Strategy and game theory practice exercises with answers

Springer Texts in Business and Economics Felix Munoz-Garcia Daniel Toro-Gonzalez Strategy and Game Theory Practice Exercises with Answers Springer Texts in Business and Economics More information about this series at http://www.springer.com/series/10099 Felix Munoz-Garcia Daniel Toro-Gonzalez • Strategy and Game Theory Practice Exercises with Answers 123 Felix Munoz-Garcia School of Economic Sciences Washington State University Pullman, WA USA Daniel Toro-Gonzalez School of Economics and Business Universidad Tecnológica de Bolívar Cartagena, Bolivar Colombia ISSN 2192-4333 ISSN 2192-4341 (electronic) Springer Texts in Business and Economics ISBN 978-3-319-32962-8 ISBN 978-3-319-32963-5 (eBook) DOI 10.1007/978-3-319-32963-5 Library of Congress Control Number: 2016940796 © Springer International Publishing Switzerland 2016 This work is subject to copyright All rights are reserved by the Publisher, whether the whole or part of the material is concerned, specifically the rights of translation, reprinting, reuse of illustrations, recitation, broadcasting, reproduction on microfilms or in any other physical way, and transmission or information storage and retrieval, electronic adaptation, computer software, or by similar or dissimilar methodology now known or hereafter developed The use of general descriptive names, registered names, trademarks, service marks, etc in this publication does not imply, even in the absence of a specific statement, that such names are exempt from the relevant protective laws and regulations and therefore free for general use The publisher, the authors and the editors are safe to assume that the advice and information in this book are believed to be true and accurate at the date of publication Neither the publisher nor the authors or the editors give a warranty, express or implied, with respect to the material contained herein or for any errors or omissions that may have been made Printed on acid-free paper This Springer imprint is published by Springer Nature The registered company is Springer International Publishing AG Switzerland Preface This textbook presents worked-out exercises on Game Theory, with detailed step-by-step explanations, which both undergraduate and master’s students can use to further understand equilibrium behavior in strategic settings While most textbooks on Game Theory focus on theoretical results; see, for instance, Tirole (1991), Gibbons (1992) and Osborne (2004), they offer few practice exercises Our goal is, hence, to complement the theoretical tools in current textbooks by providing practice exercises in which students can learn to systematically apply theoretical solution concepts to different fields of Economics and Business, such as industrial economics, public policy and regulation The textbook provides many exercises with detailed verbal explanations (97 exercises in total), which cover the topics required by Game Theory courses at the undergraduate level, and by most courses at the Masters level Importantly, our textbook emphasizes the economic intuition behind the main results, and avoids unnecessary notation when possible, and thus is useful as a reference book regardless of the Game Theory textbook adopted by each instructor Importantly, these points differentiate our presentation from that found in solutions manuals Unlike these manuals, which can be rarely read in isolation, our textbook allows students to essentially read each exercise without difficulties, thanks to the detailed verbal explanations, figures, and intuitions Furthermore, for presentation purposes, each chapter ranks exercises according to their difficulty (with a letter A to C next to the exercise number), allowing students to first set their foundations using easy exercises (type-A), and then move on to harder applications (type-B and C exercises) Organization of the Book We first examine games that are required in most courses at the undergraduate level, and then advance to more challenging games (which are often the content of master’s courses), both in Economics and Business programs Specifically, Chaps 1–6 cover complete-information games, separately analyzing simultaneous-move and sequential-move games, with applications from industrial economics and regulation; thus helping students apply Game Theory to other fields of research v vi Preface Chapters 7–9 pay special attention to incomplete information games, such as signaling games, cheap talk games, and equilibrium refinements These