Welcome to the first lesson in our new module, "Mechanism Design for Relationship Agreements." In the previous module, we focused on analyzing existing relationship structures as cooperative games, evaluating their stability by identifying the core. We were essentially analysts, predicting whether a given arrangement could survive. Now, we shift from being analysts to being architects.
This module is about Mechanism Design—the art and science of designing the rules of the game to achieve a specific, desired outcome. Instead of taking the rules as given, we will learn how to construct them to align incentives, encourage honesty, and ensure agreements are self-enforcing.
Today's lesson addresses the foundational challenge of this field: framing the creation of relationship rules as a principal-agent problem and identifying the two critical obstacles that arise from it: adverse selection and moral hazard. These concepts are central to economics and contract theory, and they provide a powerful lens for understanding why even well-intentioned relationship agreements can fail.
The Principal-Agent Framework
At its core, a principal-agent problem arises whenever one person (the "principal") delegates a task to another person (the "agent"), but their incentives are not perfectly aligned. The principal wants the agent to act in the principal's best interest, but the agent has their own goals and motivations. The principal cannot perfectly monitor or control the agent's actions, creating a strategic dilemma.
To get a clear, general overview of this concept, please watch the following short video.
This video from the channel of economist Ashley Hodgson provides a concise introduction to the principal-agent problem, explaining the core dynamic and the roles of the principal and agent.
Watch the first part of the video, from the beginning to the main explanation. Pay attention to how the roles of principal and agent are not fixed titles but are defined by who is trying to incentivize whom.
As the video explains, the roles can be fluid. In a relationship negotiation:
- If Hannah (the hotwife) is designing a set of rules for a new dynamic, she is the principal, and Charles (the cuckold), who will operate under these rules, is the agent.
- Conversely, if Charles is setting boundaries that he expects Hannah to respect during her dates, he is the principal, and she is the agent.
- When they are jointly trying to screen a new potential partner (a "bull"), they act as a single principal, and the new partner is the agent.
The central difficulty in any principal-agent relationship is asymmetric information: one party knows something relevant that the other does not. This information imbalance creates two major problems, which occur at different stages of the agreement. The next video introduces these concepts using the classic example of insurance markets.
Asymmetric Information, Adverse Selection & Moral Hazard | Economics Explained
This video from INOMICS clearly defines asymmetric information and then explains its two main consequences: adverse selection and moral hazard.
Please watch the following segments: First, the definition of asymmetric information. Next, the explanation of adverse selection, which is a problem of hidden information before a deal is struck. Finally, the explanation of moral hazard, which is a problem of hidden action after the deal is struck.
Now that we have the formal definitions, let's apply them directly to our context of relationship agreements.
1. Adverse Selection: The Problem of Hidden Information
Adverse selection arises from hidden information before an agreement is made. It's a problem of screening or selection—how does a principal choose the right agent when the agent's true "type" is hidden?
Scenario: Screening a New Partner
Imagine Hannah and Charles want to open their relationship to a new secondary partner, Ben. They are looking for someone who is not only sexually compatible but also emotionally mature, respectful of their primary relationship, and honest about their sexual health. Ben's true character and intentions are his private information. He is a "type" that is unknown to Hannah and Charles.
- Good Type: A Ben who genuinely shares their values and is looking for the kind of connection they are offering.
- "Lemon" Type: A Ben who is just saying what he thinks they want to hear to facilitate a casual encounter, with no intention of respecting their dynamic long-term.
This is a direct parallel to Akerlof's "Market for Lemons." If Hannah and Charles cannot reliably distinguish between the good types and the lemons, they might become generally distrustful. They might offer terms (e.g., extensive pre-date vetting, slow pacing) that a "good type" Ben finds onerous, while a "lemon" Ben might deceptively agree to anything. If the process is too difficult or the risk of picking a lemon is too high, Hannah and Charles might give up searching altogether. The market for good partners adversely selects against itself.
A powerful real-world example of this dynamic occurs in the context of sexual health disclosures on dating apps. The following paper analyzes this as a signaling problem, which is the theoretical framework for understanding how agents reveal hidden information. Given your background in Bayesian decision theory, you'll recognize the core logic of updating beliefs based on signals.
[PDF] Signal Appropriation of Explicit HIV Status Disclosure Fields in Sex ...
This paper applies signaling theory to understand how users of sex-social apps manage the disclosure of sensitive private information (HIV status). It provides a rigorous framework for thinking about hidden information and its consequences.
Please read the following two sections: Start with Section 2.4, "Signalling Theory". This provides the theoretical foundation from economics, including Akerlof's "Market for Lemons" and the concept of signaling. Pay close attention to the discussion of the "full disclosure principle" and the "privacy unraveling effect." This is a key game-theoretic insight: in a market where good types want to distinguish themselves, a refusal to disclose information can itself become a powerful negative signal. Someone who is silent about their intentions might be assumed to have bad ones. Next, read Section 4.3, "Non-Disclosure Cues". This section shows the unraveling effect in action, where participants fear that not disclosing their status will be interpreted as a signal that their status is undesirable.
