Management Engineering - Business & Industrial Economics

Network Economics

Divided by topic

Network economics: the hardware/software paradigm, winner -takes -all - markets and technological standard wars. We observe after a bit of competition, at a certain point a firm take the lead (eBay in this case) in a market that gained lot s of importance, the other instead lose pace. Still today eBay is the leader of the market, even if the market decreased in terms of importance. This is very typical in the network market, e.g., ICT. This evolution in terms of competitive dynamics into the market cannot only be ascribed to the fact that eBay was better performing its business than the other. It’s very typical of network market, due to its inner characteristics. Structural characteristics of the market that force the firms to put in action so me action, strategies that lead to certain performances and in terms of performance of the whole market. This type of dynamics is observable also in the above 2 examples. Smartphone OS → incredible gap compared to the other market . The car industry in w hich in terms of size are more comparable to the gap. * example explained below In terms of the dichotomy of performance comparing the market leader and others. Winner takes all market → firm able to become leader takes all the market, most of it. it leaves only pinouts to all the others. No second -best prize medal because firms compete only for the first position, the gold medal. An introductive example : The story of the garbage bin GARBAGE BIN •In the town there are six guys that like playing videogames, Alan, Bud, Charlie, David, Eliah, and Frank. •Their utility function is characterized by externalities: their preferences depend on the intrinsic value of the game console plus a bonus that depends on the number of other guys owning the same type of console (why? because they can exchange games, challenge friends, and have a greater variety of games). Ui= utility function w measures the strengths of this network externality. Utility function increase as Nk increase s. They might prefer other consoles in principle. •k = a, b, c (3 different consoles) Each firm invested in R&D for its own console and managed to patent some product features. Each firm also invested in marketing and advertising. STANDARD WAR: war among competitors that want to sell the products at the expense of the product of other competitors. All expenses are not re -deployable investments. If they convinc e people to buy the console it’s fine, otherwise , the value will decrease a lot. •Yesterday, the guys received the catalogues, illustrating technical features •Therefore, tonight the guys will read the catalogues, assign their preferences (intrinsic value X k) and make their consumption decisions. •They will assign different preferences to different consoles since they value different characteristics. What is the market structure (potential) on the Sunday night when they fall asleep? 2 A, 2 B, 2 C •Charlie’s cat is very hungry on the Monday morning and happens to turn the garbage bin upside down. •Charlie wakes up suddenly, it is 8 am. The other guys in town are still sleeping. •Charlie thinks “since I am awake, I might go to the games shop and buy my console”. •He goes and his purchase decision is obviously to buy C. This little move by Charlie, because of the presence of Network externalities, change all the context. Now console C has 1 user. Increase of other p references of 0.2. Initially , the preference was based only on the intrinsic value, now also the network effect is considered. At the end of the way the ultimate reason behind this very peculiar situation is due to the process we have seen in this simplif ied example of the garbage bin. The story is simplified; therefore, we have some questions… Open questions •What if Charlie was the dull guy in town? (C is the worst game console) → in the real world this is rare. •What about companies A & B? → they lost their battle and probably also the war. They should open a new war/battle through innovation (radical, a breakthrough) in order to take a revenge compared to the leader, where everyone starts again from the same level. •Are companies happy with such a high -risk profile of competition? (Winner takes it all) → not happy with high -risk profile competition. One typical way is by performing strategies of coopetition , a mixture of collaboration and competition . To avoi a high - risk profile, they try to collaborate in the first place, ex -ante entering the market, ex - ante competition, in order to come out with the same technological standard. Afterward,s they compete to affirm their on product at the expense of other competitors. WTA /WTM: Winner Take All/Most. FURTHER CONSIDERATION Of course, there are missing elements in the story, but you can already infer the role of past and casual events in today’s situation: How can we define business processes where a cat and a garbage bin are so important for determining the final outcome of the market? This typology of market can be defined as path dependency. We have to analyse the initial condition of the market. Only if we understand what happened at the beginning of this process, we may understand why we ended up in the situation we are at the end. These processes envisage what in chaos theory is called the butterfly effect: small modifications /changes in the initial conditions may lead to a compl etely different outcome. of the nature Typical characteristics of a network market, and the dynamics that affect competition in this market, exploring and exploding the concepts above. Our focus : first 3 bullets (4 th type refers to production) Network externality Definition of network externality : a good exhibits a network externality when the positive change in the utility a consumer derives from it raises as the number of consumers that purchase the same product increases. Examples: telephone, email, hardware -software (all the goods belonging to them. e.g. opera ting system, software run by the OS), party,etc. (the first 3 examples are linked to the ICT word) Difference between direct vs. indirect network externality: Direct externality: the value of the good increases automatically as the number of users increase s (e.g. phone or e -mail). Generally, the good has no intrinsic value. Indirect externality: the value of the good increases as the number of users increases but only in the presence of economies of scale in production (of software) because of a greater off er of complementary products (or a better quality) of the good (admitting that consumers care for software variety). In this case, the good has often an intrinsic value. We can have both negative and positive externality but we refer to the positive one, because it is the most interesting one. Distinction between a “two -way” and a “one -way” network (Economides, 1996, IJIO) Direct 2-way network: all the networks where service AB is different from the service BA (they represent two different goods, that can be, therefore, priced differently). Each knot of the network represents a user: The value of the network is a function of the active links inside the network. In a 2 -way network composed by n knots, there are n(n -1) potential links. The entry of a new user produces a positive externality on existing users, since she