topics have experienced a significant expansion in the last two decades, both in the theoretical and applied literature Yet to this day most textbooks lack detailed worked-out examples that students can use as a guideline, leading them to especially struggle with this topic, which often becomes the most challenging for both undergraduate and graduate students In contrast, our presentation emphasizes the common steps to follow when solving these types of incomplete information games, and includes graphical illustrations to focus students’ attention to the most relevant payoff comparisons at each point of the analysis How to Use This Textbook Some instructors may use parts of the textbook in class in order to clarify how to apply certain solution concepts that are only theoretically covered in standard textbooks Alternatively, other instructors may prefer to assign certain exercises as a required reading, since these exercises closely complement the material covered in class This strategy could prepare students for the homework assignment on a similar topic, since our practice exercises emphasize the approach students need to follow in each class of games, and the main intuition behind each step This strategy might be especially attractive for instructors at the graduate level, who could spend more time covering the theoretical foundations in class, asking students to go over the worked-out applications of each solution concept provided by our manuscript on their own In addition, since exercises are ranked according to their difficulty, instructors at the undergraduate level can assign the reading of relatively easy exercises (type-A) and spend more time explaining the intermediate level exercises in class (type-B questions), whereas instructors teaching a graduate-level course can assume that students are reading most type-A exercises on their own, and only use class time to explain type-C (and some type-B) exercises Acknowledgments We would first like to thank several colleagues who encouraged us in the preparation of this manuscript: Ron Mittlehammer, Jill McCluskey, and Alan Love Ana Espinola-Arredondo reviewed several chapters on a short deadline, and provided extremely valuable feedback, both in content and presentation; and we extremely thankful for her insights Felix is especially grateful to his teachers and advisors at the University of Pittsburgh (Andreas Blume, Esther Gal-Or, John Duffy, Oliver Board, In-Uck Park, and Alexandre Matros), who taught him Game Theory and Industrial Organization, instilling a passion for the use of these topics in applied settings which hopefully transpires in the following pages We are also thankful to Preface vii the “team” of teaching and research assistants, both at Washington State University and at Universidad Tecnologica de Bolivar, who helped us with this project over several years: Diem Nguyen, Gulnara Zaynutdinova, Donald Petersen, Qingqing Wang, Jeremy Knowles, Xiaonan Liu, Ryan Bain, Eric Dunaway, Tongzhe Li, Wenxing Song, Pitchayaporn Tantihkarnchana, Roberto Fortich, Jhon Francisco Cossio Cardenas, Luis Carlos Díaz Canedo, Pablo Abitbol, and Kevin David Gomez Perez We also appreciate the support of the editors at Springer-Verlag, Rebekah McClure, Lorraine Klimowich, and Dhivya Prabha Importantly, we would like to thank our wives, Ana Espinola-Arredondo and Ericka Duncan, for supporting and inspiring us during the (long!) preparation of the manuscript We would not have been able to it without your encouragement and motivation Felix Munoz-Garcia Daniel Toro-Gonzalez Contents Dominance Solvable Games Introduction Exercise 1—From Extensive Form to Normal form Representation-IA Exercise 2—From Extensive Form to Normal Form Representation-IIA Exercise 3—From Extensive Form to Normal Form Representation-IIIB Exercise 4—Representing Games in Its Extensive FormA Exercise 5—Prisoners’ Dilemma GameA Exercise 6—Dominance Solvable GamesA Exercise 7—Applying IDSDS (Iterated Deletion of Strictly Dominated Strategies)A Exercise 8—Applying IDSDS When Players Have Five Available StrategiesA Exercise 9—Applying IDSDS in the Battle of the Sexes GameA Exercise 10—Applying IDSDS in Three-Player GamesB Exercise 11—Finding Dominant Strategies in games with I ≥ players and with Continuous Strategy SpacesB Exercise 12—Equilibrium Predictions from IDSDS versus IDWDSB Pure Strategy Nash Equilibrium and Simultaneous-Move Games with Complete Information Introduction Exercise 1—Prisoner’s DilemmaA Exercise 2—Battle of the SexesA Exercise 3—Pareto CoordinationA