Welcome back to our exploration of mechanism design for relationship agreements. In our previous lesson, we established the principle of incentive compatibility, where we design rules so that being honest is a Nash Equilibrium—it's the best strategy for you, assuming your partners are also being honest.
Today, we raise the bar. We will analyze the concept of strategy-proofness, a much stronger and more desirable property for relationship rules. A strategy-proof rule makes honesty the best policy for a partner, regardless of what the other partners say or do. This is the gold standard for creating agreements that are robust, transparent, and resistant to strategic manipulation. Understanding this concept will allow you to critically evaluate common relationship norms and identify which ones are built on a solid foundation of trust and which may inadvertently encourage gamesmanship.
Strategy-Proofness: The Dominant Strategy of Truth
In game-theoretic terms, an incentive-compatible mechanism makes truth-telling a best response to others telling the truth. A strategy-proof mechanism makes truth-telling a dominant strategy. You don't need to guess or assume what your partners will do; being honest is always your best move, period.
This is a powerful guarantee. But how can a rule be designed to make its strategy-proofness obvious? The key often lies in how the rule is described. A recent line of research proposes framing mechanisms as "menu descriptions," which elegantly exposes this property.
[PDF] Strategyproofness-Exposing Mechanism Descriptions - NBER
This working paper by Gonczarowski, Heffetz, and Thomas introduces a clear way to think about and describe strategy-proof mechanisms.
Please read the following sections: Begin with the introduction to get a high-level overview of the concept. Next, study Table 1 and its surrounding text, which presents "Traditional" vs. "Menu" descriptions for two classic mechanisms: Median Voting and the Second-Price Auction. This provides concrete examples of the reframing. Finally, read Section 2.2, "Menu Descriptions." Pay close attention to Definition 2.5 and Theorem 2.6, which formally state that a mechanism is strategy-proof if and only if it always gives a player their favorite outcome from their menu. The one-paragraph proof is the core insight: your report can't change the menu of options available to you, so your best bet is to report truthfully to get the best item from that menu.
The core idea of a menu description is to separate the process into two steps for any given partner:
- Menu Calculation: The reports of all other partners determine a "menu" of possible outcomes for you. Crucially, your own report has no influence on this menu.
- Menu Selection: Your report is then used to simply select your most preferred outcome from the menu that was presented to you.
Framed this way, it's immediately clear why you should be truthful. Lying cannot expand your menu of options, it can only risk you getting a less-preferred option from it.
A Concrete Example: The Median Voting Rule
Let's explore one of the most famous strategy-proof mechanisms: the median voting rule. This rule is perfect for making a group decision on a topic where preferences can be lined up along a single dimension (e.g., a budget, a temperature setting, a time to meet).
The following video provides an excellent walkthrough of how this rule works and an intuitive argument for why it's strategy-proof.
(AGT11E11) [Game Theory] Strategy Proof Mechanism & Single Peaked Preferences: Median Voting Rule
This video from the channel selcuk ozyurt gives a clear explanation of the Median Voting Rule.
Watch the video in its entirety. First, listen to the introduction to the concept of a strategy-proof rule. Then, follow the example of how to find the median voter based on their "peak" preferences. The key is that voters are ordered by their ideal points, not by their names or some arbitrary sequence. Finally, pay close attention to the argument for why it's strategy-proof. The lecturer explains why neither the median voter nor any other voter can benefit from misrepresenting their preference.
Let's apply this to a vee relationship. Imagine a hinge H and two partners P1 and P2 need to decide how many nights a week H will spend at P1's home versus P2's. This is a classic single-dimensional problem. Let's say P1 prefers 4 nights, P2 prefers 2 nights, and H feels that 3 nights is the ideal balance.
Their ideal points are 2 (P2), 3 (H), and 4 (P1).
The median voter is H, whose peak preference is 3. The outcome is 3 nights.
Why is this strategy-proof?
Hhas no incentive to lie; they are getting their ideal outcome.P2(who wants 2) could lie and say they want 1. The set of ideal points becomes 1, 3, 4. The median is still 3. The outcome is unchanged.P2could lie and say they want 3. The set is 3, 3, 4. The median is still 3.P2cannot manipulate the outcome to be 2. By lying, they can only risk shifting the median further away from their ideal point (e.g., if they lied and said their ideal was 5, the set would be 3, 4, 5 and the median would become 4, which is worse for them).- The same logic applies to
P1.
This rule works beautifully, but it depends on a critical assumption: single-peaked preferences. It works for budget, time, or any other quantifiable, linear choice. It fails for multi-attribute choices, like picking a restaurant (Cuisine? Price? Location?), where preferences are not easily ordered along one line.
The Limits of Perfection: Impossibility Theorems
Strategy-proofness is so powerful that it's natural to ask if we can always achieve it. The sobering answer comes from one of the most foundational results in social choice theory: Arrow's Impossibility Theorem.
This theorem states that for decisions involving three or more options, no voting system can simultaneously satisfy a handful of reasonable fairness criteria (like unanimity, no-dictatorship, etc.) without introducing the possibility of cycles or other problems. Strategy-proofness is deeply related to this. The Gibbard-Satterthwaite theorem, an extension of Arrow's work, proves that for general preferences (i.e., not necessarily single-peaked), the only deterministic strategy-proof voting rules are dictatorial.
