[author]Wang Fuhua

[content]

Game analysis of collective litigation

*Written by Wang Fuhua

Professor, KoGuan school of Law, Shanghai Jiao Tong University

Abstract: The game analysis of collective litigation can provide the basis of microeconomics and social psychology for the improvement of the collective litigation system. The game in collective litigation is mainly carried out between class members, the class plaintiff and the defendant, the class and the court, as well as the class and the attorney, in each of them the players have specific litigation strategies and benefits. Generally speaking, the two sides of the game can achieve game equilibrium in procedural choice and substantial disposition decision, and find a compromised strategy to satisfy the interests of all players. In various kinds of game in collective litigation, restraining the “ free-rider” strategy, adopting the “ opt-out” rule, and incentivizing the high-value victims are conducive to maintaining and consolidating the group and realizing the value of collective litigation. In the game between the class plaintiff and the defendant, the choice of tactics of instituting and responding to an action reflects the necessity of coordinating the litigation expectations of parties and promoting litigation cooperation. The settlement game between class plaintiff and the defendant reveals the influence of the litigation cost and the accuracy of procedural operation on the settlement, and the need for the court to coordinate and supervise the settlement of the action. The game of case management in collective litigation shows that the court should be reasonably incentivized to accept class action, and the game of contingent fee needs to be reasonably regulated to prevent the abuse of collective litigation.

1.The structure and types of game analysis of collective litigation

In 1960, the American scholar Thomas C. Schelling put forward the idea of game theory as a framework for social science research, which opened the era of game theory as a tool for social science research. In the 1980s, American scholars Beberchuk and Kornhauser respectively published research results on the use of game theory to examine the civil procedure system, and game theory officially entered the field of civil procedure law as a research tool. During the same period, some U.S. courts began to advocate the application of game theory to class action trial practice, as "concepts of legal economics must be involved or relied upon when assessing the costs and benefits of class actions." In the past few decades, the application of game theory in theory and practice has provided another possibility for people to understand the laws of the group litigation system, reflect on the practice of litigation and improve legislation.

1.1 The structure of the group litigation game

Legal game analysis reveals the path of behavior interaction between people under certain rules. This kind of analysis can effectively reveal the motives and interrelationships of each litigant. Under the framework of game theory, a game consists of three elements: (1) a set of participants; (2) a set of policies for each participant; and (3) the participant's return for each possible strategy. If the litigation behavior is substituted into the above-mentioned game structure, it can be found that the procedural choice behavior of the litigation subject is usually a decision and a choice, which fully satisfies the requirements of the game elements and constitutes a typical game action. Moreover, compared with the usual litigation game, the group litigation process contains more game forms, more complex game strategies and more game nodes.

1.1.1 Game subjects in collective litigation

Participants in the game are decision-makers who choose actions to maximize their expected utility (returns). The main body of the group litigation game has the following characteristics: First, there are various types, including many victims (group plaintiffs), litigation representatives, lawyers, opposing parties, courts (judges) and other litigation subjects. The combination of participants in a specific game action is specific and may include a game between members of a group, a game between the plaintiff and the defendant, a game between the group and the representative, a game between the parties and the court, a game between the parties and the lawyer, and so on. As for the subjects in which the game action is carried out, it depends on the matters to be decided by both parties to the game, which can be the game of whether to sue or not, the game of whether to join the group, the game of electing representatives, the game of whether to accept the group dispute, and the game of whether to settle. In class litigation, various procedural choices, substantive dispositions, and even case management behaviors are all presented as game activities.

Second, the subject of the group litigation game has the ability to make rational decisions. Class litigation is a complex form of litigation, and the procedural operation and litigation outcome involve the interests of multiple parties, in which collective action decision-making is not available in ordinary litigation. On the one hand, there are a variety of game decisions in class litigation, and each subject in the game will pursue the most favorable procedural choice and substantive disposition result, and choose the most beneficial strategy, such as proposing a litigation settlement plan that is beneficial to its own party. If you can't find the best strategy, at least the worst strategy should be eliminated to achieve a rational choice. But on the other hand, whether the optimal strategy of the game agent can be achieved depends on the strategy of the other party. In this regard, the large-scale group litigation and the representative mechanism have increased the difficulty of the game subject to predict and judge the opponent's procedural choice and substantive disposition, and under complex conditions, the game subject must also adjust its litigation expectations according to the other party's strategy, substantive and procedural rules, probability of winning the lawsuit and procedural benefits (which can be obtained through court interpretation or legal services), make decisions and achieve game equilibrium.

Thirdly, there is a complex interactive relationship between the subjects of the group litigation game. In class litigation, whether it is collective action within the group, confrontation between the original defendant or game activities between other subjects, it all reflects a high degree of interaction. Since no one's choice is given, the outcome of each person's decision will be influenced by the decision of others, resulting in structural conflicts of interest in the relationship between the representative and other group members (the represented), the group lawyer and the parties, and the current rights holder and the future rights holder. At the same time, this shows that in the strategic game of procedural selection and substantive disposition, all parties must combine litigation confrontation and litigation cooperation well, and not only consider their own choices, but also need to consider the choices of others. In other words, the game analysis of the group litigation system reveals the reasons for the non-cooperation of the litigation subjects on the surface, but fundamentally explores ways to promote litigation cooperation.

1.1.2. Gambling strategies in collective litigation

The subject of a class action has the opportunity to choose a procedure and a substantive disposition, and each person should choose the best countermeasure for the other party's possible actions when choosing an action, which is the game strategy. The group litigation game strategy has the following characteristics: First, the group litigation game strategy corresponds to the litigation behavior, including two types of strategies: procedural selection and substantive disposition. For example, many victims choose to sue individually or join a group; whether the group has declared its withdrawal from the group after it has been formed; whether or not they agree to nominate a representative; whether the representative is proposed to be replaced after the nomination; agree or disagree with litigation mediation (settlement); Agree or oppose the appeal. Similarly, there are different game strategies in the game between the group plaintiff and the defendant, and between the group plaintiff and the court. In some games, the game subject will choose the "strict advantage strategy" or eliminate the "strict disadvantage strategy" at the same time to maximize the benefits. Other group litigation games will form a linear game strategy chain, in which the game subject makes the optimal choice in the continuous selection of procedures and substantive disposition nodes, and carries out repeated games.

Secondly, the "Nash equilibrium" of the game strategy can usually be found in the class action game. As with all gambling activities, the interdependence of various subjects in a class lawsuit determines that the game contains a conflict between individual rationality and collective rationality. Each subject not only makes choices and dispositions based on the maximization of its own interests, but also adjusts its choices according to the actions of the other party, until any party to the game unilaterally changes its own strategy under this strategy combination will not improve its own returns, so as to achieve a Nash equilibrium. Taking litigation settlement as an example, both the plaintiff and the defendant want to get greater litigation benefits, but the outcome of litigation settlement does not depend only on their own choice, but also on the choice of the other party. In the case that one party's settlement strategy remains unchanged, both parties may eventually agree to settle, or both parties may not agree, and these two "Nash equilibriums" are a stable combination of litigation settlement games. As for one party agreeing to the litigation settlement and the other party not, it is not a game equilibrium.

Thirdly, the path of group action game decision-making is relatively complex. Generally speaking, the strategy in a static game is relatively simple and intuitive. For example, in the decision of whether to join the group, the victim often adopts the "herd strategy", in which others sue themselves and others participate in the group, and others participate in the group themselves, which alone will form a game equilibrium. However, most class litigation games have more nodes, and it is more complicated to make optimal choices and find equilibrium paths. Taking the decision of stakeholders to join the group as an example, the game nodes at least include registering to join the group, withdrawing from the group, and requesting the court to apply the judgment in favor of the lawsuit.