Exercise 4—Cournot game of Quantity CompetitionA Exercise 5—Games with Positive ExternalitiesB Exercise 6—Traveler’s DilemmaB Exercise 7—Nash Equilibria with Three PlayersB Exercise 8—Simultaneous-Move Games with n ≥ PlayersB Exercise 9—Political Competition (Hoteling Model)B Exercise 10—TournamentsB 1 10 12 16 17 19 21 25 25 26 29 30 31 34 37 39 43 46 49 ix x Contents Exercise 11—LobbyingA Exercise 12—Incentives and PunishmentB Exercise 13—Cournot mergers with Efficiency GainsB Mixed Strategies, Strictly Competitive Games, and Correlated Equilibria Introduction Exercise 1—Game of ChickenA Exercise 2—Lobbying GameA Exercise 3—A Variation of the Lobbying GameB Exercise 4—Mixed Strategy Equilibrium with n > PlayersB Exercise 5—Randomizing Over Three Available ActionsB Exercise 6—Pareto Coordination GameB Exercise 7—Mixing Strategies in a Bargaining GameC Exercise 8—Depicting the Convex Hull of Nash Equilibrium PayoffsC Exercise 9—Correlated EquilibriumC Exercise 10—Relationship Between Nash and Correlated Equilibrium PayoffsC Exercise 11—Identifying Strictly Competitive GamesA Exercise 12—Maxmin StrategiesC 52 54 56 61 61 62 67 71 73 75 78 80 83 86 95 97 101 Sequential-Move Games with Complete Information Introduction Exercise 1—Ultimatum Bargaining GameB Exercise 2—Electoral competitionA Exercise 3—Electoral Competition with a TwistA Exercise 4—Trust and Reciprocity (Gift-Exchange Game)B Exercise 5—Stackelberg with Two FirmsA Exercise 6—First- and Second-Mover Advantage in Product Differentiation Exercise 7—Stackelberg Game with Three Firms Acting SequentiallyA Exercise 8—Two-Period Bilateral Bargaining GameA Exercise 9—Alternating Bargaining with a TwistB Exercise 10—Backward Induction in Wage NegotiationsA Exercise 11—Backward Induction-IB Exercise 12—Backward Induction-IIB Exercise 13—Moral Hazard in the WorkplaceB 107 107 108 110 113 115 117 122 125 127 129 130 132 134 137 Applications to Industrial Organization Introduction Exercise 1—Bertrand Model of Price CompetitionA Exercise 2—Bertrand Competition with Asymmetric CostsB Exercise 3—Duopoly Game with A Public FirmB Exercise 4—Cournot Competition with Asymmetric CostsA 145 145 146 151 154 157 330 10 More Advanced Signaling Games Stay Sane inc Sane inc prob.=p Exit Entrant Nature Stay Crazy inc prob.=(1-p) Crazy inc Exit Fig 10.12 Separating strategy profile (with responses) Stay Sane inc Sane inc prob.=p Exit Entrant Nature Crazy inc prob.=(1-p) Stay Crazy inc Exit Fig 10.13 Pooling strategy profile Part (c) Pooling PBE Let us now examine if a pooling strategy profile, where both types of incumbent fight, can be supported as a PBE Fig 10.13 illustrates this strategy profile First, in this strategy profile the entrant’s beliefs cannot be updated using Bayes’ rule, and must coincide with his priors, i.e., l ¼ p Intuitively, since both types of incumbent are choosing to fight, the entrant cannot refine its beliefs about the incumbent’s type upon observing that the incumbent fought In this case, the entrant chooses to stay if and only if pðaE f E ị ỵ pị2f E ị ! pf E ị ỵ pịf E ị ẳ f E or alternatively, if p ! f fỵE a  p: We separately consider the case in which the E E entrant responds staying ðp ! pÞ and the case in which it responds exiting ðp\pÞ Exercise 5—Entry Deterrence Through Price WarsA 331 Stay Sane inc Sane inc prob.=p Exit Entrant Nature Stay Crazy inc prob.=(1-p) Crazy inc Exit Fig 10.14 Pooling strategy profile (with responses)—Case Stay Sane inc Sane inc prob.=p Exit Entrant Nature Crazy inc prob.=(1-p) Stay Crazy inc Exit Fig 10.15 Pooling strategy profile (with responses)—Case  Figure 10.14 depicts the case in which the entrant responds Case p ! p staying (blue shaded arrows) In this setting, the sane incumbent would choose to Fight1 (as prescribed) if aI À f I ! 2aI , or alternatively ÀfI ! aI , which cannot hold, since aI [ and À fI \0 by definition Hence, the pooling strategy profile in which both types of incumbent fight cannot be sustained as a PBE when p ! p: Case 2, p\ p Let us next check the case in which p\p, which indicates that the uninformed entrant responds exiting, as Fig 10.15 depicts (see blue shaded branches) In this context, the sane incumbent chooses Fight1 (as prescribed) if and only if m À f I ! 