adds 2n new potential links. (e.g. TLC network made of 10 people: …) We have an empirical rule to estimate the value of a network: Metcalfe’s Law : if a network is composed by n users and each user assigns a value to the network which is proportional to the number of users. → the value of the network V = f(n 2-n) → If n is large (it means that n 2>>n) → V = f(n 2) Excursus (A continuous formulation of Metcalfe’s Law –Rohlfs 2003, p. 215) f(n)=n (it’s a linear function) v(i)=c (it’s a constant) Metcalfe’s law ➔ Relationship between number of users and value of the network implied by the law ➔ True relationship In the 2 nd picture we have a much lower increase of VR when n is high. Here the 2 nd derivative is negative instead of positive. Distinction between a “two -way” and a “one -way” network (Economides, 1996, IJIO) Indirect 1-way network : When one of AB or BA is unfeasible or does not make economic sense, or when there is no sense of direction in the network so that AB and BA are identical, then the network is called a one -way network. Hardware -software paradigm (respectively A and B): the greater the usage of a hardware, the greater its attractiveness in terms of developing software, an increase in the number of software programs further raises the attractiveness of the hardware and increases the number of new adopters and so on (“ bandwagon effect ” or “(network) positive feedback ”). (non -disclosure agreement is when, for instance, a group of people keep secret the content of a meeting.) CONVEX Relationship Between number of users a nd the value of the network CONCAVE Relationship STORY BETWEEN IBM PC and APPLE and other companies We are in the ‘70s in the US. IBM had dominated the informatic industry, it was the main leader for what concern the production mainframe right after the WW2, it dominated the era of microcomputers. IBM didn’t believe in the concept of PC, it didn’t believe i n the possibility that every individual could have a computer. Mainframe were very large computer, that were usable only by very sophisticated and proficient individuals, the same concept applies to microcomputers. The concept of PC was behind any imaginat ion for IBM in that time: “There is a worldwide market for PC which is just made by 5 people” probably a quote of Thomas J. Watson, Chairman of IBM at that time → one the biggest misjudgement of the history of management. This judgement, true or not, is sug gestive of the attitude that IBM had, which didn’t believe in the commercial appeal of a PC . When big leaders don’t make a first step, there’s room for other to do it, and in particular for start -ups. In the ‘70s lots of start -ups, such as Commodore, Atari , Radio Shack, and in particular Apple, entered into this new business of PC. The judgement of IBM changed with the launch in the market in 1997 of Apple II, which was immediately successful in the market, especially because of a so -called killer applicati on, i.e. “Visical”, that was a spreadsheet. People bought Apple II in order to run this spreadsheet. IBM understood the misjudgement/mistake in their evaluation. The problem was that they didn’t have in house any technology and competence in order to come out very rapidly with their own PC, so they were forced to look around, beyond the boundaries of their firm to see if some technology could be outsourced to them. At that time, we had that the 2 most important part of a personal computer, the microprocesso r, and the OS. The microprocessor’s main producer was Intel, with an 8-bit microprocessor. The most widely used OS was the CP/M, produced by Digital Research founded by Gary Kildall in Palo Alto, S ilicon Valley . The time was mature for migration from an 8 -bit microprocessor to a 16 -bit microprocessor. IBM realised that that was the window of opportunity that they could exploit in order to enter into the market. They asked Intel to furnish them the new microprocessor with 16 -bit, that was 8086 model by Inte l. IBM wanted an exclusive license agreement, Intel refused the exclusive license agreement, Intel had willingness to provide IBM microprocessors but without this kind of agreement, they wanted to sell it to someone else. IBM was able to force Intel at lea st to have a second supplier, and Intel found AMD. IBM cannot get from an exclusive basis this microprocessor, but at least Intel had to make produce the microprocessor also by another firm, AMB, not allowing them to hold up IBM. In that moment started the competition on Lead Time between Intel and AMD, so Intel came out with a new microprocessor and AMD tried to understand what was going on, and after 6 months they came out with the same microprocessor produced by Intel, but Intel was already a step ahead, producing a new version of microprocessor. IBM tried to establish a relationship with Digital Research: they wanted to try to convince DR to update their OS so that is able to work with the new version of the microprocessor (16 -bit). In the august 1980, IBM manager flew to Palo Alto (California) trying to meet Gary Kildall in order to propose the business to him (see printed paper). The story is surrounded by many legends (someone says that Kildall refused to meet IBM’s manager ; someone says that the 2 never met; someone say that Kildall was unavailable going around with his airplane and delegate his wife Dorothy, she was not very much in business and therefore refused to sign a non -disclosure agreement, which is usual in business in this type of re lationship; someone says that they met but didn’t reach any agreement). We don’t know how the story went, what we know for sure is that the agreement was not stipulated and IBM went to their 2 nd choice, i.e. the magnificent Bill Gates, that has just bought for an amount of money 50,000 - 100,000 dollars, from Tim Paterson, entirely (so he own also the proprietary rights) the QDOS (Quick and Dirty Operating System). IBM asked Bill Gates to rapidly produce an OS for IBM’s PC with 16 -bit microprocessor, and Bil l Gates agreed. IBM wanted to buy the new OS, but B. Gates refused to that, furthermore Gates refused like Intel to provide an exclusive right agreement to IBM (note the cleverness of Bill Gates, that noticed that things would be for him better if he refu sed this agreement). In ’97 Apple II, in ’81 IBM launched in the market with its PC, which was very much a product assembled with other parties where IBM only put its brand: the only proprietary thing s of IBM was a particular shape of the keyboard and th e ROM BIOS - chip (basically, a chip that connects input and output devices). The fact that the product was branded IBM made a lot of difference since computers in 1981 was bought by very few people: most of the potential consumer didn’t own any personal com puter, the fact that IBM entered the market exerted a great incentive for software developers to write for its platform rather than for Apple and others. Soon after the entry of IBM into the market, the quantity of software available for the IBM Intel - Micr osoft was much larger than the variety of software available for the alternative platforms, but then more software available for these platforms oriented consumers to go for IBM’s platform rather than for the alternatives. Then more consumers means more in centives for software developers to write for that platforms