1.1.3. Gambling gains in collective litigation

In game theory, the payoff refers to the reward that each participant receives for a given combination of strategies. Specific to the group litigation game, the benefits have the following characteristics: First, the group litigation game is aimed at the benefits of the game behavior. In various forms of games, such as "group-opposing party", "many victims-many victims", "group-representative", "group-lawyer", etc., the combination of the game strategy of the subject of the litigation and the game strategy of the opponent determines the litigation income. The benefits may be positive, such as reducing the cost of litigation through the choice of procedure and obtaining the desired amount of compensation through litigation settlement, or they may be negative, if the class action increases the cost of litigation, the cost of organizing the group and the litigation representation is too high, and the probability of intra-group conflict is too high, the gain of the class action may outweigh the loss.

Second, the benefits of group actions determine the game decision. The purpose of the game strategy is to maximize the benefits of litigation. Like other litigation games, the procedural choice and substantive disposition of the subject of the litigation are not only consistent with the litigation objectives pursued, but also based on the assumption of pursuing the maximization of interests. When disposing of their own civil rights and litigation rights, and when making litigation decisions, other litigants mainly consider their own personal interests and pursue better litigation results and procedural interests. As with the usual litigation games, the game benefits of class actions can be divided into two categories: substantive gains, that is, the difference between the benefits brought about by the outcome of winning the lawsuit or substantive disposition and the litigation costs, and procedural benefits, that is, the procedural benefits brought about by the choice of procedure or the litigation costs avoided.

Third, the benefits of the class action game are related to incentives. Litigation behavior in class actions needs to be coordinated, and the key to achieving coordination between the parties is how to form a common expectation, for which incentives need to be adopted. Taking the election of a representative litigator as an example, the United States has established the "lead plaintiff" rule: the plaintiff who first files a class action lawsuit can act as a group representative and become the representative of the litigant, thereby increasing his influence on the litigation and future litigation revenue. This provides the rules of the game for many victims to compete to become representatives (lead plaintiffs). On the contrary, the situation in China is that the parties lack the necessary institutional incentives to serve as representatives, the representative has little income, the workload and responsibility of the representative litigation is large, and many parties to group disputes are unwilling to pay for others, resulting in difficulties in collective action, which also reflects the need for legislation to adopt incentive measures to activate the group litigation system. Similar incentives are needed in the areas of case acceptance and representation.

In short, based on the correspondence with the above-mentioned three elements of game subject, game strategy and game benefit, group litigation has all the basic conditions of game elements and game analysis.

1.2 Types of collective litigation games

Games in class actions can be played simultaneously or sequentially, and it is often the case that the parties to the litigation use a mixture of the two game types.

1.2.1 Static game and dynamic game of group litigation

A considerable part of the class action game is static decision-making, that is, static game. Static game refers to the fact that both sides of the game act "simultaneously" when carrying out the punishment behavior. "Simultaneous" here is not only a concept of time, but also a concept of information, which specifically refers to whether one party knows what action the other party has taken when it acts. Many procedural choice behaviors in class actions are static games. Taking the game of competition as the representative of the group as an example, with sufficient incentives, the party with the information of the competitor can formulate the optimal strategy to become the litigation representative. This information includes the amount of victim support the adversary is known to the candidate on behalf of the candidate, how and what the adversary communicates with the many victims, the size and individual data of the victims, and the total amount of damage suffered by the group. If the opponent's strategy is very competitive, the candidate may choose to pay the cost of the representative at his own expense, collect evidence at his own expense, contact the person represented, etc., and propose a more competitive plan than the opponent, so as to make his strategy consistent with the litigation behavior.

A dynamic game refers to a game in which one party acts first, one party acts later, and the party that acts later knows the choice of the party that acts first. This kind of game, in which the two sides take turns to act, is also known as the sequential game. Corresponding to a class lawsuit, the game between the two parties on the jurisdiction of the case belongs to this kind of game, and the game nodes are: the plaintiff chooses the court when filing a lawsuit, the defendant raises a jurisdictional objection, and appeals the ruling on the jurisdictional objection. Taking pre-litigation jurisdiction as an example, both the operator and the consumer will choose the court that is favorable to them, rather than putting themselves in the other party's shoes. The standard contract provided by the operator usually stipulates the jurisdiction of the court in favor of the operator, because this can reduce the litigation cost of the operator. However, once many consumers choose such a court, they will have to pay higher litigation costs (such as transportation costs, etc.) after being involved in litigation, which is a strict disadvantage strategy that must be eliminated first when agreeing on a competent court. However, after the game, the above-mentioned parties' agreed jurisdiction strategy will form a balanced result - agreeing on an objective and neutral court jurisdiction. This is because, even if the court jurisdiction in favor of the operator is agreed upon at the time of conclusion of the contract, and thus the procedural interests of the consumer are harmed, it is only the first round of the game. Many consumers can raise jurisdictional objections, ask the court to exclude the jurisdiction clause in the agreement, or even continue to assert jurisdictional objections through appeals and retrials, and start a follow-up game.

The dynamic game of group litigation has the following characteristics: First, the actions of the litigation subjects in the dynamic game of group litigation have a sequence. Taking the game between the plaintiff and the defendant as an example, firstly, the group victim can choose to file a group lawsuit or not, and the defendant will take its own actions according to the decision of the plaintiff group, such as raising a causation defense or a party ineligible defense, settling with the plaintiff group or accepting mediation, etc. Both parties have a complete plan for procedural selection and substantive disposition, and adjust their choices according to the litigation actions of the first actor, and the two parties take turns to act, forming a tree-shaped game structure.

Second, the dynamic game of class litigation is usually a repetitive game. The group action has the characteristics of stages, and the related games are also continuous and repeated, and the games of each node constitute a stage game, and the results of the game in the previous stage do not change the structure of the game in the later stage. The game between the plaintiff and the defendant alone includes the parties' defence of qualifications, the defense of jurisdictional objections, conciliation or mediation in the first-instance trial procedure, and the game in the first-instance trial, second-instance trial, retrial and enforcement procedure in the whole process of litigation. These games are carried out sequentially and are embedded in larger games to form dynamic game models such as "centipede games". The intra-group game, the plaintiff and the defendant, and even the group and the court often do not stop at the first game, but in the initial game stage, the two parties choose to partially cooperate and carry out the game to multiple litigation stages to obtain higher returns.

Thirdly, the dynamic game of group litigation will be affected by the probability distribution of exogenous events. In addition to the participants' game strategies, the game process and outcome are also affected by some random factors that are not controlled by the parties, among which the influence from judicial policy is the most important. For example, in 2001, the Supreme People's Court promulgated the Notice on the Temporary Inadmissibility of Securities-related Civil Compensation Cases (Fa Ming Chuan [2001] No. 406), stipulating that securities civil compensation cases would not be accepted for the time being. In 2002, the People's Court promulgated the Notice on Issues Concerning the Acceptance of Civil Infringement Dispute Cases Arising from False Statements in the Securities Market (Fa Ming Chuan [2001] No. 43), which stipulates that "the people's courts should not accept civil compensation cases for false statements in the form of class actions". In 2005, the People's Court issued the Notice on Issues Concerning the Acceptance of Joint Litigation Cases by the People's Courts (Fa [2005] No. 270), which stipulates that "if the court accepting the case deems it inappropriate to accept it as a joint litigation, it may accept it separately".

1.2.2 Complete information game vs. incomplete information game in collective litigation

When carrying out litigation acts, the subject of a group lawsuit must make a judgment based on the information it knows, and information such as the parties' litigation rights and obligations, the litigation claims in the pleadings, the factual reasons, and the evidentiary materials and legal basis will all be exchanged among the parties and become the basis for their respective decision-making. The scope of information communication and disclosure of facts between the subjects of a class action is relatively wide, including legal procedural rules, facts of transaction or infringement, degree of personal injury, causal relationship, and the creditworthiness, business and property status of the opposing party. According to the parties' grasp of game information, group litigation games can be divided into complete information games and incomplete information games.