2aI Hence, as long as this condition holds, the pooling strategy profile in which both types of incumbents fight can be sustained as a PBE if the prior probability of the incumbent being sane, p, is sufficiently low, i.e., p\p 332 10 More Advanced Signaling Games Exercise 6—Entry Deterrence with a Sequence of Potential EntrantsC The following entry model is inspired on the original paper of Kreps and Wilson (1982)8 Consider an incumbent monopolist building a reputation as a tough competitor who does not allow entry without a fight The entrant first decides whether to enter the market, and, if he does, the monopolist chooses whether to fight or acquiesce If the entrant stays out, the monopolist obtains a profit of a > 1, and the entrant gets If the entrant enters, the monopolist gets from fighting and −1 from acquiescing if he is a “tough” monopolist, and −1 from fighting and from acquiescing if he is a “normal” monopolist The entrant obtains a profit of b if the monopolist acquiesces and b − if he fights, where 0\b\1 Suppose the entrant believes the monopolist to be tough (normal) with probability p (1 À p, respectively), while the monopolist observes his own type Part (a) Depict a game tree representing this incomplete information game Part (b) Solve for the PBEs game tree, and solve for the PBE of this game Part (c) Suppose the monopolist faces two entrants in sequence, and the second entrant observes the outcome of the first game (there is no discounting) Depict the game tree, and solve for the PBE [Hint: You can use backward induction to reduce the game tree as much as possible before checking for the existence of separating or pooling PBEs For simplicity, focus on the case in which prior beliefs satisfy p b:] Answer Part (a) See Fig 10.16 Part (b) The tough monopolist fights with probability 1, since fight is a dominant strategy for him; while the normal monopolist accommodates with probability 1, since accommodation constitutes a dominant strategy for this type of incumbent Indeed, at the node labeled with MonopT on the left-hand side of the game tree in Fig 10.16, the monopolist’s payoff from fighting, 0, is larger than from accommodating, −1 In contrast, at the node labeled with MonopN for the normal monopolist (see right-hand side of Fig 10.16), the monopolist’s payoff from fighting, −1, is strictly lower than from accommodating, Hence, the entrant’s decision on whether or not to enter will be based on: EUE ðEnterjpÞ ẳ pb 1ị ỵ pịb ẳ b p; and EUE Not Enterjpị ẳ Therefore, the entrant enters if and only if b À p [ 0, or alternately, b [ p We can, hence, summarize the equilibrium as follows: The entrant enters if b [ p, but doesn’t enter if b p: The incumbent fights if tough, but accommodates if normal Kreps, David and Robert Wilson (1982) ``Reputation and Imperfect Information,'' Journal of Economic Theory, 27, pp 253-279 Exercise 6—Entry Deterrence with a Sequence of Potential EntrantsC 333 Nature Monopolist is Tough Monopolist is Normal Prob=p Prob=1-p Entrant p NE 1-p NE E E a a Fight Fight Acc -1 b-1 b-1 Acc b -1 b Fig 10.16 Entry deterrence with only one entrant Part (c) The following game tree depicts an entry game in which the incumbent faces two entrants in sequence (see Fig 10.17) Hence, applying backward induction on the proper subgames (those labeled with MT and MN in the last stages of the game tree) we can reduce the previous game tree to that in Fig 10.18 For instance, after the second entrant chooses to enter despite observing a fight with the first entrant (left side of game tree), the tough monopolist chooses between fighting and accommodating in the node labeled MT In this case, the tough monopolist prefers to fight, which yields a payoff of zero, rather than accommodate, which entails a payoff of −1 A similar analysis applies to the other node labeled with MT (where the second entrant has entered after observing that the incumbent accommodated the first entrant) However, an opposite argument applies for the nodes marked with MN in the right-hand side of the tree, where the normal monopolist prefers to accommodate the second entrant, regardless of whether he fought or accommodated the first entrant, since his payoff from accommodating (zero) is larger than from fighting (−1) 334 10 More Advanced Signaling Games Nature MonopN MonopT Entrant1 1-p p NE NE UM U E1 U E2 a 0 Monop T a 0 E E Fight Fight Acc Acc Entrant2 after Fight q ne Monop N r e 1-q Entrant after Acc ne 1-r -1+a b-1 Fight ne a b-1 Fight b-1 b-1 e e ne -1+a b Acc -1 b b Fight b-1 b-1 Acc -1 b-1 b-1 a b -1 b b e Fight Acc -1 b-1 b-1 b b Acc b b Fig 10.17 Entry deterrence with two entrants in sequence In addition, note that the first entrant behaves in exactly the same way as in exercise (a): entering if and only if b [ p Hence, when p b, the first entrant enters, as shown in exercise (a) Figure 10.19 shades this choice of the first entrant (green shaded branches) Therefore, upon entry, the first entrant gives rise to a beer-quiche type of signaling game, which can be more compactly represented as the game tree in Fig 10.20 Intuitively, all elements before MonopT and MonopN can be predicted (i.e., the first entrant enters as long as p b), while the subsequent stages characterize a signaling game between the monopolist (privately informed about its type) and the second entrant In this context, the monopolist uses his decision to fight or accommodate the first entrant as a message to the second entrant, in order to convey or conceal his type Exercise 6—Entry Deterrence with a Sequence of Potential EntrantsC 335 Nature p 1-p E NE NE a 0 E a 0 Fight Fight Acc Acc 1-q q ne e r ne a b-1 0 b-1 b-1 -1+a b ne 1-r e ne e a b 0 b b b-1 b-1 -1+a b-1 e b b Fig 10.18 Reduced-form game (For every triplet of payoffs, the first corresponds to the monopolist, the second to the first entrant, and the third to the second entrant.) Let us now check if a pooling strategy profile in which both types of monopolists accommodate can be sustained as a PBE Pooling PBE with Acc Figure 10.21 shades the braches corresponding to such a pooling strategy In this setting, posterior beliefs cannot be updated using Bayes’ rule, which entails r ¼ p As in similar exercises, the observation of accommodation by the uninformed entrant does not allow him to further refine his beliefs about the monopolist’s type Hence, the second entrant responds entering (e) after observing that the monopolist accommodates (in equilibrium), since pðb À 1Þ þ ð1 À pÞb [ p Á þ ð1 À pÞ Á $ b [ p which holds in this case 336 10 More Advanced Signaling Games Nature p 1-p E NE NE a 0 E a 0 Fight Fight Acc Acc 1-q q ne e r ne a b-1 0 b-1 b-1 -1+a b e ne 1-r e ne e a b 0 b b b-1 b-1 -1+a b-1 0 b b Fig 10.19 Entry of the first potential entrant e r 0,b,b a,b,0 ne e ne Entrant2 after Acc -1+a,b,0 Acc 1-r MonopT Fight MonopT (prob.=p) Nature MonopN (prob.=1-p) Acc MonopN Fight e 0,b-1,b-1 ne a,b-1,0 q Entrant2 after Fight 0,b-1,b-1 e 0,b,b 1-q ne -1+a,b-1,0 Fig 10.20 A further reduction of the game If, in contrast, the second entrant observes the off-the-equilibrium message of fight, then this player also enters if qb 1ị ỵ qịb [ q ỵ qị $ b [ q Exercise 6—Entry Deterrence with a Sequence of Potential EntrantsC e r 0,b,b a,b,0 Entrant2 after Acc -1+a,b,0 ne e ne MonopT Acc 1-r Fight MonopT (prob.=p) Nature MonopN (prob.=1-p) Acc MonopN Fight e 0,b-1,b-1 ne a,b-1,0 q Entrant2 after Fight 0,b-1,b-1 337 e 0,b,b 1-q ne -1+a,b-1,0 Fig 10.21 Pooling strategy profile—Acc e r 0,b,b a,b,0 ne e ne Entrant2 after Acc -1+a,b,0 Acc 1-r MonopT Fight MonopT (prob.=p) Nature MonopN (prob.=1-p) Acc MonopN Fight e 0,b-1,b-1 ne a,b-1,0 q Entrant2 after Fight 0,b-1,b-1 e 0,b,b 1-q ne -1+a,b-1,0 Fig 10.22 Pooling strategy profile—Acc (with responses) Hence, the second entrant enters regardless of the incumbent’s action if off-the-equilibrium beliefs, q, satisfy q\b, as depicted in Fig 10.22 (see blue shaded branches in the right-hand side of the figure) Otherwise, the entrant only enters after observing the equilibrium message of Acc The tough monopolist, MT , is hence indifferent between Acc (as prescribed) which yields a payoff of 0, and deviate to Fight, which also yields a payoff of zero A similar argument applies to the normal monopolist, MN , in the lower part of the game tree Hence, the pooling strategy profile where both types of incumbents accommodate can be sustained as a PBE A remark on the Intuitive Criterion Let us next show that the above pooling equilibrium, despite constituting a PBE, violates the Cho and Kreps’ (1987) Intuitive Criterion In particular, the tough monopolist has incentives to deviate towards Fight if, by doing so, he is identified as a tough player, q ¼ 1, which induces the entrant to respond not entering In this case, the tough monopolist obtains a payoff of a, which exceeds his equilibrium payoff of In contrast, the normal monopolist doesn’t have incentives to deviate since, even if his deviation to Fight deters entry, his payoff from doing so, −1 + a, would still be lower than his equilibrium payoff of 0, given that −1 + a < or a < Hence, only the tough monopolist has 338 e r 0,b,b a,b,0 Entrant2 after Acc -1+a,b,0 ne e ne Acc 1-r MonopT Fight MonopT (prob.=p) Nature MonopN (prob.