rather than alternatives, and more software more the incentives to join IBM’s platform (bandwagon effect). This effect led to a winner takes all market result. These dynamics was also facilita ted by the fact that IBM was forced to adopt an open standard architecture , while instead the one of Apple was a close standard architecture . Apple since then till today, wants to have the control of every single piece produced for Apple: it’s a Fordist way of thinking, Ford wanted all the co al mine control for its production processes and also the railway system that transported coal from the mine to the steel mill. The second remark is that because of this open standard architecture by IBM, IBM Intel -Microsoft platform affirm itself at the expense of all the others. But who gains most was Intel -Microsoft because after the initial huge gains in terms of profits of IBM, IBM left the PC business because other companies (clones) such as Hewlett -Packard, Dell, and Lenovo stipulated agreement with Intel and Microsoft for the production of microprocessor and OS, basically reverse engineered the only 2 proprietary thing of IBM and entered in the market with their own PC. Consumers at the beginning din’t rush to buy these PC from newcomers since they were still not very branded compared to IBM. But when they realised the convenience these companies star ted to erode market share of IBM, and the latter than left the market. 3rd remark: here we have the exemplification of the strong non ergodicity of this business process. If things went differently at the very initial period, we could have a very differ ent story and outcome of the industry (Butterfly effect) . (Si definisce ergodico un processo statistico che passa per tutti i punti possibili di lavoro. ) 3 remarks 1) Bandwagon vs. negative feedback If someone is experiencing a positive feedback and has competitors, these latter are probably experiencing a negative feedback The higher is the number of users of a platform, the higher is the incentive for the developers of complementary goods (sw) to create contents for such platform, the hig her is the attractiveness of the platform , the higher is the number of users that use the platform ( bandwagon effect or positive network feedback) and so on… The lower is the number of users of a platform, the lower is the incentive for the developers of com plementary goods (sw) to create contents for such platform, the lower is the attractiveness of the platform , the lower is the number of users that use the platform (negative network feedback) and so on… 2) These dynamics also affect “Two -sided markets” •Rochet -Tirole(2006, RAND): Examples: on -line auctions, credit cards, crowdfunding “markets where one enable interactions between end -users, and try to get the two sides “on board” Subject to an indirect network externality: e.g. eBay platform as hardware, sellers as software producer, buyer as consumer. The more the number of sellers, the more the incentive of consumers to go to the market, due to the higher willingness of the people to sell. Then, higher the number of consumers, highe r the number of people willing to sell. And so forth and so on. Negative same side externality: seller wants to enjoy monopolist power; consumer would like to be the only consumer (monopsony power). (Monopsony power exists when a single buyer or an association of buyers can dictate the prices they pay to suppliers, or control other aspects of the relationship that exists between themselves and their supplier s.) 3) EMPIRICAL ESTIMATES OF NETWORK EXTERNALITIES Large number of studies in the economics -managerial literature that document the presence of network externalities: •Indirect : Software (Dranoveand Gandal, 1999; Gandalet al., 1999), ATMs (Salonerand Shepard, 1995), CD players (Gandalet al., 2000), Yellow Pages (Rysman, 2002), Videogames (Clements and Ohashi2005), Cryptocurrency (Gandaland Halaburda, 2014), Electrical Vehicles (Li et al. 2017). •Direct : Telecommunications services (Majumdarand Venkataraman1998), Mobile phones (Baraldi2012). How to estimate them? (There are multiple ways to do this. We focus on the following one) Complementary goods approach. The complementary goods approach (e.g. Gandal et al. 2000) derives a system of equations and uses the number of software available as a variable in hardware adoption regressions and vice versa. [In the hedonic approach (e.g. Gandal, 1994, Sarnikar, 2002), the installed base is treated as a product characteristic that will have a po sitive effect on prices if there are network effects]. [In the adoption (e.g. Koski, 1999, Gandal et al., 1999), the installed base at t -1 carries a positive expected sign in the adoption or diffusion equation a t.] [the timing approach (e.g. Saloner and Shepard, 1995) establishes that firms with higher expected network effects will adopt a technology earlier and proxies expected network benefits by the number of potential (internal) users of the techn ology.] There are multiple way s to try to infer whether in a market there ’s a significant presence of Network Externalities or not, w e focus on complementary good approach in trying to esti mate the presence of indirect network externality in the market. we focus on it because it is useful in the DVD vs DVX case we will deal with. W e need data in a given time period about the usage and sales of hard ware units and software sold. Researchers estimate 2 equations . In the first the dependent variable is represented by the sales of the hardware and independent variable of interest is the variety of software available one period before. Then , of course, we need to control for several control variable s. The second equation choose the variety of so ftware as dependent variable and the independe nt variable of interest is the hardware sold the period before . W e control for the control variable we wish if we have strong indirect network exte rnality in the market, we should observe the estimation of these 2 coefficients are both positive and statistically significant (of beta 1 and beta 2, both should positive and statistically significa nt, that means that software affects positively the sales of hardware and the sales of hardware positively affect the variety of software available for that time ). Basically , if you have the depend variable on the Y and the independent variable on the X , we observe many data and they should be better interpolated by a lina that shows us a posi tive an d statistically significant coefficient. 