1.2.2.1 Complete information game. A situation where a group litigation entity fully grasps the characteristics and types of information of the other party when carrying out litigation acts is a complete information game. On the one hand, the right to choose procedures stipulated in the Civil Procedure Law belongs to public knowledge, which is usually open, transparent, and popular, and has the characteristics of complete information; on the other hand, in a group lawsuit, the parties have a better grasp of the whole picture of the facts of the dispute through continuous communication and the submission of facts and evidence, and have a clear judgment on what kind of procedural system to apply to resolve group disputes, which is also conducive to the realization of a complete information game. But in fact, the complete information of the class action game is not easy to achieve. Although the parties in a class lawsuit have the willingness to disclose information to each other, and the court will also compel the defendant such as a business operator or a hospital to submit evidence in a group environmental lawsuit, a drug injury lawsuit, a consumer lawsuit, or a medical lawsuit, in most cases, the group litigants cannot grasp all the information, and the defendant who has the evidence will not take the initiative to disclose the evidence, but on the contrary, he will try his best to conceal or destroy the evidence, which makes it difficult for most group action games to be carried out under the condition of complete information.

1.2.2.2 Incomplete information game. If the subject of a class lawsuit does not fully grasp the characteristics and types of information of the opponent, it is an incomplete information game. Incomplete information games in class lawsuits are common. For example, when a victim of a class action "opt-in" to a group, he or she usually makes a decision to join the group even if the other victim does not join the group, because he understands that the other victims really mean to join the lawsuit. Another example is that when many victims decide to file a class lawsuit, they often do not know the litigation strategy, property status and willingness of the opponent (perpetrator), and the perpetrator usually does not know the size of the number of many victims, the real intention of litigation, and so on. The above-mentioned incomplete information game is obviously related to the parties' motives for concealing evidence and not disclosing information. Motivated by self-interest, even those who are able to disclose their own information will remain silent or pretend to be unable to do so. Relevant research has shown that when some parties are indeed unable to disclose information, the game will favor the parties who are able to disclose the information but remain silent. Similar situations are more common in victim-internal games, "court-group" games, and "group-lawyer" games.

In short, the class litigation game can be divided into four combinations: static game under complete information, static game under incomplete information, dynamic game under complete information and dynamic game under incomplete information. In the following sections, we will analyze several combinations: the static game of class litigation is suitable for the situation where both parties make decisions at the same time, and can effectively analyze the strategic choice issues such as the victim's "free ride" and the lawyer's remuneration for winning the lawsuit. The dynamic game of class litigation is suitable for the situation where the game subjects take turns to act at different stages of litigation, and can be used for the analysis of case acceptance game and litigation settlement game.

2.The game in collective litigation

Whether a large number of victims in a group dispute can form a group in litigation and be recognized by the court as qualified parties is a prerequisite for the initiation and advancement of the group litigation procedure, and will ultimately determine the scope of res judicata of the group action judgment. To a large extent, a class action lawsuit is a process in which the plaintiff group is constantly shaped and clarified, and whether to join, how to join, and how to withdraw all constitute game actions.

2.1 The "free-riding" strategy of intra-group games

The many victims in a class action belong to a single decision-making unit, and they are an interactive and interdependent group that requires more coordination and cooperation. According to the law of collective action, the larger the victim group, the more asymmetrical the costs and benefits of individual litigation, the less likely it is that they will cooperate in litigation, and the more difficult it will be to form a group. From the perspective of institutional function, if all victims choose the route of group action to resolve disputes, such collective action can not only reduce the cost of litigation, but also deter defendants and achieve "wholesale justice". However, collective action often presents difficulties, and the disposition of substantive rights and procedural choices often arise in the "prisoner's dilemma" or even the "multi-prisoner dilemma" – each party in the group has an incentive to be lazy or retain contributions, expecting to enjoy the benefits of others' contributions. Especially in the initiation of class actions, the parties are more inclined to adopt the "free ride" strategy, that is, to determine their own litigation strategy by observing the efforts and returns of other members of the group. In the context of this "dispersion of responsibility" or "responsibility entrustment", victims want to act as bystanders, do not organize or participate in the litigation group, and determine their own strategies after observing the efforts and returns of other parties. Obviously, the "free-rider" strategy is a completely information-static game.

From the perspective of public law, since the state has established a group litigation system, it must devote itself to its application, giving full play to its advantages in improving the efficiency of litigation and achieving "wholesale justice", rather than allowing social slackness such as "free riding". The measures to limit the "free-rider" strategy are varied: first, to reduce the gambling benefits of the strategy. If the victim does not participate in the group and has nothing to lose and does not have to pay the "future costs", he is likely not to join the group and waits for a "free ride" or sues separately on his own. If the "future cost" of the victim not participating in the group is high, the cooperative strategy will yield benefits. Therefore, if we want to change the "free-rider" situation, it is necessary to raise the cost of non-cooperation of the parties in legislation. Second, incentives and penalties reduce the willingness to "free ride". In terms of incentives, the corresponding mechanism for compensation for expenses should be clarified, and the court should support the litigation representative's claim for compensation for reasonable litigation costs. In terms of sanctions, consideration could be given to expanding the validity of judgments, so that victims who deliberately do not participate in litigation cannot sit back and enjoy the outcome of group actions, so as to encourage them to participate in collective action and achieve a one-time settlement of group disputes.

2.2 "Opt-in" and "opt-out" game rules

Once a group action case is accepted by the court, it will directly affect the litigation status of many victims: on the one hand, each victim has to decide whether to participate in the lawsuit; On the other hand, as groups are formed, the ability of individual victims to control large lawsuits will be weakened. Moreover, after the formation of the group, each victim is faced with the choice of control and counter-control strategies - either to remain in the group and continue the group lawsuit, or to leave the group and seek redress on their own. Therefore, group members should choose between two strategies, "joining" and "opting out" to maximize their litigation benefits. This is a static game under completely informational conditions.

"Opt-in" means that a victim intending to enter a particular class action must have a clear and affirmative access to the class action proceeding within the prescribed manner within the time prescribed by law. Only when the victim makes such an expression of intent in a timely manner will he or she finally be bound by the class action judgment. China's Civil Procedure Law adopts this model, which has the advantage of being in line with the principle of litigation disposition and is simple and easy to implement, first issuing a notice of right registration, and then reviewing the qualifications of the registered entity, so as to facilitate the right holder to exercise the right to sue. "Opt-out" refers to the assumption that a particular individual is a member of a large group of victims, and that he or she remains a member of the group unless he or she expressly withdraws from the proceedings and excludes himself from the effect of the group and the judgment. The "opt-out" is a mechanism that facilitates the protection of the right of action of the many victims as a whole, and it is never worse off as a member of a group than a member of a non-group, and will eventually be compensated to a greater or lesser extent, so it is a more advantageous position for victims.

