The Prisoner's Dilemma: Why Is Cooperation So Difficult for Us?

Unpacking the tragic logic where individual bests lead to group worsts, and exploring ways to design better cooperation.

• Understand the core concepts of Prisoner’s Dilemma and Nash Equilibrium.
• Analyze why dilemmas occur in reality, such as during the Cold War and climate change.
• Discover principles where human irrationality can actually lead to cooperation.

Introduction: The Unsolvable Puzzle of Cooperation

Think of a shared kitchen in an office. Everyone wants a clean environment, but no one is willing to take the initiative to clean. This trivial dilemma illustrates the essence of the Prisoner’s Dilemma, where ‘group benefit’ and ‘individual convenience’ clash. The Prisoner’s Dilemma is a powerful model that explains not only simple kitchen issues but also corporate price wars, arms races between nations, and humanity’s failure to respond to climate change.

Are we inherently selfish, or is there a hidden code for cooperation? In this article, we embark on a journey to find answers through the cold mathematics of game theory and the human insights of behavioral economics.

1. The Prisoner’s Dilemma: The Trap of ‘Rational’ Choices

Consider the classic scenario where two accomplices are interrogated in separate rooms. The investigator offers each a tempting proposal.

Rules and Outcomes of the Game

• Choices: Either keep faith with the partner by ‘Cooperating (Silence)’ or betray the partner by ‘Defecting (Confessing)’.
• Outcomes (Sentences):
  • Both cooperate: 1 year each
  • One defects, one cooperates: Defector 0 years, cooperator 10 years
  • Both defect: 5 years each

This situation can be summarized in the following table.

Table 1: Payoff Matrix of the Prisoner’s Dilemma (Unit: Sentence)

	Suspect B: Cooperate (Silence)	Suspect B: Defect (Confess)
Suspect A: Cooperate (Silence)	(1 year, 1 year)	(10 years, 0 years)
Suspect A: Defect (Confess)	(0 years, 10 years)	(5 years, 5 years)

The Inevitable Logic of Betrayal

Let’s think from Suspect A’s perspective.

• If B stays silent, I benefit from defecting (1 year vs 0 years).
• If B confesses, I still benefit from defecting (10 years vs 5 years).

A strategy that is always advantageous to me, regardless of the other’s choice, is called a Dominant Strategy. In this game, the dominant strategy is ‘defection’. Both rational suspects choose to defect, resulting in both receiving a 5-year sentence. This state, where all participants choose their best strategy in response to others, leaving no incentive for anyone to change their strategy alone, is known as Nash Equilibrium.

This is not due to individual logic being flawed, but rather a trap of rationality created by a system structure where trust and communication are impossible. This dilemma was most dramatically illustrated during the nuclear arms race between the U.S. and the Soviet Union during the Cold War. Both countries would have benefited from disarmament (‘cooperation’), but fearing betrayal from the other, they poured immense resources into nuclear weapons development (‘defection’).

2. The Dilemma of Many: Group Projects and the Tragedy of the Commons

Dilemmas do not only occur with two individuals. The ‘free-rider’ problem seen in university group projects is a prime example of a Public Goods Game involving multiple participants.

If everyone puts in effort, they receive an A+, but the most rational choice for an individual is to do nothing and ride on the efforts of others. This temptation leads to the ‘Tragedy of the Commons’. In a common pasture without an owner, if each shepherd lets one more cow graze for their own benefit, the pasture eventually becomes barren, harming everyone.

This model accurately describes the climate change crisis.

• Public Good: A stable Earth climate
• Free-Rider Incentive: Each country hopes that others will bear the costs of reducing carbon emissions while they continue to pollute for economic gain.
• Outcome: Global reduction efforts falter, and ultimately everyone faces the tragedy of climate disaster.

[Insight 1] As the number of participants increases, the responsibility is diluted, reducing the burden of defection (free-riding). While defection is clearly visible with two people, in the climate change issue involving 200 countries, the deviation of one country is less noticeable. This is why designing a coercive system, such as a carbon tax or international agreements, is essential to solve the dilemma rather than relying on informal trust.

3. Human Glitches: We Are Not as Selfish as We Think

“Someone offers you $10 if you give them $100. Would you accept this offer?”

According to classical game theory, accepting any offer greater than 0 is rational. However, reality is different. At this point, behavioral economics emerges, revealing that humans are not simple calculating machines but are influenced by social preferences such as fairness, reciprocity, and altruism.

Evidence from the Laboratory: The Ultimatum Game

• Setup: The proposer suggests how to split the money, and the responder can accept or reject. If rejected, neither receives anything.
• Reality: Unfair offers below 20-30% of the total are mostly rejected. People are willing to forgo their own benefits to punish unfair treatment.