1 At the beginning we have the most important causes of the future success of a product (butterfly effect). Innovation in a network market: the start -up problem The demand for a network good (Aggregate demand shape and how it’s done) Res pect to the standard demand curve, the demand for a network good does not only depend on price but also on the agents’ expectations about total consumers of the good. (Consumer expectation about the penetration of companies. The value of all the products d epends also on the choice of the consumers because of the network externalities.) ➔ Some basic (and key) implications -Everybody in the economy is convinced that no body will buy the network good: the network good will be effectively unsold (very pessimistic expectations) irrespective of a high or low price: “network failure”. -Consumers want to buy a hardware if many software are available for that hardware. But software developers will write software only if a great number of people has already bought the hardware since they are potential buyers: if nobody makes the first move the network good will be effectively unsold (again irrespective of a high or low price): “network failure” → Chicken -egg paradox equilibria (note how these markets are characterized to a great extent by “self -fulfilling prophecies”) -The aggregate demand could be upward sloping (Rolfhs1974 , BJE): If a firm is able to influence the expectations of potential consumers about the penetration that the good will have (and so its value in the future), is totally plausible that as the expected size of the network increases, consumers are willlingto pay a higher price for joining the network. -Demand presents a critical mass effect (Rolfhs1974, BJE): Critical mass in nuclear engineering Under radioactive decay conditions, the amount of uranium necessary to start a self - sustaining process of production of neutrons that maintains unchanged its quantity. Any larger amount of uranium will cause an explosive nuclear chain reaction. Any smaller amount of uranium will cause nothing, and it will soon decompose. Critical mass in sociology https://www.youtube.com/watch?v=r0Ls24Fbl0g In sociology this effect is due to psychological causes. Many social phenomena when we cannot observe intermediate outcome → description through this model. See also the Appendix -Demand presents a critical mass effect (Rolfhs1974, BJEM): Critical mass in network economics For any given price charged by the firm(s) it is the minimal amount of consumers which join the network and are satisfied of this choice. Any larger amount will trigger the bandwagon effect, any smaller amount will bring to a network failure. Here, the effect is routed in the technological features. Lower the price, lower the amount of people that will join the network and are s atisfied of this choice. If you increase price you have to convince more people. Endogenous to the behaviour of the firm: the firm can choose the critical mass that it wants to achieve. STYLIZED EXAMPLE in network economics Firm A has invented (and potent ially patented) a network good (exhibiting direct externalities) and is going to commercialize it under a monopolistic regime. Trade -off •High price and a high number of people to convince •Low price and a small number of people to convince Suppose: •p = 1€; at that price a representative agent may be willing to buy the good if it can be enable to communicate with (say) 100 people. •p = 100 €; at that price one may be willing to buy the good if it can be enabled to communicate with (say) 1.000.000 people. If the choice of the firm is p = 1, the firm has to convince 100 people If the choice of the firm is p =100, it has to convince 1.000.000 people Whatever the choice, if the firm wants to sell the good it has to attract the critical mass (which is increasing in price). If p =1€but the firm convinces only 99 individuals, someone of these individuals will be unhappy (someone who gives to the good a value of 1€only if 100 individuals had joined the network); so he will leave; the network size will shrink to 98 individuals, again someone of these individuals…… If p =1€ but the firm convinces more than 100 individuals, e.g. 101, the value of the network good raises, and some more agents will want to buy the network good, network size becomes 102 , so the value of the good raises……….. → AIM : Attract Critical Mass so to innescate the BANDWAGON EFFECT Aggregate demand for a network good: a stylized model (Rohlfs1974, Bell J of Econ and Man; see also Shapiro and Varian 1999) Consider a network that is of interest for N potential agents. They are indexed by xwhich is uniformly distributed on the interval [0,1]. Each agent faces the binary decision of whether to buy the good or not. The good exhibits positive network externalities Since a continuum of potential consumers exists in the interval [0,1], there will be therefore a particular consumer, indexed by x* , such that she is indifferent between buying and not buying the good. This consumer is found by: Hence all the consumers that have a higher willingness to pay for the service ( x≤ x* ) will buy the good, while all the agents that have a lower willingness to pay( x >x* ) won’t. Perfect foresightis assumed (Remember self -fulfilling prophecies) …and then the inverse aggregate demand curve for a network good becomes 2 stable equilibria: a)Network of zero size (“network failure” caused by pessimistic expectations) b)Network of large size(CRITICAL MASS is reached and positive feedback is triggered) Critical mass : for any given price, it is the minimal amount of consumers which join the network and are satisfied of this choice MARKET DYNAMICS (Shapiro& Varian1999) “[…]It is plausible to assume that when people are willing to pay more than the cost of the good, the size of the market expands and, when they are willing to pay less, the market contracts. Geometrically this is saying that when the demand curve is above the supply curve, the quantity goes up and, when it is beneath the supply curve, the quantity goes down.” Typical path of a network good in a dynamic framework: X = % users (penetration rate) 1°phase: launch ( 0 –Xl); 2°phase: rapid growth(positive feedback, Xl -Xm ); 3°phase: maturity(universalservice policies, X m –100%). Unsuccessful network good has a flat line a a s graph. Compared to a standard good diffusion, the network good: 1. reaching the critical mass: the initial phase can be more problematic, i.e., the launch phase. 2. because of the bandwagon effect once reached the critical mass the growth is faster. 3. the product become so important that the amount of people that adopt the network good are higher. How to solve the start -up problem? STRATEGIC POLICIES Compatibili ty with competitors: share the effort but be careful since interlinking can generate a free -riding problem Promotional prices (very often below costs) Advertising (size and nature) https://www.you tube.com/watch?v=M5fNInc9OoE https://www.youtube.com/watch?v=jhhyivR -SrY → Nature of adv is different for network market compared to other kinds of goods. (e.g. below snob effect vs be part of a community). Appendix Critical mass in sociology Adapted example from Schelling (1978): Micromotives and Macrobehaviour Background Suppose a University Professor is particularly nasty and deci de to schedule a series of seminars on Saturday afternoon. Attendance by students is not strictly mandatory, but of course «being there» (or not being there) may favourably (or negatively) impress the Professor. Knowing all that, each student will decide whether to participate or not on the basis of a mi nimum threshold of attendance to these seminars by the class. Students’ expectations are adaptive over time (i.e. what students expect in terms of seminars’ participation in the present week will strictly depend on the actual attendance registered in the p revious week). Needless to say, there is no possibilities for coordination among students. The class is composed by 10 students. Students are heterogeneous and they have different minimum thresholds. More specifically, let us assume that these threshold s exhibit a distribution close to the