The "opt-in" rule gives many victims the option of a strategy that allows them to sue individually and without being bound by a class action judgment if they wish – unless they voluntarily agree to participate in the action, the class action decision has no effect on them. However, this rule can also lead to "negative returns" in the game: first, victims often lack motivation to litigate, and in the absence of incentives, they will use "free riding" as the optimal strategy and do not actively participate in litigation. This is in line with the logic of the game of life – waiting for someone else to make a choice of strategy until the strategy proves to be successful or unsuccessful. Second, the "opt-in" rule will weaken the attractiveness of the group litigation system and is not conducive to the one-time resolution of group disputes. In contrast, the incentive for parties to sue and settle under the "opt-out" rule is enhanced, because if they do not claim compensation through a class action, they will lose their right to sue for the same infringement. Third, the "opt-in" rule is also not in favor of the defendant, because if the victim does not join the group, he will not be bound by the judgment, and he can file a lawsuit alone. In this case, the defendant has to respond to the lawsuit one by one, and the fatigue of responding to the lawsuit will lead to an increase in the marginal cost of dispute resolution and increase the litigation burden of the defendant. But equally, the drawbacks of the "opt-out" rule are obvious, and it has been criticized by law and economics for its inefficiency, irrationality, and low returns. Moreover, if all parties in the group opted out, the outcome would be worse than if they all opted in. It can be seen that the two rules of "opt-in" and "opt-out" have their own advantages and disadvantages, and the adoption of different rules will inevitably have different impacts on the function of the group action system. As to whether the two are beneficial to the group parties, it is necessary to distinguish the specific circumstances: for victims who claim a higher amount of compensation, "opt-in" should be an advantageous strategy, since they can obtain better procedural safeguards by filing alone. However, if the amount of compensation is low, the litigation costs that each member needs to pay may exceed the amount of compensation, and "opting out" has a positive effect on the protection of the rights of these people.

On the one hand, when the scope of interested parties is not clear, some of them are allowed to sue first, initiating group action procedures, so that rights holders who have been confirmed by the parties can obtain timely relief; After accepting a case, if the court finds that the party to the lawsuit has not yet been determined, it may issue a public announcement explaining the circumstances of the case and the litigation claims, and notify the rights holder to register with the court within a certain period of time, so as to provide the right holder who has not yet entered the litigation procedure with an opportunity to participate in the litigation. The problem, however, is that the "opt-in" rule imposes a heavy registration and examination burden on the courts, as well as a greater risk, so the courts lack motivation. As an attempt, China has begun to shift to the "opt-out" rule in the field of securities class litigation, which not only takes into account the protection of investors' substantive rights and procedural safeguards, but also alleviates the burden of court case management. Similarly, in antitrust class actions, the opt-out system is also believed to maximize the aggregation of victims of monopolistic behavior, form a scale advantage in litigation, and strongly incentivize litigation.

The Securities Law amended in 2019 and the Provisions of the Supreme People's Court on Several Issues Concerning Representative Litigation in Securities Disputes (Fa Shi [2020] No. 5, hereinafter referred to as the "Provisions on Securities Representative Litigation") have established a dual mechanism for participating in class litigation: (1) the "implied participation + express withdrawal" model, which is applicable to representative litigation in securities disputes initiated or participated by investor protection institutions, and (2) the "express participation + express withdrawal" model, which is applicable to ordinary representative securities litigation. In the field of securities class litigation, "opt-out" replaces "opt-in" with a better game strategy for the investor group: first, in the case of investors with a more specific range of investors represented by investor protection agencies, this rule is more conducive to the formation of investor litigation groups, which is in line with the law of public choice in the securities field, and the participation rate of class litigation can be greatly increased. Second, the "opt-out" mechanism is represented by the group, which can provide competent and loyal services to the many victims represented, remove the economic barriers to litigation, prevent the group attorneys from making excessive profits, and facilitate the establishment of links between the group and its members. In addition, the mechanism still allows the parties to decide the scope of the group independently, which embodies the concept of party autonomy and the principle of self-attribution, and helps to protect the litigation rights of the parties.

2.3 The game between high-stakes and low-stakes victims

The dynamic game within the group is mainly between the "high-value victims" and the "low-value victims", which is a completely information-dynamic game. A high-value victim is a victim whose expectation of litigation benefits is higher than the expectation of litigation costs, and whose claims are "strong claims" because the damage suffered by such victims is serious and the amount of compensation is high. Low-value victims refer to victims whose expectations of litigation benefits may be less than the expected litigation costs, and whose claims may outweigh the losses in their separate litigation rights, and their litigation claims are "weak litigation claims".

The game between high-value and low-value victims is usually a game of non-cooperation. This is because, first of all, the best option for each victim is to "free ride", not to participate in the collective action themselves, but to sit back and enjoy the results of the litigation of others, which is a typical manifestation of non-cooperation. Second, in the case of a class action lawsuit, it seems that the best strategy for victims should be to register to participate in the lawsuit in order to reduce their own litigation costs or litigation risks and maximize the benefits of litigation. The problem, however, is that without the deterrence of class actions, victims are less likely to receive reparations, so they have to base their choices on the basis that other victims also litigate, and cooperate with other victims to litigate together and face the defendant. In contrast, the most injured high-value victims are the most willing to sue, but the lower victims have the right to decide whether to join the group and profit from it, which can result in losses for the high-value victims (litigation for the benefit of other victims without additional returns). In other words, while the size of the group increases after the participation of low-value victims, high-value victims may fear that they will pay higher litigation costs and that the benefits will be less than expected. In contrast, the benefits of litigation for low-value victims will be greater than expected. In this case, solo litigation is undoubtedly the best strategy for high-value victims, and he does not hesitate to withdraw from the group, even if the cost of individual litigation is high.

Figure.1 Extended game model of high-value victim and low-value victim

Figure 1 shows that under the opt-in rule, the dynamic game between high-value victims and low-value victims can be inversely summarized in two aspects: first, assuming that the cost of a group action is -10, the cost of litigation for the high-value victim who initiates a class action first is -10, and when the low-value victim chooses to join the lawsuit, the litigation cost of the high-value victim is shared and reduced to -5. At this time, the cost of litigation for low-value victims is also shared, down to -5. If the low-value victim does not choose to join the class action lawsuit filed by the high-value victim and files the lawsuit on his own, the litigation cost is still -5 because the litigation cost is lower due to the small amount of the subject matter, the litigation cost is not required to be represented by a lawyer, and the class action winning judgment can be applied. Second, assuming that the high-value victim does not file a class action, the low-value victim is faced with two options: file a class action lawsuit and bear the litigation cost of -10 by himself, and at the same time give the high-value victim a chance to "free ride", and the latter's litigation cost drops to -5, or file a separate lawsuit with the high-value victim, which is a lose-lose situation, and the cost of rights relief for both is -10. This inverse inductive game model shows that, on the one hand, although the high-value victims do not know the litigation strategy of the low-value victims, they must understand their position in the game, and filing a group lawsuit and calling on the low-value victims to join it is the optimal response and the optimal strategy for the low-value victims. On the other hand, it is not the optimal strategy for low-value victims to file a class action lawsuit, although they have the same belief as high-value victims on the issue of group rights protection, but there is a higher litigation cost to file a class action lawsuit, and the optimal litigation strategy is to join a class action lawsuit filed by a high-value victim.

From the above game model, we can also observe the characteristics of group dynamic games: the parties have to act sequentially to indicate whether to join the group, or decide whether to leave the group after joining the group. In the case where members of the high-value victim (group 1) and the low-value victim (group 2) are brought together in the same class lawsuit, there will be a dynamic game between the two groups. Since the party who suffers a low amount of damage due to the same infringement (weak claim) shares the litigation proceeds of the party with a high amount of victim (strong claim), the cost of a group lawsuit, especially the cost of filing a group lawsuit first, is mainly borne by the high-value victim, and the litigation income usually does not increase, and if he is unable to lead the action of the group, he may withdraw from the group. This "adverse selection" strategy will have a domino effect, and some high-value victims will drop out of the group, and their proportion in the group will decrease, which will trigger the withdrawal of more members and further weaken the ability of the group to play.

