On Software Piracy 185 Copyright © 2005, Idea Group Inc. Copying or distributing in print or electronic forms without written permission of Idea Group Inc. is prohibited. The profit of the original software firm is: * * O π = ()( ) 2 ) 3 4 )( 1 ( 1 1 4 qq qq θθ θ −−− −− (8) and that of the pirate is: () () 2 ** 34 1 qq qq P θ π −− − = (9) The following result summarizes the impact of the presence of the pirate in the market under network externality. Proposition 1 In the presence of network externality, when the pirate is present in the market, the demand for the original firm is higher than its demand under protection, while price under piracy is lower than under protection. Formally: D * O > D * NP and P * O < P * NP . Proof: Follows after comparing (and simplifying) (2) with (6) and (1) with (4) respec- tively. Q.E.D. So under network externality, the presence of the pirate has a positive effect on the original firm’s demand as expected, but a dampening effect on the price due to compe- tition. Under this, we are interested to see how these two opposing effects combine and what would be a more profitable situation for the original firm between piracy and protection. Protection versus Non-Protection We compare between the profits of the original software firm under protection and non- protection. Proposition 2 In the presence of network externality given a choice between employing protection and non-protection, it is always profitable for the original software developer to protect its software. TLFeBOOK 186 Poddar Copyright © 2005, Idea Group Inc. Copying or distributing in print or electronic forms without written permission of Idea Group Inc. is prohibited. Proof: To show that π * NP – π ** O ≥ 0 Observe that: π * NP – π ** O = ()( ) () −− −− − − 2 34 114 4 1 )1( 1 qq qq θ θ θ = ()( ) 2 2222 3414 98108 qq qqqqq θθ θθθ −−− ++−− The denominator of the above expression is non-negative. We have to show that the numerator is non-negative for all θ and q. Simplifying the numerator, we get (1 – θ )[8q – q 2 (9 θ – 1)], to make it positive we must have q q 9 8 + ≤ θ which is always true for all q ∈ (0.1) and ∈ 2 1 ,0 θ . Note that q q 9 8 + is decreasing in q. Q.E.D. This result is interesting since under network externality, when the pirate is present in the market, even if there is a positive effect on the demand of the original firm, yet the more profitable situation for the firm, is to protect. Proposition 3 The original software developer has got higher incentive to protect its product in the presence of network externality as oppose to the case of without any network external- ity. Proof: It is easy to see that the incentive to protect increases with the degree of network externality. Gain from protection under network externality is () () 2 344 98 qq qqq θ θ −− −+ = G (say). Observe that G is an increasing function of θ . Q.E.D. Discussion In this part, we tried to argue that the prevalence of network externality in the software user market cannot generally be held as a reason for software piracy. We showed that in some situations, even with very strong network effect, protection instead of allowing piracy, is the optimal measure for the original software developer. To this end one might argue that in our model since deterring the pirate (or protection) is costless to the original TLFeBOOK On Software Piracy 187 Copyright © 2005, Idea Group Inc. Copying or distributing in print or electronic forms without written permission of Idea Group Inc. is prohibited. software firm, the original firm will always deter the pirate and enjoy the monopoly market simply because monopoly profit is always higher than duopoly profit. To this end, readers would also like to notice that the presence of the pirate increases the demand of the original firm through network effect compared to the full protection case. Thus, this is clearly a positive effect of allowing piracy. Although there exists the dampening effect on the price under piracy due to competition, but a priori it is not quite clear which effect dominates and eventually which situation would be more profitable to the original firm. Digging a little bit deeper and contrasting with previous results in the literature (discussed earlier), we realize that the market structure, the nature of competition and the demand structure play a very crucial role to drive these results. For example, when the market structure is monopolistic with two types of consumers, software piracy allows price-discrimination among the different classes consumers (see Conner and Rumelt, 1991; Takeyama, 1994; Slive and Bernhardt, 1998). 