normal Gaussian distribution. In particular, for 1 student the minimum threshold is equal to 2 students; 1 student has 3 as minimum threshold; 2 students have 4; 2 students have 5; 3 students have 6; and lastly, 1 stud ent has a minimum threshold of 8. Analysis 4 students represents the critical mass for this cycle of seminars : 4 students are willing to participate only if the participants to the seminars amount to 4 students. This is an equilibrium. But is it stable ? Let’s try to perturb it. Let us suppose that at the first seminar instead of 4, 5 is the number of attendees. The following week (as you can see from the graphs of the previous slides), the number of students who will attend the seminars will be equal to 6; the Saturday of the 3°week, students will become 9, and at the Saturday of the 4°week, every student will attend the seminar. Let us suppose that instead of 4, 3 is the number of attendees. The following week (as you can see from the graphs of the prev ious slides), the number of students who will attend the seminars will be equal to 2; on the Saturday of the 3°week only one student will attend, starting from the Saturday of the 4°week, every student will not attend the seminar. The tipping point between success and failure is represented by the achievement of the critical mass. APPENDIX Network goods , Diffusion path examples Some examples 1) Video -telephone: network failure in U.S. in ’70 and Italy in ‘80 -Launched in the early 70s by AT&T in the Chicago area. -Price not too high (86 $ flat monthly rate) -Service quality not too bad -Major failure causes: a) Privacy b) Rate of return regulation (cost -plus scheme) -If something new is technically feasible does not necessarily mean that will be appreciated by consumers (invention does not imply innovation) -Low incentives to invest in marketing and related expenditure to reach the critical mass. 2) Fax -1843 invented by Bain -1865 commercial fax service between Paris and Lyons -1925 AT&T commer cialize the fax in U.S. -1982/1990 rapid growth (c.a. 80% penetration rate among offices) 3) Telephone (invented in 1870s) US President Rutherford Hayes «It's a great invention, but who would ever want to use one?». One to one Vs. Network nature Exploitation of patents in the short -term, i.e. high prices. Rent of the equipment: $14/month (equivalent to $330 in 2000s) vs. production cost: $4 [See Rohlfs2003] VCR=videocassette recorder or video recorder (video recorder is an electromechanical device that records analog audio and analog video from broadcast television or other source on a removable, magnetic tape videocassette , and can play back the recording. ) Network economics: the hardware/software paradigm, winner -takes -all - markets and technological standard wars_3 FURTHER REMARKs ON THE POSSIBLE NON -EMERGENCE OF “WINNER TAKES ALL (or MOST)” MA RKET EQUILIBRIA Outcome of the paper shows that where people are really interested in communicating in few people, and not with all the market, the insurgence of winner takes all equilibrium is not very likely. Beta close to 1, higher possibility of winner takes all. beta near to 0 → the probability of winner takes all decrease E.g. I’m interested in communicating only with my family, the same for another person. in this situation, the arising winners takes all become unlikely because one family may be served by 1 social network and the other family may be served by another social network. The winner takes all positions, is the increasing presence of the phenomena known of multi -homing. for what concern consumers' slide an example is t he series provided: I do not have to choos e single platform for enjoying one typology of content, but I can have more than 1 platform. On the other side we can have multi -homing on the software producer side where 1 software is producing for 1 typology of hardware but the conversion cost among different software is very low. So, a game that can be read by different hardware. Me may end up with some local markets where we have some dominant position, dominant position in a local market doesn’t mean that i have the dominant position in another market. The 3rd possible reason why we do not observe winner takes all market position is the fact that in some market we can have the existence of geographical bounded network externalities. The presence of 1 of the reasons do not exclude the existing of another one, for the description the arising of winner takes all. Strategic choice between compatibility and incompatibility (Even in the case o incompatibility we cannot always be sure that a standard war will arise and winner takes all position will emerge into the market ) COMPATIBILITY or NON COMPATIBILITY: this is the question First , let us define compatibility: 2 products are defined compatible if they “can work together”: the output of a given bra nd can be used by other brands. If this is the case we say that different brands adopt the same (technological)standard. In the opposite case, we say that products are incompatible. There is downward compatibility when a new release of a product is compatible with the old one, but it is not the other way around (example Windows Office). A product is “one -way compatible”if it can work together with the output of a rival brand, but it is not the other way around (example: Linux -Windows). (Compatibilit y offered in this case by the small network. while the large network is protected. Linux is a small network, software written for windows can be used by Linux products). If the firms are adopting the same technological standard and the analysis conducted till now apply. In order to better understand the consequences of the strategic choice (comp. Vs non - comp.) let us focus on the characteristics of a network market The success of a standard will depend on its capacity to solve the start -up problem (i.e. attract the critical mass) and trigger the bandwagon effect The success of a standard will depend on its capacity to solve the start -up problem (i.e. attract the critical mass) and trigger the bandwagon effect. If firms choose the non comp. path, the y have to bear in mind the following 3 important facts. Observations Formally (see Farrell and Saloner1985, Economic Letters; Cabral IO) 2 incompatible goods: A and B. Population (N =1) is formed by agents who prefer A (a) and agents who prefer B (b) wit h a+b=1. Utility of a: v(x) if buy A and v(x) –α if buy B Utility of b: v(x) if buy B and v(x) –α if buy A X=number of consumer that have bought that specific typology of good, Let us suppose that market is perfectly competitive so firms charge the same price, and also suppose that being the only one to acquire the good will bring a zero benefit. Monopoly will be an equilibrium if v(1) –α≥ 0. Now suppose a sequential entry of consumers into the market. Consumers b will buyA only if v(a) –α> v(b) [s o v(a) -v(b)> α]and Consumers a will buy B only if v(b) –α> v(a). Alfa measures the disutility in buying the least preferred good. consumer of type A prefer A than B, if they buy B they have higher disutility in consuming that good. This situation can be represented by the diagram below. The achievementof a criticalmass in a standard war Once the market touch this technological absorbing bullet we say that the market is locked in that specific product. consumer b will buy a. v(a) - v(b) ->everybody will buy a rather than b. this is what happens in standard war. we observe that in a certain po int eBay reach the critical mass and triggered bandwagon effect at expense of competitors. at a certain point eBay product touch the technological absorbing barriers and lock the market. If you are bale toa ttrach early adopters you have better possibilit ies to lock in the market throgugh your product rather than the one of competitors. 