Characterizing societal challenges using game theory: enlightenment in a time of confusion

KAMIJO Yoshio

Specialized field

Game theory, Economics


Exploring the impact of racial diversity on the reward and punishment approach

In everyday life, people compete and sometimes cooperate; it's like a kind of game. Game theory is the science of predicting the behavior of opponents in scenes involving deals, through mathematical analysis of rational decision making. Mathematical modeling can give us an understanding of complicated human behavior and social phenomena, so game theory is being used for solving all sorts of problems, in areas ranging from economics to business to everyday affairs.

In his work as a game theorist, Professor Kamijo has been exploring various real-world societal phenomena using this theory. In particular he is focusing on research related to reward and punishment. There are two main methods for evoking motivation in members of an organization or a group: compensation type methods, that give rewards to the upper part of the group structure, and punitive type, that punish the bottom.

Prof. Kamijo explains, "We use these two types of motivators somewhat unconsciously, and there is no clear rule for choosing the appropriate one." Is that kind of selection rational? How does selection affect a group's overall performance? Game theory analysis tells us that reward is appropriate for cases where there is a wide disparity in capacity among group members, for example in a physical education class; and that punishment is suitable for cases where the members are largely similar in capacity, e.g., a football team.

This theme of rewards and punishments was originally considered a topic in economics, but Prof. Kamijo was the first researcher to develop a successful formula using the reward and punishment model with a game theory approach.

Currently, Prof. Kamijo has extended the scope of his research interest to include the influence of racial diversity on the reward and punishment model.

He explains, "For example, the United States population is composed of diverse races, and the disparity in ability among members is large, while in Japan diversity is low. I observed that game theory would suggest that America is a rewarding society that rewards the top, and Japan is a punitive type of society that punishes the bottom. I feel that this rings true in reality."

Prof. Kamijo is now conducting a survey to demonstrate this theory by comparing the reward-punishment behavior of Japan and the United States.

Using online games to verify game theory predictions

Through game theory analysis of situations involving multiple subjects, it is possible to arrive at guidelines for judgment for any given scenario. However, when it comes to applying game theory to problems in real society, a number of difficulties can arise.

Prof. Kamijo says, "People can't be sure how many options they have in a given situation, nor can they be clear about the degree of satisfaction resulting from choosing an option. People usually make vague choices based on their assumptions about a situation. It seems to me that only in rather limited situations can people anticipate the degree of satisfaction that would result from each choice."

In his thinking about the extent to which game theory is applicable to reality, Prof. Kamijo came up with the idea of using online games as a means of experimentation.

One of the biggest problems when applying game theory to real-world behavior is that it is extremely difficult to measure people's true benefit from strategy choices. In the real world, it is possible to measure benefit by observing the winning percentage in a survival competition among individuals in a controlled environment. However, when we confront real decision scenarios, individual character and complicated rules are involved, so strategy outcomes are often difficult.

Prof. Kamijo explained his rationale: "In online card games, the users compete in an anonymous environment, and the influence of individual characteristics and the relationships among players is relatively small, so I thought people's choices or tactics would not be too fragmented, at least in some types of card game."

Prof. Kamijo verified the game theory approach in the context of a competitive online card game. Each user freely made decks of 40 cards by selecting cards from a deck of 400. Then the number of decks was reduced to those that are frequently used, numbering about five. Then about 700 games were played in a round-robin competition, and the percentage of wins was calculated. In the course of play, individual statistical surveys were conducted to record the decks used by the opponents. Prof. Kamijo conducted game theory analysis of the data and examined the results to see when the data agreed with actual conditions.

He says, "The data and the actual conditions did not agree in this study, but I felt I might be able to use this methodology to determine how well game theory predictions line up with reality."

This use of online games for verification of game theory predictions appears promising. Prof. Kamijo says, "It is significant that we could predict the number of victory and defeat outcomes of actual games."

"In the field of experimental economics, studies often observe how people behave in a constructed bargaining scenario by assigning numerical values to outcomes, but not many studies observe the payoffs people can get for different choices in actual situations.

"If a researcher only wants to validate a theory, all that is required is to observe how people behave by assigning numerical values to outcomes, but if the researcher wants to determine outcomes in a real-world environment, direct measurement is necessary. In the virtual reality of online games, the game theoretic payoff was actually checked to determine whether people's behavior does in fact coincide with the predictions of game theory. I think this is quite a progressive approach. Expansion of this sort of research will lead to the development of new theories."

Demeny voting to correct for policy distortion in super aging societies

Prof. Kamijo is also tackling major social problems in his research. One of the topics is the prevention of policy distortion using a new voting method called Demeny voting, in which children have the vote, and parents vote as their proxies. When the number of senior citizen voters increases due to declining birthrate and aging population, policies favorable to the elderly tend to be chosen, i.e. the skewed population creates policy that is indifferent to the future of the young generation. Demeny voting is a groundbreaking idea devised in 1986 by demographer Paul Demeny in order to correct greying based policy distortion. Although some countries have considered adopting the system, none has yet done so.

Prof. Kamijo has been studying the potential outcomes of implementation of this new voting system. One specific research theme is how the shift to a multiple vote system would impact on the actions of those individuals with only one vote.

Prof. Kamijo speculates, "There are two possibilities that I can think of. The first is that people with one vote might think that the value of their vote is diminished and that individual influence has declined as a result of the alloting of multiple votes to some individuals. The second possibility is that people might come to have positive feelings about the notion of voting for the future. How will these two outcomes influence the behavior of people with one vote? To investigate that question, I conducted an experiment."

To analyze a choice between two policies, A, which is favorable for the future and B, which does not take the future into consideration, an experiment was conducted in which the subjects were offered two possible effects, and asked which one they would vote for. Those who reported that they thought that the value of their vote would go down voted for B and those who imagined they were voting for the future voted for A.

Prof. Kamijo explains, "Those who think that Demeny voting will only strengthen the rights of parents tend to select the option favorable for their generation. Meanwhile, those who are told that their vote would be a vote to improve the future reacted in favor of the other option. This suggests that when introducing Demeny voting, a thorough explanation of the concept is necessary, since success depends on the acceptance of Demeny voting by society as a whole." Thus, it can be said that Demeny voting, which was expected to serve as a fresh step towards a sustainable society, was shown to have a potential defect.

Prof. Kamijo gives his perspective, saying, "Since the social services system is strained to capacity, I believe it is important to generate society-wide debate with an objective clarification of the strengths and weaknesses of Demeny voting, a system which could shake the foundation of democracy."

Further pursuit of this research direction will be a catalyst for serious discussion of the mechanism of the review of the current configuration of democracy. As a result, even if Demeny voting were not adopted, there is a possibility that a new social system would emerge from those discussions. Prof. Kamijo says, "I am aiming at realizing a sustainable society that gives consideration to the situation of future generations." His numerous studies looking well into the future are intended to support that realization.