Game Theory and Economic Behavior Unit 9 ReviewCooperative Games: Coalitions & Bargaining

Key Concepts

• Cooperative games involve players working together to achieve a common goal or maximize their collective payoff
• Coalitions are groups of players who agree to cooperate and coordinate their strategies
• Characteristic function $v(S)$ assigns a value to each coalition $S$, representing the total payoff the coalition can achieve
• Transferable utility (TU) games allow players to freely transfer payoffs among themselves, while non-transferable utility (NTU) games have restrictions on payoff transfers
• Shapley value is a solution concept that assigns a unique payoff to each player based on their marginal contributions to all possible coalitions
• Core is another solution concept that represents the set of payoff allocations that no coalition can improve upon
• Bargaining models, such as the Nash bargaining solution and Rubinstein bargaining model, analyze how players negotiate and reach agreements

Cooperative Game Theory Basics

• Cooperative game theory focuses on how players can work together to achieve better outcomes than acting individually
• In cooperative games, players can form binding agreements and make joint decisions
• The main elements of a cooperative game include the set of players, the characteristic function, and the payoff allocation
• The characteristic function $v(S)$ assigns a value to each coalition $S$, representing the total payoff the coalition can achieve
  • For example, in a three-player game, $v({1,2}) = 10$ means that players 1 and 2 can achieve a total payoff of 10 by forming a coalition
• Superadditivity is a property of cooperative games where the value of a union of disjoint coalitions is at least as great as the sum of their individual values
• Convexity is a stronger property than superadditivity, requiring that the marginal contribution of a player to a coalition increases as the coalition grows
• The core of a cooperative game is the set of payoff allocations that no coalition can improve upon by deviating and forming their own coalition

Coalition Formation

• Coalition formation is the process by which players decide to work together and coordinate their strategies
• Players form coalitions to increase their bargaining power, share resources, or achieve a common goal
• The stability of a coalition depends on factors such as the payoffs, the ability to make binding agreements, and the presence of externalities
• The grand coalition is the coalition that includes all players in the game
  • In some games, the grand coalition may not always form due to incentives for players to deviate and form smaller coalitions
• Coalitional rationality requires that the payoffs allocated to each player in a coalition are at least as much as they could achieve by forming their own coalition
• The core of a cooperative game represents the set of payoff allocations that are coalitionally rational and stable
• Dynamic coalition formation models consider the process of coalition formation over time, allowing for the possibility of players joining or leaving coalitions

Types of Coalitions

• Winning coalitions are those that can achieve the desired outcome or goal of the game
  • For example, in a voting game, a winning coalition is one that has enough votes to pass a proposal
• Blocking coalitions are those that can prevent a winning coalition from forming by refusing to cooperate
• Minimal winning coalitions are winning coalitions that cannot lose any member without losing their winning status
• Oversized coalitions are winning coalitions that have more members than necessary to achieve the desired outcome
  • Oversized coalitions may form to increase stability or to distribute payoffs among more players
• Connected coalitions are those where all members are directly or indirectly connected to each other through communication or cooperation links
• Balanced collections of coalitions are sets of coalitions where each player appears in the same number of coalitions, weighted by the coalitions' values

Bargaining Models

• Bargaining models analyze how players negotiate and reach agreements in cooperative games
• The Nash bargaining solution is a classic model that satisfies several desirable properties, such as Pareto optimality, symmetry, and independence of irrelevant alternatives
  • It maximizes the product of the players' gains relative to their disagreement payoffs
• The Rubinstein bargaining model considers a sequential bargaining process where players make alternating offers until an agreement is reached
  • The model predicts that the player who makes the first offer has an advantage, and the agreed-upon payoff division depends on the players' discount factors and the time between offers
• The Kalai-Smorodinsky bargaining solution satisfies monotonicity, individual rationality, and a fairness property called restricted proportionality
  • It selects the Pareto optimal point where the ratio of the players' gains relative to their ideal payoffs is equal
• Bargaining models can incorporate various factors, such as outside options, asymmetric information, and deadlines, to analyze their impact on the negotiation process and outcomes

Solution Concepts

• Solution concepts in cooperative game theory predict or prescribe how players should divide the payoffs among themselves
• The Shapley value assigns a unique payoff to each player based on their average marginal contribution to all possible coalitions
  • It satisfies desirable properties such as efficiency, symmetry, linearity, and null player property
• The core is the set of payoff allocations that no coalition can improve upon by deviating and forming their own coalition
  • It represents stable outcomes where no group of players has an incentive to break away from the grand coalition
• The nucleolus is another solution concept that minimizes the maximum dissatisfaction of any coalition
  • It selects the payoff allocation that lexicographically minimizes the excess of each coalition, where excess is the difference between a coalition's value and the sum of its members' payoffs
• The kernel is a solution concept that ensures that each player's maximum surplus over another player is equal to the other player's maximum surplus over them
  • It represents a form of bilateral stability between any pair of players
• The bargaining set is a generalization of the core that allows for objections and counter-objections between players, capturing a notion of credible threats in the negotiation process

Applications in Economics

• Cooperative game theory has numerous applications in economics, including the analysis of market power, resource allocation, and public goods provision
• In the context of market power, cooperative game theory can model the formation of cartels or mergers among firms to increase their profits or market share
  • The stability of such agreements depends on factors such as the distribution of profits, the threat of entry, and antitrust regulations
• Resource allocation problems, such as the division of a common resource or the sharing of costs and benefits in a joint project, can be modeled as cooperative games
  • Solution concepts like the Shapley value or the core can provide insights into fair and stable allocations
• Public goods provision can be analyzed using cooperative game theory, considering the incentives for individuals to contribute to the public good and the possibility of free-riding
  • The core and other solution concepts can help design mechanisms to ensure the efficient provision of public goods
• Cooperative game theory has also been applied to problems in operations research, such as supply chain management, inventory pooling, and transportation network design
  • By modeling the interactions among different agents in these systems, cooperative game theory can help identify optimal strategies and stable allocations of costs and benefits

Challenges and Limitations

• One challenge in cooperative game theory is the assumption of binding agreements and the ability to enforce cooperation among players
  • In reality, players may have incentives to deviate from agreements or engage in non-cooperative behavior
• The computational complexity of finding solution concepts, such as the Shapley value or the nucleolus, can be high for games with a large number of players or coalitions
  • Approximation algorithms and heuristics may be needed to obtain solutions in practice
• Cooperative game theory often assumes complete information about the characteristic function and the players' preferences
  • In many real-world situations, players may have incomplete or asymmetric information, which can affect the formation and stability of coalitions
• The assumption of transferable utility may not always hold in practice, as there may be restrictions on how payoffs can be divided among players
  • Non-transferable utility games, which allow for more general utility functions and preferences, can be more challenging to analyze
• Cooperative game theory typically focuses on static, one-shot interactions among players
  • Dynamic models that consider the evolution of cooperation and competition over time may be needed to capture more realistic scenarios
• The choice of solution concept can have a significant impact on the predicted outcomes and policy implications
  • Different solution concepts may lead to different recommendations, and the appropriateness of a particular concept may depend on the specific context and assumptions of the game