2) Early adopters are more important(for reaching the critical mass) the higher are the network externalities(high v’, i.e. the first derivative of v with respect to the n umber of users) and the lower is the love for variety (low α: it means that in the eyes of consumer the 2 products are very similar). 3) Same logics applies to the first mover advantage. QWERTY vs DVORAK Questionable whether there was path dependent inefficiency or not………..less questionable is the fact that the process has shown a higher degree of path dependency since we are adopting this system invented 150 years ago. QUESTION(S): WHY A KEYBOARD IS A NETWORK GOOD? AND OF WHAT TYPE? This k eyborad is a sort of playtform that makes communicate computer users and computer producer, we have cros side ext going on. more the user, more the incentives to produce pc with qwerty. the more pc with qwerty the more the incentive of people to use qwerty . Keyboard as a platform making communicating 2 side of market, or keyboard as hardware, pc are producing software and consumer are consumer. it's a indirect externality goings on. 2 forces may spur standardization → • Firms have to bear high costs in R&D to develop a standard and in large measure they are sunk because highly specific. • Note also that marketing expenditure and coordination complementors costs can be extremely high when a firm choose incompatibility and opt for a go -it-alone strategy So it is not unusual that in network markets standards are defined through international organizations (UL, ITU, NIST in the USA) or alliances between firms (“coopetition logic”). Or firms choose an “open standard” policy, i.e.the firm that developed the standard makes it available to other firms (very low licensing fees for the patented technology they developed). [See the Appendix for a thorough analysis of the “open” vs. “closed” standard strategy] (3 typologies of compatibility: 1. third neutral institution owned and managed by governmental body who choose the standard in the market. (This is the case of TLC services). 2. alliances between firms: coopetition logic. firms without the intervention of an y external body decide by their own to form these alliances to establish which are the standard in the market. 3. …) The logico f open standard througha game -theoretic approach(Grindley1995, Pepalletal. 2009) a) Battle of the sexes game (2 strategies: lead or follow) We cannot predict who will the leader and who the follower Result : Both firms will prefer to cooperate rather than fight each other b) The rather fight than switch game Envisage a standard war: both firms want to lead Result : Both firms prefer to fight rather than cooperate Moving from the 1 to the 2nd game. little change pose that the Nash equilibrium envisage a standard war. its' one where both want to lead. we observe that actually they end uop with an expected pay off (3,3), which is pareto dominated by the NashNequilibria in the first game since 6 and 4 are always greater than 3 and 3. a standard war lead to a inferior level of expected pay off because standard war are costly (marketing and expenditures to arrive before the competitors). the other reason: standard war may originate inertia on both consumer and software producer in the case of network good with indirect network externality. The equilibrium of the first game Pareto dominates the equilibrium of the second game The difference between the two games stems from the greed for profit of the firm that develops the standard Lesson: This is an example in which firms’ are better off if are cooperative or not to greedy. Firms can switch from the second to the first game if the leading firm: -Set very low licensing fees (policy of open standard) -Reduce absorption costs of the follower (in knowledge -intensive sectors these may be large, e.g. The Micronascase, in lecture on Transaction costs). Appendix Strategi c positioning decisions (Grindley1995) 4 Strategies Sponsor/defend : development of a standard and restriction of the use towards concurrent firms (high licensing fees) Give away : possibility for the concurrent firms to use the standard without restriction or with low licensing fees License in : use of a standard developed by another firm Clone : use of an open standard without restriction Network economics: the hardware/software paradigm, winner -takes -all -markets and technological standard w ars_4 No comp : phenomena of inertia totally apply. Consumer may wait to see who represent the winner standard. If everybody waits, there are lower probability to achieve critical mass, which will never occur. The firm that establishes itself as market standard, gets a big pie and fully benefit from it without sharing it with anybody. CASES ON STANDARD WARS (3 cases) Arthur (1996), Harvard Business Review (“Increasing returns and the new world of business”). It is casino gambling, where part of the game is to choose wh ich games to play, as well as playing them with skill. We can imagine the top figures in high tech — the Gateses of their industries — as milling in a large casino. Over at this table, a game is starting called multimedia. Over at that one, a game called Web s ervices. In the corner is electronic banking. There are many such tables. You sit at one. How much to play? you ask. Three billion, the croupier replies. Who’ll be playing? We won’t know until they show up. What are the rules? Those’ll emerge as the game u nfolds. What are my odds of winning? We can’t say. Do you still want to play? High tech, pursued at this level, is not for the timid. (Above other factors, such as courage → gift of precognition, in seeing before others how things will unfold) In fact, the art of playing the tables in the Casino of Technology is primarily a psychological one. What counts to some degree — but only to some degree — is technical expertise, deep pockets, will, and courage. Above all, the rewards go to the players who are first to make sense of the new games looming out of the technological fog, to see their shape, to cognize them. Bill Gates is not so much a wizard of technology as a wizard of precognition, of discerning the shape of the next game. Rohlfs, 2003 , 2003, Bandwagon effects in high -technology industries, p. 45: [in the case firms choose NO COMP, the most rationale strategy] “may be to signal an irrational commitment to win the race (of the standard) at all costs. This is analogous to a well -known result of oligopoly theory; namely, the most rational strategy for an oligopolist may be to signal an irrational strategy commitment to punish price cutters -regardless of the ruinous cost that it may incur by doing so.” What is important Factors that can positively influence the st andard war Size Financial resources Commercial strength (marketing and distribution channels) Brand and reputation The IBM -APPLE case testifies that these factors matter, the VHS -Betamax