Impossibility & Social Choice - Intro to Political Economy, Lecture15
This lecture from Duke University's political economy course provides an excellent overview of Arrow's Theorem and its profound implications.
Please watch the following segments: The explanation of the revelation problem and Arrow's Theorem. The key takeaway is that for general preferences, the only way to guarantee a determinate, rational outcome that satisfies certain fairness axioms is to have a dictator. A review of the desirable characteristics and the statement of the theorem's conclusion. The shocking result is that any mechanism satisfying the first four properties (unanimity, transitivity, unrestricted domain, independence of irrelevant alternatives) must be a dictatorship.
This presents a stark choice in designing relationship agreements:
- Option A: Restrict the domain of decisions to ones with special structure, like single-peaked preferences, where non-dictatorial strategy-proof rules like Median Voting exist.
- Option B: Use a dictatorial rule. For example, "The hinge decides on weekend plans." This is strategy-proof (the partners' preferences don't influence the outcome, so they have no incentive to lie), but it's explicitly hierarchical.
- Option C: Accept that your rule is not strategy-proof and is therefore vulnerable to strategic manipulation.
Analyzing Vee Relationship Rules
Many "common sense" rules or norms in polyamory fall into Option C. They sound good on paper but are not designed to be strategy-proof, leaving them open to being used as tools of control, as highlighted in the infographic below.
Let's analyze some rules, including some mentioned in the infographic, using the concept of strategy-proofness.
-
Veto Power: The infographic lists "Veto (demanding end to another relationship)" as a potential element of control. As a mechanism, a veto is notoriously not strategy-proof. A partner's true "type" is their level of happiness or distress. A veto rule incentivizes a partner to misrepresent their distress—to exaggerate their unhappiness or to make strategic threats of a veto—in order to control the hinge's behavior or extract concessions. Honesty is not the dominant strategy; strategic leveraging is.
-
Hierarchical Rules ("Primary" Partner): A rule stating "The primary partner's preferences are considered first" can be modeled as a Serial Dictatorship.
P1(the primary) chooses their best option from the whole set of possibilities. ThenP2chooses their best option from what remains. As discussed in theGonczarowskipaper, this mechanism is strategy-proof. Each person's best strategy is to honestly state their preferences to get the best outcome available to them at their turn. The trade-off is not in stability, but in fairness. It is explicitly non-egalitarian, which may or may not be acceptable to the participants, but it is at least transparent and not manipulable by individuals. -
"Radical Honesty" as a Rule: This is a norm that encourages revealing all information and feelings. However, without a mechanism that makes it safe and productive, it is not strategy-proof. It is a "cheap talk" game. If a partner learns that revealing a certain feeling (e.g., insecurity) leads to punishment or a fight, they are incentivized to conceal that feeling in the future. The infographic's example, "Demanding detailed accounts of sex with other lovers," is a coercive version of this. It demands information but provides no guarantee that the information won't be used against the person revealing it, thus failing the strategy-proofness test spectacularly.
A Higher Bar: Group Strategy-Proofness
Even if a rule is strategy-proof for individuals, can it be manipulated by a coalition? A mechanism is group-strategy-proof (GSP) if no group of partners can coordinate their lies to make everyone in the group better off.
Group Strategyproofness in Mechanism Design - Emergent Mind
This summary from Emergent Mind defines the concept of group strategy-proofness.
Please read the introductory definition and Section 1, "Formal Definitions and Conceptual Framework." Focus on the core idea: GSP rules out coordinated manipulations by coalitions.
GSP is an extremely demanding property. For a vee, it would mean that the two non-hinge partners could not collude with their reported preferences to manipulate the hinge. For many mechanisms, achieving GSP is impossible without resorting to dictatorial or randomized outcomes. This is often a theoretical frontier, but it's worth being aware that even individually-sound rules can sometimes be vulnerable to group strategy.
Conclusion
Today we've delved into strategy-proofness, the cornerstone of designing truly manipulation-resistant agreements. By making honesty a dominant strategy, these rules remove the need for partners to second-guess each other and allow for more efficient and trustworthy decision-making.
- Key Takeaways:
- Strategy-Proofness (SP) is a stronger property than incentive compatibility; it means truth-telling is a dominant strategy, optimal regardless of others' actions.
- Menu descriptions are a powerful way to frame rules that makes their strategy-proof nature intuitive and transparent.
- The Median Voting Rule is a classic, non-dictatorial SP mechanism, but it requires that choices can be mapped to a single dimension with single-peaked preferences.
- For general choices, the Gibbard-Satterthwaite theorem shows that strategy-proofness usually requires a trade-off: either accept a dictatorial rule or a limitation on the types of decisions you can apply it to.
- Many common relationship "rules" like veto power or unstructured "radical honesty" are not strategy-proof and can be co-opted for control and manipulation.
In our next and final lesson of this module, we will examine the Vickrey-Clarke-Groves (VCG) mechanism. VCG is a remarkable and powerful strategy-proof tool for making collective decisions that maximize the total value for the group, but it comes with its own complexities—namely, the use of "payments." We will explore how this concept can be translated into the non-monetary world of relationships.