Video game consoles: Sega vs Nintendo" Platform for development, attract investment.
Media" Readers dislike advertisement, but like lower prices
Match-making" Who does the choice?
Ownership in two sided markets(Gaudeul and Jullien 2001)
! Consider two sides, A and B, of a market, who can contact each otheronly through an intermediary.
Side A receives surplus fN when N members of side B join, while
side B receives surplus Fn when n members of side A join.
Prices for joining are set by the intermediary at p for side A and P forside B.
! Net profit for the intermediary is (p-c)n+(P-c)N, assuming marginal costof operation is a constant c, the same for both sides.
! Total welfare is W=(f+F)Nn-a(n)-A(N)-c(N+n)
a(n) and A(N) are the sum of access costs for all participants in themarket
Participants may be more or less eager to join, and this may depend onhow many joined already (gregariousness and then congestion)
Ownership in two sided markets(suite)
! State-owned subsidized monopoly intermediary
! Not-for profit monopoly intermediary
! For-profit monopoly intermediary
! Competing intermediaries with global database of clients
! Competing intermediaries with proprietary database ofclients
Divide and conquer strategy
Policy issues
! Below-cost pricing must be judged by considering both sidesof the market. Not necessarily predatory.
! Ownership of intermediaries, of databases, and ability tomulti-home must be considered.
! Buyer-side and seller-side intermediaries
Possible limitations in choice
Market design becomes a private matter
Required reading: summary
! Wright (2004) “One sided logic in two-sided markets”
Prices may not reflect costs
Above cost prices may be sustained in a competitive equilibrium
Below cost prices do not indicate predation
Competition may decrease efficiency
Imbalanced prices for the two sides does not indicate cross-subsidy (i.e. exploiting strength in one market to obtain dominancein the other). In fact, each side may benefit from an imbalancedpricing structure.
Conclusion
! Interdependence of the participants in a market.
within one’s own side (network effects)
vis-à-vis the other side (two-sided markets)
! Consumers are limited by the choice made by otherconsumers, or by the intermediary.
! Principles for the regulation of intermediaries are lacking.
! The effect of subtle changes in market rules is not wellunderstood
Need for empirical work, focus on Internet markets.