=1-p) Acc MonopN More Advanced Signaling Games Fight e 0,b-1,b-1 ne a,b-1,0 q Entrant2 after Fight 0,b-1,b-1 10 e 0,b,b 1-q ne -1+a,b-1,0 Fig 10.23 Pooling strategy profile—Acc (with responses) incentives to deviate, and the entrant’s off-the-equilibrium beliefs can thus be restricted to q = upon observing that the monopolist fights Intuitively, the entrant infers that the observation of Fight can only originate from the tough monopolist In this case, the tough incumbent indeed prefers to select Fight, thus implying that the above pooling PBE violates Cho Kreps’ (1983) Intuitive Criterion.(Q.E.D) Let us finally check if this pooling strategy profile can be sustained when off-the-equilibrium beliefs satisfy, instead, q ! b, thus inducing the entrant to respond not entering upon observing the off-the-equilibrium message of Fight, as illustrated in the game tree of Fig 10.23 (see blue shaded branches in the right-hand side of the tree) In this case, the MT has incentives to deviate from Acc, and thus the pooling strategy profile where both MT and MN select to Acc cannot be sustained as a PBE Pooling PBE with Fight Let us now examine the opposite pooling strategy profile (Fight, Fight), in which both types of monopolists fight, as depicted in Fig 10.24 Hence, equilibrium beliefs after observing Fight cannot be updated, and thus satisfy q = p; while off-the-equilibrium beliefs are arbitrary r ½0; 1Š after observing the off-the-equilibrium message of accommodation Given these beliefs, e r 0,b,b a,b,0 ne e ne Entrant2 after Acc -1+a,b,0 Acc 1-r MonopT Fight MonopT (prob.=p) Nature MonopN (prob.=1-p) Acc MonopN Fig 10.24 Pooling strategy profile—Fight Fight e 0,b-1,b-1 ne a,b-1,0 q Entrant after Fight 0,b-1,b-1 e 0,b,b 1-q ne -1+a,b-1,0 Exercise 6—Entry Deterrence with a Sequence of Potential EntrantsC e r 0,b,b a,b,0 ne e ne Entrant2 after Acc -1+a,b,0 Acc 1-r MonopT Fight MonopT (prob.=p) Nature MonopN (prob.=1-p) Acc MonopN Fight e 0,b-1,b-1 ne a,b-1,0 q Entrant2 after Fight 0,b-1,b-1 339 e 0,b,b 1-q ne -1+a,b-1,0 Fig 10.25 Pooling strategy profile—Fight (with responses) upon observing the equilibrium message of Fight, the entrant responds entering since pð b 1ị ỵ p b [ p ỵ pị $ b [ p which holds by definition If, in contrast, the entrant off-the-equilibrium message of Acc, then it responds entering if observes the r b 1ị ỵ r b [ p ỵ r Þ Á $ b [ r Hence, if b > r, the entrant enters both after observing Fight (in equilibrium) and Acc (off-the-equilibrium path) If, instead, b r, then the entrant only responds entering after observing Fight, but is deterred from the industry otherwise We next separately analyze each case Case Figure 10.25 illustrates the entrant’s responses when b>r, and thus the second entrant enters after Acc In this context, the tough monopolist is indifferent between Fight, obtaining zero profits, and accommodating, which also yields zero profits A similar argument applies to the normal monopolist in the lower part of the game tree Hence, in this case the pooling strategy profile (Fight, Fight) can be supported as a PBE if off-the-equilibrium beliefs, r, satisfy r < b Case If, instead, r ! b, then the entrant is deterred upon observing the off-the-equilibrium message of Acc, as Fig 10.26 depicts The pooling strategy profile cannot be sustained in this context, since MN has incentives to deviate towards Acc, obtaining a payoff of a, which exceeds its payoff of zero when he Fights Hence, the pooling strategy profile (Fight, Fight) cannot be supported as a PBE when off-the-equilibrium beliefs satisfy r ! b: Separating PBE (Fight, Acc) Let us next examine if the separating strategy profile in which only the tough monopolist fights can be sustained as a PBE Figure 10.27 depicts this strategy profile In this case, entrant’s beliefs are updated to q = and r = using Bayes’ rule, implying that, upon observing Fight, the entrant is deterred from the market since b – < 0, given that b < by definition Upon observing Acc, the entrant is 340 e r 0,b,b a,b,0 Entrant2 after Acc -1+a,b,0 ne e ne Acc 1-r MonopT Fight MonopT (prob.=p) Nature MonopN (prob.=1-p) Acc Fight MonopN More Advanced Signaling Games e 0,b-1,b-1 ne a,b-1,0 q Entrant2 after Fight 0,b-1,b-1 10 e 0,b,b 1-q ne -1+a,b-1,0 Fig 10.26 Pooling strategy profile—Fight (with responses) e r 0,b,b a,b,0 Entrant2 after Acc -1+a,b,0 ne e ne Acc 1-r MonopT Fight MonopT (prob.=p) Nature MonopN (prob.=1-p) Acc MonopN Fight e 0,b-1,b-1 ne a,b-1,0 q Entrant after Fight 0,b-1,b-1 e 0,b,b 1-q ne -1+a,b-1,0 Fig 10.27 Separating strategy profile (Fight, Acc) e r 0,b,b a,b,0 ne e ne Entrant2 after Acc -1+a,b,0 Acc 1-r MonopT Fight MonopT (prob.=p) Nature MonopN (prob.