case are an example of how a small firm can beat a big company. If you start a standard war and you are exerting a lot of eff ort in advertising and marketing, but then for some reasons you slow down the freed om of your adverting and marketing expenditure , you risk to or iginate and spread out in market rumors that you are not really convinced about your hot of winning, and this market this might be crucial in determin ing your loss , because th is market are heavily characterised by self fulfilling proficies: if many people start believe something th en that something may happe n (if many people sta rt believe you may lo ose, then you will effectively loose ). Sometimes particip ating in th is standard w ar are also called “all in ” by firm , “all in ” may rein for ce the expectation of consumer that you will b e winning standard, then you will be the winning standard. It’s very diffic ult to predict how a standard war will end up at the end of the day, there are some exogenous factors which may orient the battle/the war. O ne of the most prominent is the size of a company in terms of financ ial resources available, capacity to access large distributi on channel , and also manufactu ring capacity of a firm: in this sense a large firm has an advantage toward a star tup , even for the expectat ions in this market ; if IBM or Sony vs startup , clearly the market expect that the winning s tandard will be the one of the large firm rather the one of the startup. Anyway, we have exa mples where David defeated Golia, so a small company because of the implemented strategy was able to win. In this market first mo ver adva ntage exist, but both the history of IBM against Apple shows us that it is not crucial in order to determine the victor y: the case where VHS standard was the winning standard against the one p roposed by Sony with Betamax. x 1) VHS VS Betamax PREMISE 1970 partnership between Sony, JVC/Matsushita (smaller compared to Sony, affiliated to Matsushita, i.e., Panasonic) for developing a prototype of VCR for scientific purposes ( UMatic on a patent -sharing agreement , used for video cassette recording: promising but not suitable for mass market. The only commercial application → record scientific information). Both were unsatisfied. Sony accused JVC for participating only for stealing knowledge. 2 different technological trajectory fol lowed: IMPORTANT DATES 1975 Launch of Betamax (Sony) in the U.S. 1976 Launch of VHS in the U.S (JVC/Matsushita and other partners) 1988 Sony adopted standard VHS. Time -shifting, recording tv programs. Not a network good → network market The community was much larger for JVC: buying blank cassettes to record tv programs was used to record movies, and most of the movies last more than 1h. afterward Sony improved their products improving the length of fruition of the cassettes. In 1977 → new model for Sony, 2h length. JVC came out with 4h length cassettes → community of JVC continued to growth, because of the need of the recording of sport events, such as Superbowl (which last more than 2 h. time -shifting). The market afterward became a network marke t because in 1976 -1977, Andre Blay asked to many companies to allow him to have the permission to produce a series of movies. Only Twenty century Fox allowed. Then other entrepreneurs followed what he did. In the same years, George Atkinson came up with th e idea of renting the movie → rental home video business started. These reproduced movies are the software for the hardware (products of JVC and Sony). Target choice: they choose to produce movies primarily with VHS (JVC) for the larger customer base. The co mmunity of VHS started to grow even more → bandwagon effect. Main takeaway: 1.Difference between invention and innovation. Videocassettes used for another purpose, considering the one they were produces for → discernible ex -post 2.Power of software producer on who will be the winning hardware. If multi -homing was prevailing in the software side, we would have not observed a winner takes all dynamics. 3. Look at the chart above. Panasonic → lead and open standard. DVD replaced videocassettes recording. ‘90s : DVD (Digital Versatile Disk) Sony, Toshiba, Panasonic, and others with content producer (such as Columbia, Warner, MGM, and Polygrow) participated to this format. First DVD Sept. 1996: heavily commercialized by the 2 nd e-retailer that was, still is, Best -Buy. The launch of the DVD had a good initial success, in particular due to the great storage capacity. 2 major movie studios pretty concerned about DVDs, Disney and Universal. The concern was about piracy, DVD has som e technologies that allowed piracy, but they were insufficient according to the 2 companies mentioned above. Best -buy main competitors → Circuit city wanted to stop consumption of DVDs, avoiding Best -buy acquiring more consumers. In sept. 1997: Circuit city announced an alternative technology to DVDs. To launch it Circuit city made agreement with hardware producers (Zenith) and software producer (Universal). In the fall of 1998 DIVX was launched into the market. Strategic announcement made in order to influe nce expectations that is not immediately followed by effective entry of the product in the market: Vaporware. It offers one -way compatibility with DVD, (possibility to read DVD risk with DIVX hardware, not the other way round). June 1999 → retired from the m arket. Takeaways: 1) it was proved, that to really stop the sale of the DVD was the announcement made in advance → Vapoware*. *Proved through the estimation made by Dranove and Gandal. Entry an Denise occur when the bandwagon effect already started The announcement made before tech absorbing barrier (critical mass). It was possible because if we t ake th is model with the technological absorbing bar rie r (for DVD in thi s case ) and te ch abs barrier of DV X, the announcement, the vaporware occur s before D VD had reach the technologic al absorbing barrier . W e can also see it in another way, looking at the graph of the right. DVX announcement Entry and demise occurs occur when the tech abs orbing barrier for DVD was achieved so the bandwagon effect started and that is unstoppable On Y in % we have t he penetration rate Critical mass Vaporware: an intangible thing like vapor can have more tangible result than tangible things like the entry of a new alternat ive product in the market o r its exit. Still used very much today, e.g. sony announce 1.5 y ears before the next plastation model. 