The key takeaway is that in a situation with adverse selection, the absence of a signal is itself a signal. If Hannah asks a potential partner, "What are your long-term intentions regarding our dynamic?" and he evades the question, the "privacy unraveling effect" suggests she should infer the worst. This forces agents who are "good types" to find credible ways to signal their quality before the agreement is made.
2. Moral Hazard: The Problem of Hidden Action
Moral hazard arises from hidden action after an agreement is in place. The principal and agent have agreed to a "contract," but the principal cannot perfectly monitor the agent's subsequent behavior.
Scenario: Enforcing Relationship Rules
Let's say Charles (principal) and Hannah (agent) agree on a clear rule for her dates with Ben: "No alcohol consumption beyond two drinks to ensure clear-headed consent and decision-making." Hannah agrees to this rule.
The agreement is made. Now, Hannah goes on her date with Ben. Charles is not present and cannot observe her actions. The moral hazard is the risk that Hannah, now "insured" by the privacy of the date, will change her behavior. She might have a third or fourth drink, thinking, "Charles will never know, and it won't be a problem."
Her behavior is altered by the existence of the agreement combined with the lack of observability. The "hazard" is not that Hannah is inherently untrustworthy, but that the structure of the situation creates an incentive to deviate from the agreed-upon terms.
This is a common issue in many contexts. A compelling parallel comes from the business model of online dating platforms, which can be viewed as a large-scale principal-agent problem.
[PDF] Price of Anarchy in Algorithmic Matching of Romantic Partners
This paper frames the relationship between a dating site and its users as a principal-agent problem and quantifies the conflict of interest.
Please read the following sections: Read the introductory paragraph of the paper, from the beginning of Section 2. This explicitly frames the situation as a principal-agent problem with potential moral hazard. The "agent" (the platform) has different goals (maximizing engagement) than the "principal" (the user, who wants to find a perfect match and leave the platform). Now, read the first paragraph of Section 4, "Discussion". This section clearly articulates the moral hazard: to maximize revenue, the platform is incentivized to suggest matches that lead to short-lived relationships, encouraging users to return. This is a hidden action that works against the principal's interest.
Just as the dating platform has an incentive to provide mediocre matches to keep users engaged, a partner in a polyamorous dynamic might have a short-term incentive to bend a rule for immediate gratification, even if it undermines the long-term stability of the primary relationship (the principal's goal).
A Dialogue on Rule-Making
Let's see how these two problems manifest in a negotiation between Hannah and Charles as they prepare for her first meeting with Ben.
Hannah: "Okay, so Ben and I are set to meet for coffee on Thursday. I told him our ground rules: no physical intimacy on the first meeting, and the focus is on seeing if we have a personal connection."
Charles: "I appreciate that. But my concern isn't what he agrees to now. It's about what he's actually looking for. He seems great, but what if he's just a 'lemon'—someone who's incredibly charming and agrees to anything just to get in the door, but has no real intention of respecting our dynamic long-term? How do we know he's a 'good type'?"
Hannah: "That's a fair point. That's the adverse selection problem. We're selecting from a pool of candidates without knowing their true nature."
Charles: "Exactly. And then there's the other side of it. You two agree to 'no physical intimacy.' But I can't be there. Once you're there, the vibe might be great, and the temptation to break that rule might be strong. It's not that I don't trust you, but the situation creates a moral hazard. You're essentially 'insured' against my immediate reaction because I'm not there to see it."
Hannah: "I see. So we have two distinct problems. First, how do we filter out the people with hidden bad intentions before we even start? Second, once we have an agreement with someone we think is good, how do we create a situation where the rules are followed even when no one is watching?"
This dialogue perfectly separates the two challenges. Adverse selection is about filtering the lemons. Moral hazard is about ensuring the chosen agent doesn't start acting like one after the deal is done.
Conclusion
In this lesson, we have established the foundational framework for mechanism design by identifying the core challenges that arise when setting relationship rules.
Key Takeaways:
- Principal-Agent Problem: Setting and following relationship rules can be modeled as a principal-agent problem, where one party (the principal) designs an agreement for another party (the agent) to execute under conditions of misaligned incentives and imperfect monitoring.
- Asymmetric Information: The root of the problem is that the agent possesses private information, either about their "type" (hidden information) or their actions (hidden action).
- Adverse Selection (Hidden Information): This occurs before an agreement is made. The principal struggles to distinguish high-quality agents from low-quality "lemons," which can lead to poor choices or even market collapse (e.g., being unable to find a good partner).
- Moral Hazard (Hidden Action): This occurs after an agreement is made. The agent, being unobserved, may be tempted to act in their own self-interest in ways that violate the spirit or letter of the agreement.
Preview of the Next Lesson:
Now that we have formally defined these two critical problems, our next task is to learn how to solve them. In the upcoming lesson, we will explore one of the most powerful ideas in mechanism design: "Apply the Revelation Principle to design contracts that incentivize partners to truthfully report their preferences and risk tolerance." We will see how we can design rules so cleverly that it becomes in the agent's own best interest to reveal their hidden information truthfully.