=1-p) Acc MonopN Fight e 0,b-1,b-1 ne a,b-1,0 q Entrant2 after Fight 0,b-1,b-1 e 0,b,b 1-q ne -1+a,b-1,0 Fig 10.28 Separating strategy profile (Fight, Acc), with responses instead attracted to the market since b > Figure 10.28 illustrates the entrant’s responses (see blue shaded arrows) In this setting, no type of monopolist has incentives to deviate: (1) the tough monopolist obtains a payoff of a by fighting (as prescribed) but only from deviating towards Acc; and similarly (2) the normal monopolist obtains by Exercise 6—Entry Deterrence with a Sequence of Potential EntrantsC e r 0,b,b a,b,0 Entrant2 after Acc -1+a,b,0 ne e ne Acc 1-r MonopT Fight MonopT (prob.=p) Nature MonopN (prob.=1-p) Acc Fight MonopN e 0,b-1,b-1 ne a,b-1,0 q Entrant2 after Fight 0,b-1,b-1 341 e 0,b,b 1-q ne -1+a,b-1,0 Fig 10.29 Separating strategy profile (Acc, Fight) e r 0,b,b a,b,0 ne e ne Entrant2 after Acc -1+a,b,0 Acc 1-r MonopT Fight MonopT (prob.=p) Nature MonopN (prob.=1-p) Acc MonopN Fight e 0,b-1,b-1 ne a,b-1,0 q Entrant2 after Fight 0,b-1,b-1 e 0,b,b 1-q ne -1+a,b-1,0 Fig 10.30 Separating strategy profile (Acc, Fight), with responses accommodating (as prescribed) but a negative payoff, −1 + a, by deviating towards Fight, given that a < by definition Hence, this separating strategy profile can be sustained as a PBE Separating PBE (Acc, Fight) Let us now check if the alternative separating strategy profile, in which only the normal monopolist fights, can be supported as a PBE (We know that this strategy profile sounds crazy, but we want to formally show that it cannot be sustained as a PBE.) Figure 10.29 illustrates this strategy profile In this setting, the entrant’s beliefs can be updated to r = and q = 0, inducing the entrant to respond not entering after observing Acc, but entering after observing Fight, as depicted in Fig 10.30 (see blue shaded branches) Given these responses by the entrant, the tough monopolist has incentives to deviate from Acc, which yields a negative payoff of −1 + a, to Fight, which yields a higher payoff of zero Therefore, this separating strategy cannot be supported as a PBE Index A All-pay auction, 237, 246, 251, 252 Auctions, 237–239, 244–247, 251, 252, 254, 255 Anti-coordination game, 17, 63, 68, 95, 97 B Backward induction, 107–109, 125, 129, 130, 132, 136, 171, 332, 333 Bargaining game, 80, 82, 108, 109, 127, 129, 202, 203 Battle of the Sexes game, 16, 26, 29–31, 62, 100, 101 Bayesian Nash Equilibria (BNE), 217 Bayes’ rule, 92, 96, 257, 263, 265, 279, 282, 290, 306, 330, 339 Belief updating, 257, 265, 267–271, 274, 339, 341 Beliefs, 257, 258, 260–262, 264, 265, 267, 269–271, 273, 277, 280–282, 293, 300, 302, 312, 324, 329, 335, 341 Bertrand game, 47, 118, 145, 148, 150, 153, 154, 195, 233 Best response function, 20, 32, 34, 36, 38, 39, 51, 52, 55, 57, 66, 68, 70, 85, 117, 123, 124, 126, 130, 148, 154, 155, 158, 161, 169, 175, 209, 214, 285, 309 Best responses, 25, 26, 29, 41, 44, 45, 78, 83, 87, 115, 135, 197, 204, 220, 223, 225 Blume, 154 Branch (of a game tree), 7, 133, 270, 272, 274, 275, 307, 338 C Cartel, 183, 188–194 Cheating (in repeated game), 185, 186 Cho and Kreps’ (1987) intuitive criterion, 258, 298–303, 322, 325, 337 Collusion, 188, 192, 194, 196, 208, 211, 215 Competition in prices, 47, 121, 124, 145, 146, 151, 154, 195 Competition in quantities, 26, 31–34 Complete information, 25, 26, 145, 218, 230, 231, 303, 308, 309, 327 Conditional probability, 92, 93, 277, 291 Concavity, 131 Continuous action, 151 Convex hull of equilibrium payoffs, 85, 86, 91 Convexity, 206 Cooperation, 183–187, 191, 195, 211, 212 Coordination games, 17, 30, 53, 78, 309 Correlated equilibrium, 62, 83, 85, 86, 91, 93, 95–97 Cournot game, 26, 31, 32, 118, 145, 160, 188, 192, 218 Credible punishment, 192 D Defect (in repeated games), 183, 185–187, 207 Deviation, 38, 49, 146, 147, 150–154, 183, 184, 190, 192, 193, 195, 196, 207–209, 212, 215, 293 Direct approach, 38, 239, 241–243, 247–249 Direct demand, 122, 124, 213, 283 Discrete action, 151 Dominance solvable games, 1, Duopolists, 56 Duopoly, 145–147, 152, 156, 157, 171, 230 E Efficiency gains, 26, 56, 59 Entry deterrence, 327–329, 332–334 © Springer International Publishing Switzerland 2016 F Munoz-Garcia and D Toro-Gonzalez, Strategy and Game Theory, Springer Texts in Business and Economics, DOI 10.1007/978-3-319-32963-5 343 344 Envelope theorem, 161, 168, 237, 239, 241, 242, 247–249 Expected utility, 64–66, 74, 96, 101–103, 105, 106, 139, 199, 204, 228, 236, 239, 251, 263, 280, 282, 291, 317, 319 Extensive-form game, 2, F