But like other games, repeated games between group members can also find game equilibrium. This is because the victims in group disputes are "rational economic persons" who will take into account both immediate and long-term interests, so it is possible to find an equilibrium point and maintain cooperation between group members through repeated games. First, "adverse selection" is made in the context of information asymmetry, when one part of the victim is unable to determine the damage of the other part of the group. Due to information asymmetry, the aggregation of high-value victims and low-value victims into a single procedure weakens the motivation of the former. Therefore, Figure 1 connects the two nodes of the low-income victim (group 2) with dotted lines to emphasize the information that the low-income victim should know when he or she decides to join the group. Second, high-value victims are the backbone of class lawsuits, and if they leave the group, then class lawsuits will be less meaningful. Therefore, in order to mobilize the enthusiasm of high-value victims, it is necessary to change their litigation situation: First, it is necessary for the court to classify the victims' claims according to the strength of their claims, and each group to appoint representatives and entrust lawyers respectively, so as to protect the interests of different types of parties. The second is to adopt reasonable incentive measures, and fully consider the particularity of the litigation claims of high-value victims in the litigation process and litigation results, so as to increase their reasonable gaming benefits. For example, it may be considered to give priority to compensating them for the costs they have already incurred when the benefits of winning the lawsuit are distributed, so as to fully mobilize their enthusiasm for litigation, and even consider providing some cost incentives at the beginning of the class action to avoid the litigation dilemma caused by the withdrawal of high-value victims from the group.

In conclusion, it is difficult to separate the connection between the litigation game and other aspects of the interaction between the parties in the process of group formation, because the situation of many members after a dispute has arisen can be a "prisoner's dilemma". As for whether to take collective action, some games require group members to make decisions independently at the same time, while others require group members to make decisions sequentially. However, in any case, the formation mechanism of the group will determine the operation effect of the system, and whether the group can be successfully formed depends on whether there are fair and reasonable incentive measures. As a legislative optimization measure, on the one hand, it is necessary to adopt reasonable incentive measures to change the unreasonable phenomenon of "free riding", and on the other hand, it is necessary to mobilize the enthusiasm of high-value victims in litigation and give play to their backbone role in resolving group disputes.

3.A game between a group plaintiff and a defendant

In the class litigation game action, the game between the plaintiff and the defendant is the most intense, and the game nodes are also the most, covering two types: static game and dynamic game. The game between the two is the most typical between the two nodes of litigation initiation and litigation settlement.

3.1 Litigation initiation strategies for group plaintiffs and defendants

The static game between the plaintiff and the defendant in a class action lawsuit is manifested in the strategic choice of suing and responding to the lawsuit. Both parties have their own strategies of strict superiority: one party to the group pursues a winning outcome and tries to recover the defendant's ill-gotten gains, but toughness and confrontation often make the group party unable to get too many benefits from winning the case, and has to accept the sub-optimal outcome. The defendant also pursued the goal of litigation without paying a penny, but had to lower its psychological expectations and adjust its strategy according to the other party's strategy, so as to finally achieve game equilibrium.

3.1.1. Litigation strategies for group plaintiffs

In the game between the plaintiff and the defendant in a class action, the plaintiff as the victim has an advantage and can obtain more procedural benefits. First of all, through a litigation procedure to hear and adjudicate common factual and legal issues, the group plaintiff can obtain the benefit of reducing the cost of litigation, and with the increase of the size of the group and the number of common issues, the cost advantage will be further amplified. Second, class actions can bring economies of scale to class members, offset the defendant's advantages, prompt the defendant to rationally consider the claims of all victims, and increase the chances of winning the lawsuit by increasing litigation investment. Thirdly, the class action mechanism increases the benefits of the group plaintiffs, and in addition to expecting compensation from the defendant, they can also use those "negative value lawsuits" (i.e., lawsuits that lack litigation value) as a strategy to threaten the defendant, and resort to the group litigation procedure for disputes with little hope of winning the lawsuit or the amount of compensation is not high, so as to achieve the effect of "small to broad".

The advantage of litigation for group plaintiffs is the marginal diminishing cost of litigation. Suppose a mass tort causes 100 victims and suffers a loss of 80 yuan each, if the victims sue individually, they will incur 100 yuan in litigation costs, while if all the victims sue and merge into a group action, the litigation cost paid per person is only 1 yuan, which is much lower than the cost of individual litigation by the victims, and everyone can get a positive benefit. This scenario can be reduced to a "multi-prisoner dilemma" game model: each victim has a strict advantage strategy – filing a lawsuit to defend their rights. If the other victims do not file a lawsuit, the victim can file a lawsuit to reduce his or her losses from $80 to $20 ($80 for the loss, but only $100 for the costs of the lawsuit), and from $100 to $50 if one more victim also participates in the lawsuit ($50 for each of the two victims and $30 for the win). Therefore, in any case, filing a class action lawsuit will be the best strategy.

Table 1 Matrix of income from class action litigation

Obviously, class actions provide a way for many victims to achieve "wholesale justice", and the game between the plaintiff and the defendant is consistent with the legislative purpose. Conversely, in the absence of a class action mechanism, the defendant is highly likely to avoid liability when the damage suffered by each victim is small but there are many victims, and he is well aware that the victims must take into account the costs of litigation and that it is costly to seek judicial remedies by suing alone, so the victim's individual claims are often ignored. In other words, a group lawsuit can transfer the small amount of compensation requested by an individual to a large lawsuit, balance the litigation capacity between the parties, promote the development of the law, fill the legal loopholes, and turn a non-threatening individual lawsuit into a large-scale tort lawsuit, inject deterrent genes into the group lawsuit, make it have the function of private law enforcement, and save a lot of social costs in resolving large-scale disputes and regulating the market.

3.1.2. The defendant's response strategy

Once the class action lawsuit is accepted, the risk of the defendant losing the lawsuit increases, and the liability for compensation may become a reality. In the case of a large victim group, the best strategy for the perpetrator is to accept the class lawsuit and compensate the victim for the loss in order to avoid a greater liability to more victims, which is their strict advantage strategy. Otherwise, you can only sit and wait for many parties to sue one by one, become defendants repeatedly, and pay the cost of litigation repeatedly. In other words, the defendant can liquidate the liability of the defendant in a centralized and holistic manner through a class action, and the outcome of the litigation is tantamount to bankruptcy, which is a positive effect for it, or at least reduces the cost of litigation. On the other hand, the large number of victims filing class lawsuits may also be motivated by abusive motives. After all, the reduced cost of litigation can encourage people to abuse the system, even if there is little hope of success, to sue hastily and start litigation. In the case of a sloppy lawsuit by the plaintiff, the defendant's strict advantage strategy should be to find ways to increase the cost of litigation for the other party and reduce their willingness to litigate. For example, consistently refusing to settle a lawsuit increases the cost and risk aversion of the plaintiff group, reduces their expected benefits, etc. At the same time, it is also possible to achieve a game equilibrium by dividing and disintegrating the plaintiff group as much as possible, and recognizing the claims of some groups while rejecting the claims of others.

In short, the relationship between the group plaintiff and the defendant constitutes the most important game relationship in class litigation, and the game between the two at the initiation point of litigation prompts us to pay attention to the dual value of class litigation, which is not only to provide effective remedies for many victims, but also to deter potential large-scale infringements, and the social value of group litigation demonstrated by the latter is undoubtedly very important.

3.2 Litigation settlement games

As in ordinary litigation, the game between the plaintiff and the defendant in a class action is a dynamic game of incomplete information. This is because the gambling entity does not know much about the situation of the other party - in order to obtain greater benefits of winning the lawsuit, the plaintiff group may conceal the facts or exaggerate the losses, and they often use the facts that the group has formed as a game strategy to pressure the defendant, and expand the scope of litigation claims, complicating the case. The defendants had similar self-interested motives and would not disclose the relevant facts unless compelled to do so.