This may appear as a ‘bug’ in individual economic calculations, but it is a very important ‘feature’ from the collective long-term perspective. The ‘irrational’ response of feeling anger and punishing unfairness serves as a powerful adhesive that enforces social norms and builds trust.

4. The Long Game: How Does Cooperation Evolve?

Political scientist Robert Axelrod sought to uncover the secrets of cooperation through a computer tournament of repeated Prisoner’s Dilemmas. When the game is not a one-off but continues to repeat, the ‘shadow of the future’ influences present decisions. Today’s actions become tomorrow’s reputation, and trust becomes an asset.

The tournament’s winner was one of the simplest strategies, ‘Tit-for-Tat (TFT)’.

1. Always cooperate on the first move.
2. Then, mimic the opponent’s last action.

The success of Tit-for-Tat lies in four characteristics.

• Niceness: Never defects first.
• Retaliatory: Immediately punishes defection to prevent exploitation.
• Forgiving: Quickly forgives if the opponent returns to cooperation, restoring the relationship.
• Clear: The strategy is simple, allowing the opponent to learn how to cooperate quickly.

[Insight 2] The success of Tit-for-Tat is not because it is ’nice’, but because it adheres to the principle of ‘reciprocity’. From my personal experience, in business negotiations, initially showing trust and approaching generously, but responding firmly when the other party breaks promises, and restoring the relationship when they show sincerity, has yielded the best long-term results. Cooperation is not inherent; it is an emergent property created by designing a reciprocal environment.

5. Taming Free-Riders: The Double-Edged Sword of Punishment

In Ernst Fehr’s public goods game experiment, introducing a ‘punishment’ option led to surprising results. People spent their own money to punish low-contributing participants (altruistic punishment), resulting in cooperation levels soaring to nearly 100%.

However, this story does not end here. Follow-up studies conducted in 16 cities worldwide observed the opposite phenomenon in some cultures. ‘Antisocial Punishment’, where cooperative individuals are attacked, was noted. This phenomenon was particularly pronounced in societies with weak civic cooperation norms and low trust in the rule of law.

Comparison: Mechanisms that Induce Cooperation

Mechanism	Description	Effects and Implications
Altruistic Punishment	Punishing free-riders at one’s own cost	Dramatically increases cooperation levels in high-trust societies.
Antisocial Punishment	Punishing high-contributing individuals	Can destroy cooperation in low-trust societies due to jealousy or revenge.
Reciprocity (Tit-for-Tat)	Mimicking the opponent’s actions	Induces stable cooperation in repeated interactions.
Reputation and Transparency	Making individual actions transparent	Can lead to voluntary cooperation through social pressure.

This shows that the effectiveness of punishment systems depends on the cultural and institutional soil in which they are rooted. Before introducing punishment systems, fundamental norms of trust and civic cooperation must be established.

System Design Guide for Cooperation

Humans are neither angels nor demons but ‘conditional cooperators’. Instead of trying to change our nature, we should focus on designing systems that foster cooperation.

1. Cast a long shadow of the future: Encourage long-term relationships over one-off encounters so that today’s actions impact the future.
2. Increase transparency and reduce anonymity: Clearly show who did what to activate the reputation system.
3. Facilitate communication: Simple conversations can resolve distrust and lay the groundwork for cooperation.
4. Use rewards and punishments wisely: When introducing a sanction system, first build social consensus that it is fair and just.

Conclusion

• Key Point 1: The Prisoner’s Dilemma illustrates the ’trap of rationality’ where rational individuals make decisions harmful to the group in the absence of trust and communication.
• Key Point 2: Humans are not entirely selfish, and ‘irrational’ emotions regarding fairness and reciprocity are crucial keys that enable cooperation.
• Key Point 3: Successful cooperation depends not on finding good people but on designing transparent, long-term systems where reciprocal strategies like Tit-for-Tat can thrive.

Ultimately, decoding cooperation is not about changing human nature but about creating an environment where our conditional cooperative tendencies can manifest positively. What are the ‘rules of the game’ that hinder cooperation in your organization or community? What small actions can you take today to change those rules?

References

• Namu Wiki Prisoner’s Dilemma
• Wikipedia Tragedy of the Commons
• Chosun Ilbo Are Humans Selfish or Altruistic? The Surprising Answer from the ‘Ultimatum Game’
• Namu Wiki Tit-for-Tat
• Herrmann, B., Thöni, C., & Gächter, S. (2008). Antisocial Punishment Across Societies. Science. link