14 On the other hand, when the market structure is duopolistic (or strategic in general), the results regarding the existence (or not) of software piracy very much depends on the nature of competition between the competing firms. For example, when competition takes place between two symmetric firms (both are original software developer, while their products are differentiated) (see Shy and Thisse (1999)), then allowing software piracy by one group (typically low-valued users) of software users could be supported as a non-cooperative equilibrium under strong network effect. At the same time, when the competition takes place between two asymmetric firms, i.e., one firm is the original software developer and the other is just a pirate (as in this case), then allowing piracy (by the pirate) is not a profitable outcome to the original firm. Therefore, protection remains the only profitable option to the original developer. One important distinction that we would like the readers to notice here is that our study is based on retail piracy (i.e., one single pirate does all the piracy and sells to others), 15 while most of the studies (except Banerjee 2003) discussed in the literature (see section 2 and above) so far, are mainly based on end-user piracy (i.e., consumers pirate copies mainly for their own use). So there is a distinct difference in the act of coping. Hence, whether the nature of piracy actually leads to alternative outcomes that remain to be seen. A future research along this will line would be desirable. Part II The Case of Sequential Move So far we have considered a simultaneous move game between the original developer and the pirate. Now we are going to consider a sequential move game where the original firm acts as a leader and the pirate as the follower. We believe this market structure is also very common in many real life situations, where the original producer is an established firm in the business and is the market leader. In such situation if any pirate comes to TLFeBOOK 188 Poddar Copyright © 2005, Idea Group Inc. Copying or distributing in print or electronic forms without written permission of Idea Group Inc. is prohibited. operate, it naturally becomes the follower. An analysis of the leader-follower game also gives us the opportunity to compare the outcomes of the simultaneous move game scenario, which we had studied in the previous part, with the sequential version of the game. This also makes our analysis on software piracy in a strategic framework rather complete. The Leader-Follower Game of Piracy Given the distribution of the buyers which we have discussed earlier (see section 3.3), the profit function of the pirate (follower) is given by: π F P = D P . P P = P P . [ ) 1( qq PqP PO − − ] (For the expression D P see section 3.3) Thus, the reaction function is given by: P F P (P O ) = 2 O qP The profit function of the original firm (leader) is: π O = D O . P O = P O . [ ()( )( ) ()() θ θ −− −−−+− 1 1 1 q PPPqPq POPO ] (for D O see 3.3) Plugging in the reaction function of the follower in the above expression, we solve for (subgame perfect) equilibrium prices: P O L = q q q θ −− − 2 1 ; P P F = () ) 2( 2 1 θ qq qq −− − (10) Equilibrium demands are given by: () θ − = 1 2 1 L O D ; () qq D F P θ −− = 2 2 1 (11) TLFeBOOK On Software Piracy 189 Copyright © 2005, Idea Group Inc. Copying or distributing in print or electronic forms without written permission of Idea Group Inc. is prohibited. The profit of the original software firm is given by: ∗ L O π = () ) 2 ( 1 2 ) 1 ( qq q θθ −−− − (12) and that of the pirate is: () () 2 24 1 qq qq F P θ π −− − = ∗ (13) Comparison between Simultaneous and Sequential Move Game When we compare the equilibrium expressions in the leader-follower game with the simultaneous move game, we get the following results. Proposition 4 (i) In the leader-follower game, the prices of the original firm and the pirate are higher compared to the simultaneous move game. Formally, P O L ≥ P O * and P O F ≥ P P * . (ii) The demand of the original firm becomes lower while the demand of the pirate becomes higher in the leader-follower game compared to the simultaneous move game. Formally, D O L < D O * and D P F ≥ D P * . Proof: (i) Follows after comparing (10) with (4) and (10) with (5) respectively. (ii) Follows after comparing (11) with (6) and (11) with (7) respectively. Proposition 5 (i) The profits for both the leader and the follower are higher than respective simultaneous Bertrand profits. Formally, * * * O L O π π ≥ and * * * P F P ππ ≥ . (ii) The original firm (leader) gets a higher profit than the pirate (follower). Formally, * * F P L O ππ ≥ . Proof: (i) Follows after comparing (12) with (8) and (13) with (9) respectively. (ii) Follows after comparing (12) with (13). Note that point (ii) needs some attention. Usually, if the strategies are strategic complements between the competitors (which is the case here), then the follower gets higher profit than the leader. Here, that is not happening since the products TLFeBOOK 190 Poddar Copyright © 2005, Idea Group Inc. Copying or distributing in print or electronic forms without written permission of Idea