Under the condition of incomplete information, only after repeated games can the two sides reach game equilibrium. In terms of defendant litigation strategy, after the formation of a class action, the defendant may turn to a position of procedural weakness, forcing it to invest more litigation costs in order to overcome a large number of opponents. After a group has made a decision and formed a group, the defendant will consider the impact of his actions on the group, and he knows better than the plaintiff whether he will be judged to bear civil liability and whether he will face an "all or nothing" outcome. In addition, the plaintiff and the defendant have independent access to some of the information in each other's possession, and to some extent the parties tend to have similar beliefs, so that reconciliation becomes more likely. Especially for the defendant, litigation settlement is objectively conducive to achieving a one-time and complete settlement of the dispute, and can effectively protect its privacy, resolve the dispute as soon as possible and confidentially, hide the infringing information to maintain production and operation, and prevent its reputation from being continuously damaged in the lengthy litigation process. In particular, the settlement agreement has a one-time dispute resolution function, which plays a bankruptcy-like role and saves the defendant from ongoing litigation, so his strict advantage strategy is to agree to the litigation settlement. The above-mentioned litigation motives determine the following characteristics of the dynamic game of group action settlement:

First, the settlement game of class litigation reflects the balance of the game between the plaintiff and the defendant. Litigation settlement as a litigation strategy is a strictly advantageous strategy for both group plaintiffs and defendants. When the plaintiff forms a group, it gets the "bargaining chip" to negotiate or settle with the defendant, and the group representative and lawyer have the power to deter the defendant, and the settlement can be used as an "opportunistic behavior" strategy by the group plaintiff, and the settlement has significant litigation benefits. If the plaintiff's expected value of success (minus the costs of litigation) is less than the defendant's expected damages, then a mutually beneficial settlement is possible, and any settlement amount in between will be better for both parties than the trial. Therefore, when settling a lawsuit, the defendant is often the first to put forward the conditions, and the game strategy of the group plaintiffs is either to agree or disagree, and they usually choose not to agree. The defendant may then compromise further and propose better settlement terms, and after several rounds of repeated negotiations, the litigation settlement game can be declared closed until the other party agrees or completely rejects.

Second, litigation settlement strategies generate corresponding litigation benefits. For one side of the group, they may agree to settle for reasons of litigation efficiency and litigation risk. If a settlement is not agreed, the cost of litigation will increase later on, and many group members may receive less compensation. In addition, other variables (exogenous events) also come into play in litigation settlements, for example, the attorney may take advantage of the opportunity to profit at the expense of many group members who cannot effectively supervise the group members. For the defendants, they will be worried that more victims will join the lawsuit, so they want to settle the lawsuit as soon as possible through settlement, and the division and disintegration of the group has become their strict advantage strategy. But as the group size grew, the defendant eventually had to satisfy the claims of all the victims – not only for the many victims sued, but also for all potential victims (future claimants), and he had to strike another equilibrium with the group plaintiffs, which was to compensate the victims as a whole so that the dispute could be resolved once and for all.

Thirdly, the cost of litigation and the quality of the procedure determine the outcome of the settlement game of group litigation. The practice of securities class litigation in China shows that if the cost of litigation exceeds the compensation it can obtain, the plaintiff would rather give up the lawsuit; If the defendant listed company knew that the plaintiff's litigation costs outweighed the benefits, it would not settle with the plaintiff. In addition, the more accurately the proceedings are conducted, the more likely the parties are to agree on the evaluation of the outcome of the trial and the more likely they are to reach a settlement. The enlightenment of this is that through the reform of the litigation cost mechanism, the litigation costs of the parties can reflect the actual litigation costs, which can promote the settlement of group disputes through litigation settlement.

The above-mentioned process of group action settlement game is only a theoretical derivation, and the actual game situation is certainly more complex, and other factors should be taken into account. First of all, the gaming status and ability of group plaintiffs and defendants are complex. On the one hand, individual group plaintiffs lack procedural control and often play an insignificant role in litigation settlement, and are unlikely to play a role in settlement like the representative they elect, and even do not understand the litigation status of their own representation and the significance of the group action settlement agreement to themselves, so litigation settlement is easily manipulated by representatives or lawyers. On the other hand, group plaintiffs may also benefit from vulnerable defendants through abusive litigation and inefficient or predatory settlement agreements. Second, if the litigation representative and the lawyer occupy a favorable negotiating position in the litigation settlement negotiation with the defendant, they can put forward favorable negotiation conditions and obtain more litigation benefits, and this kind of settlement is called "sweetheart settlement", which is often the defendant's bribery of the litigation representative. There is also a tendency for group lawyers to pursue excessive profits, especially through litigation and settlement. To correct these gaming imbalances, courts must be legislatively empowered to monitor and review class action settlement agreements.

The judicial supervision of litigation settlement by judges can also be regarded as a "supervision game" between judges (agents) and courts (clients). The judge's strategy is to supervise (work) or not to supervise (lazy), but supervision and review can ensure the fairness, reasonableness and adequacy of the settlement of the group action, protect the legitimate rights and interests of the parties, especially one of the parties to the group, and achieve fairness and justice. Conversely, if the judge does not supervise and review the settlement agreement, although it can save judicial resources and maximize the total benefits, it may allow some parties, lawyers or litigation representatives to harm the interests of the state, society and the rights and interests of other parties. Therefore, the court should supervise (supervise) whether the judge has exercised the power of judicial review of the class action settlement agreement, and prevent the judge from neglecting the review and causing the settlement agreement to be unlawful.

While it is necessary for judges to supervise and review settlement agreements, it is difficult to implement them in practice. It is an incomplete information game, and supervision and censorship will be constrained by information. The judge supervising the review did not have sufficient information to judge whether the settlement agreement was reasonable and whether it harmed the interests of the victims who did not participate in the proceedings, because they lacked the appropriate sources of information and did not conduct the substantive hearing. Nevertheless, the judicial supervision and review mechanism for group action settlements should be maintained, which is a necessary condition for ensuring the voluntariness and legitimacy of litigation settlements and giving full play to the deterrent function of the system.

4.Case management game for class lawsuits

Group litigation is a dispute resolution method with high litigation risk, which requires precise judicial supervision and active judicial management. Broadly speaking, the judicial supervision and case management of group litigation cover almost all aspects of group formation, selection of litigation representatives, and litigation settlement.

4.1 "Prosecution-acceptance" game

Almost all group litigation legislation emphasizes the role of case management, and the court filters the cases at the initial stage of litigation, which gives rise to the game between the court and the group in the decision to prosecute and accept the case. In this game under the condition of complete information, the main body of the game is the group parties and the court, and the game strategy is the choice of the form of litigation - group litigation or individual litigation. Group parties tend to be "collective proceedings", forming groups after filing a lawsuit, which can obtain greater litigation benefits, reduce litigation costs and maintain deterrence against defendants. However, for the courts, if they do not have a suitable judicial environment and sufficient judicial resources, the optimal game strategy is to split the group action cases and accept and try them as separate litigations. However, this kind of separation must take into account the rigid constraints of the Civil Procedure Law, and cannot illegally restrict or deprive the parties of their right to sue. Obviously, the game strategy of the above two parties depends on their game profits, and when the benefits of the class action exceed the benefits of the individual lawsuit, they will choose the group action as a strict advantage strategy. Instead, they will choose to litigate separately as a strictly advantageous strategy.

On the whole, the "prosecution-acceptance" game has significant dynamic game characteristics. Not only is there a sequence of actions between the group parties and the court, but the parties are aware of the information at the time of decision-making. After a group litigant files a lawsuit, the court should make a strategy of whether to accept the case according to the group case or split the group case; If a large number of victims and the court are unable to reach an equilibrium in the game, the next round of the game will be triggered until the two sides reach equilibrium. In this process, the court first examines whether the lawsuit meets the conditions for a group action and decides whether it will be tried in the form of a group action or a separate lawsuit (split case). Subsequently, the group party has to determine its own strategy of accepting the court's decision to proceed with the lawsuit; Or raise objections through procedural objections, petitions and other complaint channels, creating pressure on the court and even damaging its reputation.