Finitely-repeated games, 184 First-order statistic, 248 Folk theorem, 184, 207 G Game of Chicken, 62, 64, 68 Game tree, 1, 2, 4, 5, 7, 80, 110, 132, 133, 138, 221, 259, 262, 270, 274, 293, 327, 332, 333, 338 Gift-Exchange game, 115, 116 Grim-trigger strategy, 207 H Heterogeneous products, 151, 152 Homogeneous products, 167, 208, 211 Hoteling, 46 I Incentives and punishment, 54 Incomplete Information, 217, 218, 223, 229, 231, 237, 258, 303, 308, 312, 332 Inequity aversion, 107 Infinitely-repeated games, 184, 188, 204, 207, 208 Information set, 4, 7–9, 259, 261, 303–306, 312 Initial node, 2, 5, 7, Inverse demand, 31, 117, 118, 125, 146, 151, 157, 160, 163, 164, 167, 179, 212 Iterative Deletion of Strictly Dominated Strategies (IDSDS), 1, 10, 11, 13, 18, 19, 226 Iterative Deletion of Weakly Dominated Strategies (IDWDS), 13, 22 K Kreps, 258, 298–304, 322, 325, 326, 332, 337, 338 L Labor market signaling, 267, 302, 312 Lobbying, 20, 52, 67–72 M Matching pennies game, 98, 99 Maxmin strategies, 62, 101, 102 Index Mergers, 56, 59, 179, 180 Mixed strategy, 61, 62, 65, 68, 69, 82, 145, 148, 152, 198, 199, 310 Mixed Strategy Nash Equilibrium (msNE), 62, 65, 69, 148, 152, 198 Monopolist, 190, 211, 213, 297, 332–335, 337–341 Monopoly, 32, 118, 121, 122, 147, 209, 212, 213 Moral hazard, 107, 137, 138, 140, 142 N Nash equilibrium, 12, 17, 25, 29, 34, 38, 44, 47–49, 55, 62, 66, 83, 86, 90, 101, 123, 135, 136, 146, 152, 157, 184, 192, 198, 204, 207, 225, 231, 238, 310 Nash reversion (in repeated games), 185 Node (of a game tree), 4, 5, 7–9, 135, 138, 259, 261, 268, 292, 329, 333 Normal-form game, 3, 4, 6, 10–15, 19, 21–23, 40–43, 45, 84, 87, 102 O Outcomes of the game, Oligopolists, 168, 188, 208, 209, 283 Oligopoly, 168, 183, 191 P Pareto coordination game, 30, 53, 78, 310 Pareto optimality, 73 Payoffs, 3, 8, 9, 17, 28, 29, 31, 41, 44, 46, 63, 75, 77, 83, 85, 86, 89, 90, 95, 103, 108, 113, 129, 135, 136, 183, 188, 197, 200, 203, 204, 206, 208, 220, 227, 276 Perfect Bayesian Equilibrium (PBE), 257, 258, 317 Pooling equilibria, 258, 299, 304 Posterior beliefs, 308, 314, 335 Price discrimination, 294, 298 Prior beliefs, 332 Prisoners’ dilemma game, 8, Private firms, 154 Private information, 284, 298, 309 Probability, 48, 54, 62, 64, 65, 68, 69, 72, 74, 76, 78, 92, 93, 96, 105, 150, 200, 218, 226, 233, 234, 238, 241, 252, 255, 265, 276, 289, 306, 307, 318, 327, 332 Product differentiation, 163, 211 Profit maximization, 118, 124, 130, 171, 177 Profitable deviation, 38, 49, 93, 146, 147, 150–154, 214 Public firms, 145, 154–156 Pure strategies, 61, 66, 69, 76, 89, 105, 199, 221, 226 Index Pure Strategy Nash Equilibrium (psNE), 25, 35, 37, 39, 44, 47, 53, 61, 62, 64, 66, 78, 152, 157, 197, 201 Q Quantity competition, 26, 31, 32, 107, 117, 154 R Repeated games, 183, 185, 188, 197, 204, 207, 208 Reputation, 304, 332 Risk, 115, 127, 237, 246, 253–255 Risk aversion, 237, 245, 246, 253, 254 Rock-paper-scissors game, S Semi-separating equilibrium, 289, 290, 292, 293 Separating equilibria, 259, 261, 314–316, 322, 326, 327 Sequence of entrants, 332 Sequential-move game, 80, 110, 117, 129, 138, 257 Sequential rationality, 107 Signaling, 257, 258, 267, 288, 298, 299, 303, 322, 334 Singh, 164 Simultaneous-move games, 18, 25, 43, 217 Social welfare, 28, 145, 154, 155, 180, 181 Spence’s (1974) job market signaling game, 258, 267, 303 345 Stackelberg game, 107, 117, 118, 120, 125, 127 Strategic complements, 36, 123 Strategic substitutes, 32 Strictly competitive game, 61, 62, 97, 98, 100–103 Strictly dominated strategies, 1, 2, 10, 11, 13, 16, 18, 21, 198, 226 Subgame Perfect Nash equilibria (SPNE), 107, 110, 111, 115, 117, 119, 127, 131, 138, 184, 201 Symmetric strategy profiles, 37, 47, 48 T Terminal nodes, 8, 132, 135, 138 Traveler’s dilemma game, 31, 37 U Ultimatum bargaining game, 108–110 Uncertainty, 283 Utility function, 53, 54, 108, 116, 117, 138, 239, 246, 253, 294 V Vives, 164 W Weakly dominated strategies, 2, 13, 15, 21–23 Wilson, 304, 332 ... Business and Economics More information about this series at http://www.springer.com/series/10099 Felix Munoz-Garcia Daniel Toro-Gonzalez • Strategy and Game Theory Practice Exercises with Answers. .. Publishing Switzerland 2016 F Munoz-Garcia and D Toro-Gonzalez, Strategy and Game Theory, Springer Texts in Business and Economics, DOI 10.1007/978-3-319-32963-5_1 Dominance Solvable Games Finally,... Battle of the Sexes GameA Exercise 10—Applying IDSDS in Three-Player GamesB Exercise 11—Finding Dominant Strategies in games with I ≥ players and with Continuous Strategy SpacesB