Group Inc. is prohibited. are vertically differentiated. The leader is selling the high quality product and the follower is selling the low quality product. The former is true when the products are horizontally differentiated. Now, we compare between the profits of the original software firm under protection and non-protection. Protection versus Non-Protection We begin with the following interesting observation: () = − = θ 12 1 L O D D * NP , where D * NP is the demand under protection. Also: * L O P = q q q θ −− − 2 1 ∗ =< NP P 2 1 , where P * NP is the monopoly price under protection. Hence, we have the following result. Proposition 6 In this leader-follower case with the presence of network externality, presence of the pirate does not make any difference to the demand of the original firm. It remains exactly the same as it was under protection, yet the price is reduced due to competition. This implies total profit of the original firm under the leader-follower game must be less than the total profit under protection. Thus, we arrive at the main result for this part. Proposition 7 In the presence of network externality given a choice between employing protection and non-protection, it is always profitable for the original software developer to protect its software, even when it is the market leader. TLFeBOOK On Software Piracy 191 Copyright © 2005, Idea Group Inc. Copying or distributing in print or electronic forms without written permission of Idea Group Inc. is prohibited. Proposition 8 The ranking of profits of the original firm in three different regimes (i.e., (i) simulta- neous move under no protection, (ii) sequential move under no protection, and (iii) protection) respectively is as follows: ∗∗∗∗ << N P L OO πππ . Thus, moving from a simultaneous move game to a sequential move game as a leader improves the original firm’s profit, yet the improved profit is still lower than the profit under protection. Hence, protection remains the optimal policy to the original developer under all circumstances. Like the simultaneous case, the following result is also true in the sequential game. Proposition 9 The original software developer has a greater incentive to protect its product in the presence of network externality as oppose to the case without any network externality. Proof: As before, the incentive to protect increases with the degree of network externality. Gain from protection under network externality is: () ( ) 2 2 34)2(12 )1( qqqq qq θθθ −−−−− − = G 1 (say). Observe that G 1 is an increasing function of θ . Q.E.D. Part III Welfare Analysis Now we are ready to do some welfare analysis. Assume that in the set up, that is, discussed in previous two parts (I and II), there is a social planner (say, the government) whose objective is to maximize society’s welfare. What would be the policy recommen- dation with respect to piracy? In other words, the question is, whether allowing piracy is welfare improving or welfare reducing from the society’s point of view. To analyze that, first we list the consumer surplus and the social welfare under various cases. TLFeBOOK 192 Poddar Copyright © 2005, Idea Group Inc. Copying or distributing in print or electronic forms without written permission of Idea Group Inc. is prohibited. Software Protection (No Piracy) 2 1 ) 1 ( 8 1 θ − = C S (14) 2 2 1 ) 1 ( 8 2 3 ) 1 ( 4 1 ) 1 ( 8 1 θ θ θ θ − − = − + − = W (15) No Software Protection (Piracy) - The Simultaneous Game 22 2222 2 )34()1(2 4 45148 qq qqqqq C S θθ θθθθ −−− +−+−+ = (16) 2 2 2 2 2 2 2 2 ) 3 4 ( ) 1 ( 2 2 2 8 11 18 8 12 qq qqqqqq W θθ θθθθθ −−− −−++−−− = (17) Comparisons Comparing between (15) and (17), we get the following result. Proposition 10 Under network externality, the society is better off with the pirate. Formally: W 2 > W 1 . Proof: It can be shown that W 2 is increasing in q for all θ ∈ (0, ½). It is also true that W 1 ( θ ) = W 2 ( θ ) q=0 . Thus, combining these two we get the result. No Software Protection (Piracy) - The Sequential Game Here again, we list the consumer surplus and the welfare for the case when the original developer is the leader in the market, while the pirate is the follower. 2 2 2 2 2 2 2 3 ) 2 ( ) 1 ( 8 4 10 2 3 qq qqqqqq C S θθ θθθθ −−− +−+−++ = (18) TLFeBOOK On Software Piracy 193 Copyright © 2005, Idea Group Inc. Copying or distributing in print or electronic forms without written permission of Idea Group Inc. is prohibited. 