The following assumptions are that in cases where the court does not have the capacity to resolve the class action, one party of the group files a complaint as a strategy to "retaliate" against the court for splitting the case. The parties have four options for the parties: (1) if the court accepts the group dispute case, the group parties will gladly accept it; If the court splits the group dispute case, then appeal or complain. (2) if the court accepts a class action, the group parties appeal or complain; If the court splits the group case, it will gladly accept it. (3) Regardless of whether the court accepts the case in the form of a class action, the group party will gladly accept it. (4) Regardless of whether the court accepts the case in the form of a class action, one of the group parties appeals or complains. The above policy combinations are represented by the model shown in Table 2.

Table 2 Dynamic game model of "prosecution-acceptance" in group litigation

The respective benefits of the group parties and the court are as follows: if the group party accepts the group case without the ability to resolve the group action lawsuit, and the group party accepts the decision, the court's benefit is -2 (because accepting the group lawsuit may affect social stability and the trial is time-consuming and laborious), and the benefit of the group party is 2 (after the group lawsuit is accepted, it will reduce the litigation cost of the group party and deter the defendant); if the court has to accept the group case, but the group party still complains, the court's benefit is -5, and the group party's complaint may also be dismissed, so the benefit is also -5If the court splits the case and hears it separately in accordance with the law, and the group party complains, the court's benefit is -5, and the group party is also -5; if the court splits the case and accepts the case separately, and the group party does not complain, the court's benefit is 2 and the group party is -2.

It is not difficult to find that some of the strategies in the above-mentioned dynamic games are obviously threatening, and are not the real meaning of the game subject, so they are not the real game equilibrium. For example, at the point where a group case is accepted, the threat of the group parties is usually that if the court does not accept the group case, it will resolve it through complaints such as petitions or creating news and public opinion, creating more pressure on the court. However, the threat of the court is usually that group litigation is easy to cause mass incidents, which is not conducive to dispute resolution, and the plaintiff's insistence on suing in a group lawsuit is to undermine social stability. But in reality, none of these threats can be trusted and are not a reasonable game strategy. Because the parties' respective ex-ante and ex-post optimal choices are inconsistent—the parties' expectations of resolving disputes and obtaining compensation are inconsistent with the behavior of complaints, and using complaints as a threat is not the optimal strategy. Similarly, the court's behavior of splitting the case is inconsistent with the optimal strategy of one-time resolution of disputes, and group litigation is more in line with the judicial goals of litigation economy and litigation efficiency. Therefore, when the conditions for group litigation are met, the strategy of rejecting a group lawsuit on the grounds of maintaining social stability is also an "unbelievable threat" or an unreasonable strategy, which is not a true Nash equilibrium and should be excluded in the game analysis.

From a rational point of view, there are two kinds of Nash equilibrium that should be excluded in the "prosecution-acceptance" game: (1) For group cases that should not be accepted, if the court accepts them as a group lawsuit, the group parties choose to accept them. The result of this equilibrium may be that the court is worried about reputational damage and no longer considers its own judicial capacity limitations and social stability factors, and the judicial risk increases, so the benefit is -2, while the group party reduces the cost of litigation and deters the defendant, and the benefit is 2. But in fact, this Nash equilibrium implies the premise that the court is forced to accept the group action case, which is tantamount to the court's subservience to the group parties. Even if the court does not accept the case as a class lawsuit, as long as the decision to split the case is made by the court within the scope of its discretion and case management power, the complaint of the group party will be dismissed, and the complaint and rights protection are time-consuming and laborious and cannot really solve the problem of accepting the class lawsuit, and the benefit is -5. On the other hand, if the court knows that the group of parties is rational, it should not be afraid of its threat. (2) In cases that should be accepted as a group lawsuit, the court splits and accepts them separately according to individual lawsuits, and the parties choose to accept them. In this case, the court's gain is 2 because the risk of a class action is avoided, which is a positive benefit, while the benefit for the group parties is -2 because they are not accepted as a class action, contrary to their initial expectations, and the group parties are under "unbelievable threat".

In contrast, for a group case that should not be accepted, the court decides not to accept the group action case but the group parties accept the result, which is a Nash equilibrium. The benefits of both parties are: the court splits the trial of group disputes according to its own circumstances in accordance with the law, which not only avoids judicial risks but also avoids complex procedures, and its benefits are 1; The litigation mentality of the group parties is correct, and even if the court splits the class action case, they can accept it and believe that they can get the same litigation result as the class action. Obviously, this Nash equilibrium is reasonable and does not contain "unbelievable threats". In other words, the litigation strategy of the court and the parties to the dispute reflects dynamic consistency. Therefore, the court can achieve a state of equilibrium in the game by accepting the case in an alternative way, such as "separate case filing + joint trial", which is the best strategic choice although this is a suboptimal outcome for both parties.

The game between the group of parties and the court on "prosecution-acceptance" only shows the litigation psychology of both parties and the actual state of procedural operation, and the result cannot replace the value judgment of the law. However, the game process and results are enough to suggest that we must improve the relevant game benefits when optimizing class action legislation, especially considering the game benefits of the courts. Because class actions consume more judicial resources, it is quite difficult to accept and try such cases. Dividing and disintegrating group lawsuits can not only reduce the difficulty of trial, but also weaken the gaming ability of many plaintiffs and prevent some difficult situations due to group gatherings. Relevant judicial practice has shown a "herd effect" - if one court splits cases from the perspective of self-interest and excludes group litigation, then other courts are bound to be even more exclusive, and ultimately make the group litigation system lose its value. To change the above situation, it is necessary to set reasonable incentives, so that the benefits of splitting cases are smaller than the benefits of group lawsuits, and to maximize the institutional value of group lawsuits: on the one hand, courts should attach great importance to the goals of safeguarding judicial authority, public interests and social stability, and accept group lawsuits to protect the interests of the group; On the other hand, it is necessary to optimize the judicial performance evaluation mechanism, adopt incentive measures, and increase the benefits of the courts' acceptance and trial of group cases. Considering that the political laws on which the operation of state organs depend are different from the economic laws on which the actions of legal persons depend, incentives in the form of "powers-duties" can be adopted to motivate trial organizations and judges who hear group action cases through remuneration, or to set up collegial panels that specialize in hearing group action cases, and to strengthen the specialization of group dispute resolution.

In short, the highly complex practice of group litigation inevitably means difficult case management, and high-quality case management can make the group litigation system effective. The analysis of the "prosecution-acceptance" game confirms the policy needs of the group action system, where both parties and the court need to receive reasonable incentives, and both should also encourage each other, which is a necessary condition for initiating a group action.

4.2 The game in the system of winning the lawsuit and getting paid

The rapid development of the class action system in the international arena is largely due to the blessing of the cost-sharing rule of "remuneration for winning the case". The prevailing remuneration rule is a mechanism that shifts the economic risk of losing a case from the group and the plaintiff's representative to the group of lawyers, who are only liable to pay the lawyer's fees if they win the case. At the same time, this rule also provides lawyers with the necessary financial incentives to put class action cases first and do their best. Under this rule, there are a variety of strategic options between the group parties and the lawyer, including: (1) a fixed standard representation fee: if the group wins, the lawyer receives the representation fee according to the standard agreed by both parties; (2) Determine the agency fee according to the agreed base: if the group wins the lawsuit, the agency fee will be increased accordingly according to the risk coefficient as a reward; (3) Determine the agency fee according to the ratio according to the winning result: if one of the group wins the lawsuit, the agency fee will be determined according to the percentage of the total compensation received by the group; (4) Determine the floating agency fee standard according to different litigation stages: Determine different agency fee ratios according to the litigation settlement, pre-trial stage and trial stage in the process of group dispute resolution.