2) ** (For Linux -Microsoft → survival strategy). According to the professor : not bad move but didn’t come with exclusive agreement with software producer : there were no break from the consumer viewpoint: every movie was available with DVD. It would be more intelligent according to t he professor to jointly apply one -way compatibility with an attitude for e xclusive agreement from the software side → less differentiated product. HD DVD – Blu -ray HD DVD evolution of DVD. Blue -ray considered a revolution in terms of performance. At that time agreement was not possible due to the reason above written. Lau nch and start -up phase CES 2006: both formats are present and presented April 2006: HD DVD commercialized at $499 July 2006: Blu -ray commercialized at $999 Fall 2006: Microsoft commercialize Xbox with HD DVD add -on at $200 Fall 2006: Sony PS3 commercialized with Blu -ray at $499 Paramount and dreamwork abandoned Blu -ray. (rumour about payment made to Paramount and Dreamworks to make only HD DVD prod uct). Market is sluggish…….. “Despite tremendous efforts and investments by both sides to gain the upper hand, sales of both formats had been sluggish by any measure. The slow consumer adoption was generally attributed to their reluctance to upgrade while the format war was still ongoing. Uncertainty was also hurting the movie studios: in 2006 one analyst firm had projected that media companies stood to lose $16 billion in revenue over seven years if a single standard failed to emerge.” [“The last DVD forma t war”, Andrei Hagiu, Harvard Business School] Inertia: before making the choice (consumption in particular) want to wait to see the outcome of the war. (Accused by Toshiba because monetary induced by Sony) Warner’s reasons “[They should] understand that this is the last physical format there will ever be. Everything is going to be streamed” - Bill Gates, Microsoft chairman U.S Sales: Blu -ray : $169 Million HD -DVD : $103 Million “We’ve been monitoring the situation with consumers for a while now and they have clearly made their choice,” Mr. Meyer said. “We followed.” The New York Times, Brooks Barnes 5/1/2008 Toshiba’s fall “Within a month following Time Warner’s move, electronics retailer Best Buy, movie rental company Netflix, and finally WalMart, all of whom had been stocking movies in both formats, announced that they would drop HD -DVD, leading to Toshiba’s surrender. On February 19, 2008, Toshiba announced that it would discontinue production of DVD players, recorders and parts i n the HD -DVD format, which it had been championing since 2002 and in which it had invested large amounts of money to develop and bring into the market” [“The last DVD format war”, Andrei Hagiu, Harvard Business School] Takeaways: 1. Given the software industr y importance hardware producer are incentivized to pay them 2. Inertia on the consumer, on the software producer also, that standard way may arise 3. Precognition gift Switching costs & technology replacement Switching cost, i.e. the cost for a user for migrating from the use of one technology towards another. In network markets, since they are often market in high technology which may require sophisticated knowledge, we have the important presence of individual switching cost suffered by individuals. Since we are talking about network good, we have also collective switching cost, because of the network externalities. The value of the new network good stem from how many others will buy the new version of the technology. Strong individual switching cost and collective switching cost make a huge difference compared to what are the switching cost of other contexts, e.g. dishwashing machine → choice based only on price of new technology and performance of new vs old technology. In high tech network market, w e have also other 2 dimension, the 3 rd and 4 th (which originates coordination problems and strong lock -in effects, i.e. people are reluctant to change. They prefer to remain with the old version: inertia for the migration from the old to the new) displayed in the slide above. E.g. Necessity of consumer to embark in learning investment, such as OS. If we introduce a new OS, to migrate toward it we need to disregard the human capital invested and learn again, learn how to use this new OS. For standard good we do not have these drawbacks. We may have other reason why that increase these costs. To categorize potential coordination problems 2 main coordination problem (3,3) Nash equilibriuim dominated by (4,4). Here we have a case of excess inertia. Both stays with the old technology but ould benefit from the migration to the new one. ½ N O N (4,4) (2,1) O (1,2) (3,3) ½ N O N (3,3) (1,2) O (2,1) (4,4) Case of excess momentum, no need for consumer to hurry up the adoption of the new technologies, because would be better to stay with the old one, nevertheless for some reasons they migrated. (3,1): consumer 1 is happy, but consumer 2 is experiencing excess momentum, he would stay better in the old situation. If the equilibrium that emerge is (2.5,2.5), consumer 1 excess of inertia, and consumer 2 stay better here. Affirmation of the new te chnology at the expense of the old one. Excess momentum is more subjective. Final thought Always keeping in mind caveats (avvertimenti) and exceptions (e.g. community of interest, local nature of network externalities, multi -homing phenomena ), we know that the choice of “non compatibility” (“standard wars”) can frequently lead to global monopo lies Thus, if WINNER TAKES ALL MARKET!!!!!!! What the LOSERS can do? But nothing is everlasting!!!!!!!....or is it? In network markets competition and rivalry could not be within the market but rather be for the market. Monopoly can only be temporary, and they can always be interrupted by technological breakthroughs (Schumpeterian competition, SM1). If this is the case, market shares can be a poor indicator of the degree of competition and rivalry that exist i n a network market. But…Network monopolies are (probably) becoming more and more stable………. Why? -Network monopolies are now conglomerates with enormous advantages (financial resources but also data) and they are persistent top investors in R&D (SM2). → confirmation of SM2 validity. (see DaikoT., DernisH., DossoM., GkotsisP., SquicciariniM., VezzaniA. (2017). World Corporate Top R&D Investors: Industrial Property Strategies in the Digital Economy. A JRC and OECD common report. Luxembourg: Publications Office of the European Union; Grassano, N., Hernandez Guevara, H., Fako, P., Tübke, A., Amoroso, S., Georgakaki , A., Napolitano, L., Pasimeni, F.,Rentocchini, F., Compaño, R., Fatica, S. and Panzica, R. The 2021 EU Industrial R&D Investment Scoreboard , EUR 30902 EN, Publications Office of the European Union, Luxembourg, 2021, ISBN 978 -92 -76 -44399 -5, doi:10.2760/559 391, JRC127360) [e.g. top big 5, i.e. GAFAM (Google (Alphabet), Apple, Facebook (Meta), Amazon and Microsoft) are the largest in terms of capitalization in their sector but also they are the one that invest more in R&D of their sector] -The narrative of a successful entrepreneur is now the one who “exits” by selling its business to an incumbent (lack of SM1). (This viewpoint is different from the one saw in the lecture with Pedota about the willingness of entrepreneurs -innovators to expand their empire) -K iller Acquisitions [(e.g. GAFAM have the strong tendency to acquire startups → the large incumbents have the corporate shooting attitude. Many M&A are made not only for acquiring the technologies developed by the startups, but to KILL the technology that is near to arise in order to not have disruptions (Killer Acquisition)] In the picture below we see the most important acquisition made by GAFAM, moreover the trend is still the same, constantly growing. 2001 -2018: Google has bought 1 firm per month, ever y month, for the past 17 years. Incumbent: be alwayson guard!!!!!! The Microsoft case is emblematic of the fact that a monopolist in a network market cannot relax (never!) and need to be always on its guard. (always, we may have threat popping out that threaten the leading position) The browser war in the mid’90s ( Microsoft vs Netscape ) Early