2 2 2 2 2 2 2 3 ) 2 ( ) 1 ( 8 3 6 7 6 9 8 12 qq qqqqqq W θθ θθθθθ −−− −+++−−− = (19) Comparisons Comparing between (15) and (19), we get the following result. Proposition 11 The society is better off with the pirate under network externality, i.e.: W 3 > W 1 . Proof: Like before, it can be shown that W 3 is increasing in q for all θ ∈ (0, ½). It is also true that W 1 ( θ ) = W 3 ( θ ) q=0 . Thus, combining these two, we get the result. Discussion Thus, in our models, we find that the existence of the pirate is always better for the society. This is true for both the simultaneous and sequential version of the game under network externality. But at the same time, we would like to warn our readers to be more careful in order to generalize this result in other situations. First of all, here we only capture a situation of retail piracy in a particular demand environment. Generally, the impact of piracy on social welfare is a far more complex issue than we captured here. For that matter, we would like to draw the readers’ attention on a comprehensive study by Chen and Png (2003) on copyright enforcement and pricing of information goods and welfare aspects. Software is one of the information goods that we are interested here. Chen and Png deal with general information goods, where the primary question was – how should the government use its various policy instruments – penalties, taxes, and subsidies – in the market for information good? This question is especially difficult because the govern- ment, in setting policy, must consider how legitimate producers will adjust their pricing and enforcement in response to government policy. So the study addresses the impact of government policy on the software publisher’s price and detection expenditure and then analyzes the consequences for social welfare. One distinct difference from their study to our study here is that in their study the pirates are end-users (consumers) and there is no retail piracy. The main findings from the study can be summarized as follows. While the publisher may consider a price reduction and an increase in detection as simply two alternative ways to boost legitimate demand, the two changes have qualitatively different welfare effects. Society prefers the publisher to manage piracy through lower prices rather than increased enforcement. Lower prices allow more people to use software, which in turn increase consumer surplus and welfare. Second, a tax on the copying medium is welfare superior to a penalty for copyright violations. Compared to the penalty, the tax has less effect on the legitimate price and leads the publisher to reduce TLFeBOOK 194 Poddar Copyright © 2005, Idea Group Inc. Copying or distributing in print or electronic forms without written permission of Idea Group Inc. is prohibited. rather than raise spending on detection. Reducing spending on detection always has a positive effect on welfare, since the cost of detection is somewhat wasteful. Lastly, it is optimal to subsidize legitimate purchases. Here also subsidy leads to reduced spending on detection. On the other hand, it is generally true that the government policies that focus only on penalties will never be an optimal choice from society’s welfare point of view. So the other policy instruments available turn out to be crucial in enhancing society’s overall welfare under the existence of software piracy. Contribution of the Study to Researchers/Instructors and Managers/Entrepreneurs Academic Purpose for Researchers and Instructors Up until now, the literature on the economics of software piracy is scattered in various directions and sometimes even with several conflicting results. This chapter puts an order to this scattered literature and connects one research agenda with the other. It explains why some results are in conflict with other results. In doing so, the study provides an in-depth analysis of certain important issues raised in the literature of software piracy. We believe the chapter will be a useful guide to an academic researcher as well as to the instructor who plans to teach a course on the economics of software piracy. Further readings on the issue of software piracy are listed in the bibliography. Business Purpose for Managers and Entrepreneurs We also believe this study will give some new insights to the business practitioners who are involved in the software industry. Every year software piracy is costing billions of dollars to the industry. So the conventional wisdom would suggest to stop piracy at any cost to save the industry. But at the same time, the legal software products are almost beyond the purchasing capability of the average software users in the developing world. So piracy remains the only way out in those regions. It is also true that easy availability of the software products in a developing country’s markets increases the know-how and the usage of software products. This in turn helps the software companies to sell their products in those markets more successfully. When the scenario is like this, it is important to the managers and entrepreneurs in the industry to come up with innovative business strategies, which are feasible as well as profitable. We believe this study will help them to formulate such business strategies. Another issue is that in the present business world, e-business and e-commerce are gradually taking the center stage, and it is needless to say that the behaviour of the software industry will have a profound effect on them as well. Managers and entrepre- TLFeBOOK [...]