The game between the group parties and the lawyer around the remuneration system for winning the lawsuit is an incomplete information static game. On the one hand, when a lawyer decides to represent a class action, he or she usually does not fully understand the facts of the damage, the process of forming the dispute, the total litigation costs, etc., and even if the lawyer has all the information, it is difficult for the client to compel them to disclose it. At the same time, based on the assumption of maximizing lawyers' interests, it is impossible for lawyers to actively and completely disclose all the information they have to many principals and make the information a "public resource". On the other hand, the group side also lacks the necessary information in the "group-lawyer" game. Group members are often strangers, lack the necessary communication with each other, and it is difficult to obtain consensus, and their responsibilities in supervising the representation of lawyers are decentralized, and the intensity of supervision is low, so it is naturally difficult to obtain complete information.

The structure of the incomplete information game reveals that the game between "group-lawyer" under the current institutional framework in China is obviously biased towards the lawyer (as long as he obtains the right to be represented). First of all, expanding the size of the representation group is a strict advantage strategy for lawyers. Because the more victims a lawyer represents, the greater the influence he exerts on the litigation, which in turn will further motivate him, driven by the maximization of the benefits of litigation representation, and the lawyer will try to include as many clients as possible in the group. Second, since the right of representation of the lawyer representing one of the group members is not obtained on the basis of the agreement between the group and each victim, the members cannot enter into a separate agency agreement with the lawyer, but through the representative of the group and the lawyer, therefore, the lawyer representing one of the group is actually directly responsible to the representative and indirectly responsible to the principal, and the party in the group is not fully aware of the situation of the lawyer, and the game between the two is an incomplete information game.

As with any group, there are always high costs associated with representing, organizing, and managing a class action. However, not every victim is willing to pay agency fees that are considered "collective goods", and if one victim no longer pays for the "collective goods" he enjoys, the burden on other victims increases, so they refuse to continue to contribute, and the "collective goods" cease to exist, especially in low-value victim-dominated class actions. As a sub-optimal option, the game between the group party and the lawyer often balances the strategy of winning the case and getting paid, so as to effectively link the agency fee with the interests of the group.

The practice of class litigation in China and the United States has recognized the feasibility and necessity of the remuneration system for winning the lawsuit, and the game equilibrium around this makes the interests of lawyers and group parties more compatible and reduces the conflict of interest in litigation representation. Suppose that a securities fraud infringement causes each investor to suffer a loss of 1,000 yuan, and the cost of defending rights through litigation is 1,000 yuan, it is obvious that each injured investor will not pay 1,000 yuan of litigation costs to sue for rights protection alone, and the litigation benefit is 0. We can further assume that if 1,000 investors are injured, the probability of the defendant performing the obligation is 60%, and the lawyer can be paid 10% of the winning interest, then the lawyer will have an agency income of 60,000 yuan. This is enough to attract a lawyer to represent the case. This may explain why lawyers do not hesitate to advance funds in mass tort litigation such as consumer torts, product liability and environmental torts, while at the same time expanding the size of the group to make the group action mechanism more commercialized and provide financial incentives for the operation of the class action mechanism.

In stark contrast to the advantages of the lawyer, the group of parties is in a disadvantageous position under the system of remuneration for winning the case. First, the amount of compensation claimed by each victim in a class action may be low, and the willingness to sue is not strong, which invisibly weakens the incentive to supervise the lawyer. The victim is indifferent to the proceedings and the outcome of the proceedings and is represented by a lawyer. The same is true for litigation representatives, where the reality is often that the lawyer appointed by one of the groups will contact the representative, rather than the other way around. Second, in a group action planned by a lawyer, both the parties and their representatives may be set aside and cannot have an impact on the litigation. Once lawyers have been granted the right to represent in litigation, they do not have to pay a cost to impress their clients, especially under the opt-out system, and they do not even have to win the trust of the latter. On the other hand, many low-value victims also lack the motivation to change or complain about their lawyers, and even acquiesce to lawyers' representation in violation of professional ethics. Thirdly, for a lawyer, the larger the group of victims, the greater his impact on the litigation, and expanding the size of the group of agents is his advantageous strategy, which will also improve his ability to "legalize and blackmail" the defendant. In addition, after representing a group of clients, the lawyer may also use the termination of the agency contract as his own gambling strategy. For example, in the course of settlement negotiations between a group lawyer and a defendant, if a large number of victims dispute the settlement agreement or do not agree to the settlement agreement, the lawyer may threaten to terminate the agency contract, even though they have no intention of resigning from the litigation agency.

Based on the above-mentioned characteristics of the "group-lawyer" game, it is necessary to regulate the litigation agency activities of lawyers. First of all, if a lawyer represents a party to a group and reaches a mediation agreement with the other party, changes the litigation claim, or admits the other party's litigation claim, the court should be given the right to review the situation where it is difficult for the group (especially in mass tort disputes) to reach an agreement and collective action. The lawyer's representation agreement for remuneration for winning the case should be the focus of review, and the court should review and approve the calculation and payment method of the lawyer's fee, so that the lawyer's representation can only be paid when the lawyer's representation brings actual benefits to many victims. Second, it regulates lawyers' representation behavior and prevents lawyers from seeking excessive benefits through group litigation. In this regard, it is necessary to learn from the "reverse auction" mechanism in class action litigation in the United States, that is, the legal services of group representation must be determined through bidding, the court will issue a tender announcement to set the minimum standard of legal services, and lawyers can obtain litigation agency qualifications through competitive bidding, thereby weakening the unreasonable advantageous position of lawyers in the litigation agency game. Third, strengthen the information disclosure mechanism for lawyers. After the group parties elected representatives, they actually withdrew from the litigation and turned to passive and limited participation in the litigation, which also lacked the ability to control the litigation and could not effectively supervise the litigation lawyers, which provided opportunities for the lawyers to manipulate the procedures and deliberately conceal and distort information and mislead the group plaintiffs. The way to change this information asymmetry is to force the lawyer to disclose all litigation information and eliminate the disadvantages of incomplete information games.

In short, the enthusiasm of lawyers to represent group litigation cases determines the effectiveness of the group litigation system, but the incentive measures for lawyers' enthusiasm for representation should be reasonable. On the one hand, the lawyer should receive sufficient incentives and have the opportunity to obtain the necessary benefits to give him the motivation to represent the group of clients, and the rule of remuneration for winning the case should be maintained. On the other hand, it is also necessary to reduce the remuneration income from winning the case to a reasonable range, so as to prevent lawyers from making excessive profits and artificially creating group cases, and avoid the waste of judicial resources.

5.conclusion

The game analysis of group litigation can provide theoretical enlightenment for improving the group litigation system from the perspectives of legal economics and behavioral psychology. First of all, the operation of the group litigation system is inseparable from the litigation cooperation of all parties. Group formation, litigation settlement, case acceptance, and even lawyer representation are all the results of the transition from adversarial competition to game equilibrium among specific subjects, and are the dynamic process from confrontation to cooperation, which is the law of group litigation. Second, all litigants, including the courts, are rational individuals in group litigation, and will pursue the maximization of their own interests and have their own optimal strategic choices. However, the group action system contains collective rationality, and even if the decision-making of the subject of the litigation may conflict with it and lead to the "prisoner's dilemma", the Nash equilibrium of the static game can be reached to find the optimal solution, so people do not have to fear that the group action will endanger social stability. Thirdly, the stage of litigation determines the dynamic and repetitive nature of the group litigation game, and both parties to the game have complete procedural choices and substantive disposition plans, and interact with each other to carry out litigation actions and have interactive paths, which also reflect the behavioral laws of the litigation subject's procedural selection and substantive disposition. Finally, the information or evidence in the class action is of great significance for each game subject to make decisions. Litigation information is not only the basis for the parties' procedural choices, but also the key to their game. Screening and refining litigation information can help make the right litigation decisions, and the more complete the information, the more likely it is to achieve Pareto optimality. To this end, full attention should be paid to the disclosure of private information and the exchange of evidence between private and public information in group litigation, so as to ensure the rational choice of the subject of litigation and promote the reasonable resolution of large-scale and complex disputes.