... passengers, networks and freight volume, and cultural events (Brunn, Williams and Zeigler, 2003; Friedman, 19 86, 1995; Short et al., 1995; Short and Kim, 1999; Knox, 1994; Knox and Taylor, 1995; Knox and Pinch, 2000; Hall 1 966 , 1984, 2001; Lo and Marcotullio, 2001; GaWC Study and Network; Smith and Timberlake, 2001; Wagenaar, Mamadouh and Dijkink, 2000) These studies are valuable in suggesting subsequent... 13,241,700 11.1 Europe – East (10) 6, 875,000 5.8 Europe – West (9) 15,230,000 13.0 South America (13) 5,783,900 4.9 Central America (8) 12,279.000 10.3 Caribbean (18) 6, 664 ,000 5 .6 Africa – West ( 16) 2,891,000 2.4 Africa – Central (9) 2,293,000 1.9 60 3,000 5 Africa – North (5) 1,229,000 1.0 Africa – East (18) 3,275,000 2.7 Asia – East (6) 5, 769 ,000 4.9 Asia – South & Central (14) 4, 065 ,000 3.4 Asia – Southeast... San Marino San Marino 1,800 Vienna Austria 1,750 Beijing China 1 ,69 0 Buenos Aires Argentina 1 ,67 0 Stockholm Sweden 1 ,63 0 Auckland New Zealand 1,570 Athens Greece 1,550 Bangkok Thailand 1,540 Delhi India 1,510 The top 15 capitals, with 46 million references, had 31% of the all capital city references The top 27 capitals had 66 million or 56% of the total Some of the cities with high rankings are mentioned... Review, 1997; Janelle and Hodge, 2000; Wilson and Corey, 2000; Leinbach and Brunn, 2001; Tijdschrift voor Economische en Sociale Geografie, 2002; Environment and Planning A, 2003), sociology (Barnett, 2001; Castells, 2001; Hargittai and Centeno, 2001; Kick and Davis, 2001; Sassen, 2001), organizational science and management (Sacks, Ventresca and Uzzi, 2001) Cities, urban regions, and networks have also... festivals in Berlin, and museums in Washington, D.C and Paris Webcam sites were provided for Moscow, Auckland, and Luxembourg Business directories were provided for Guatemala City and San Marino and stock market updates for Singapore and Madrid While most of these first pages were in English, there were also Spanish sites for Madrid and Mexico City, French sites for Paris, German for Berlin, and Chinese for... have adopted and benefited from ICT and e-commerce (Department of Communications and the Arts and AUSe.NET Australian Electronic Business Network, 1998; Papandrea, 1998; National Office for the Information Economy, 2000; Papandrea and Wade, 2000; Ernst & Young and Multimedia Victoria, 2002) These are, however, generally very brief accounts which lack detail in terms of the processes, the challenges, ... of disciplinary and interdisciplinary research initiatives (Castells, 2002; Van der Wusten, 2002) Among the topics addressed are telecommunications and the changing structures of cities (Graham and Marvin, 19 96; Wheeler, Aoyama and Warf, 2000; Brunn and Ghose, 2003), the changing geographies of the Internet (Zook, 2001; Kellerman, 2002), the most and least linked regions (Saad, House and Brunn, 2002),... Science, 37(2), 125-139 Gandal, N (1994) Hedonic Price Indexes for Spreadsheets and an Empirical Test of Network Externalities Hypothesis Rand Journal of Economics, 25, 160 -170 Katz, M and Shapiro, C (1985) Network Externalities, Competition and Compatibility American Economic Review, 75(2), 424-440 King, P S and Lampe, R (2002) Network Externalities, Price Discrimination and Profitable Piracy Mimeo:... Freetown, and Baghdad The third category has 29 cities with 26 - 50 hyperlinks per capita These are a number of capitals in large countries and mini-states in Europe, Asia and Africa Examples of cities include London, Paris, Luanda, Gaborone, Tel Aviv, Doha, Dubai, Havana, and San José The fourth category includes 16 capitals with 51 - 75 hyperlinks per capita These are mostly in southern and eastern... care) Transportation and communication linkages to and from the capital cities are also important, including direct road and rail connections to other capitals, airline connections as well as the volume of phone, mail (letters and packages), and fax traffic, the exchanges of information among members of diplomatic staff and print and visual journalists for various news organizations, and the volume of . networks and freight volume, and cultural events (Brunn, Williams and Zeigler, 2003; Friedman, 19 86, 1995; Short et al., 1995; Short and Kim, 1999; Knox, 1994; Knox and Taylor, 1995; Knox and Pinch,. Taylor, 1995; Knox and Pinch, 2000; Hall 1 966 , 1984, 2001; Lo and Marcotullio, 2001; GaWC Study and Network; Smith and Timberlake, 2001; Wagenaar, Mamadouh and Dijkink, 2000). These studies are valuable. well as small island states and political units in the Caribbean (Dominica, Curaçao, and Guadeloupe); South Indian Ocean (Mauritius and Seychelles); and Pacific Basin (Nauru, Tonga, and New Caledonia).