📈 economics

Obviously Strategy-proof Choice of Social Acts

This paper characterizes the class of unanimous social choice functions over acts as being obviously strategy-proof implementable if and only if they are dictatorial.

Original authors: Abinash Panda, Anup Pramanik

Published 2026-06-01

📖 5 min read🧠 Deep dive

CC BY 4.0

Original authors: Abinash Panda, Anup Pramanik

Original paper licensed under CC BY 4.0 (http://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

Imagine a group of friends trying to decide on a group vacation plan. The twist? They don't know the weather yet. They have to pick a "contingency plan" that says, "If it rains, we go to the museum; if it's sunny, we go to the beach."

This paper explores how such groups can make decisions when everyone has different ideas about what they want (preferences) and different guesses about what the weather will be like (beliefs).

Here is the breakdown of the paper's story, using simple analogies:

1. The Setup: The "Weather Forecast" Game

The authors are looking at a specific type of decision-making called Social Choice over Acts.

The Scenario: A group must choose a plan that covers all possible future scenarios (states of nature).
The Players: Each person has their own "utility" (how much they like the museum vs. the beach) and their own "probability" (how likely they think rain is).
The Goal: They want a rule (a Social Choice Function) that picks the best plan for the group.

2. The First Rule: "Don't Lie" (Strategy-Proofness)

In the world of game theory, a rule is Strategy-Proof if it's always in your best interest to tell the truth.

The Analogy: Imagine a voting system where lying about your favorite movie never helps you get a better result than just saying what you actually like.
Previous Research: A study by Bahel and Sprumont (2020) found that in this "weather plan" scenario, there is actually a lot of flexibility. You can have complex, fair rules where no single person controls everything, and no one is tempted to lie. It's like having a sophisticated committee where everyone's opinion matters, and honesty is the best policy.

3. The New Rule: "Obvious" Honesty (Obvious Strategy-Proofness)

The authors ask a harder question: What if the rule isn't just "honesty is best," but "honesty is obviously best"?

The Concept: This comes from a concept called Obvious Strategy-Proofness (OSP).
The Analogy: Imagine a game where you have to choose a path.
- Normal Strategy-Proofness: You have to think, "If I lie, maybe I get a better outcome, but maybe I get a worse one. If I calculate the odds perfectly, I see that lying is risky, so I should tell the truth." This requires complex thinking.
- Obvious Strategy-Proofness: The game is designed so that even if you look at the worst possible thing that could happen if you tell the truth, it is still better than the best possible thing that could happen if you lie.
- The Metaphor: It's like choosing between a guaranteed $10 bill (Truth) and a lottery ticket that might win you $100 but might win you $0 (Lying). If the worst outcome of the lottery is $0, and the guaranteed $10 is better than $0, the choice is "obvious." You don't need to be a math genius to know you should take the $10.

4. The Big Discovery: The "Dictator" Problem

The paper's main result is a shock to the system.

The Finding: While there are many complex, fair rules where people should tell the truth (Strategy-Proof), there are almost no rules where it is obviously in their best interest to tell the truth, unless the rule is controlled by a single person.
The Metaphor: The authors prove that if you want a decision-making process where the "honesty" part is so simple that even a child or a confused person can see it immediately, the only way to do it is to let one person (a Dictator) make the choice.
- If one person is the "Dictator," the rule is simple: "Whatever the Dictator wants, we do."
- Why is this obvious? Because the Dictator knows exactly what they want. If they lie, they might get something they don't want. If they tell the truth, they get exactly what they want. There is no confusion.
- If you try to make a rule where two or more people share power (like a committee), the logic gets complicated. To prove that "lying is bad," you have to do complex mental math about probabilities and outcomes. Once you add that complexity, the rule stops being "Obviously Strategy-Proof."

5. The Conclusion

The paper concludes that in this specific world of uncertain future plans:

Complexity kills simplicity. The more you try to make a rule fair and shared among many people, the harder it becomes for people to see that they should be honest.
The Trade-off: You can have a fair, shared decision-making process where people should be honest (but might not realize it easily), OR you can have a process where honesty is blindingly obvious, but that process must be a dictatorship.

In short: If you want a group decision where everyone instantly knows, "Hey, I should just tell the truth," the only way to build that machine is to let one person call the shots. Any attempt to share the power makes the "honesty" part too confusing to be "obvious."

Technical Summary: Obviously Strategy-proof Choice of Social Acts

Problem Statement
This paper investigates the implementation of social choice functions (SCFs) within the framework of social choice over social acts, originally introduced by Bahel and Sprumont (2020). In this environment, a social planner must select a contingent plan (an act mapping states of nature to outcomes) before uncertainty is resolved. Agents evaluate these acts using the Subjective Expected Utility (SEU) criterion, meaning they possess both subjective valuations of outcomes and subjective probability measures over states of nature.

While Bahel and Sprumont (2020) demonstrated that the class of unanimous and strategy-proof SCFs in this domain is rich—specifically characterized as "locally bilateral top-selection" rules—this paper asks whether these rules satisfy the stronger condition of Obvious Strategy-Proofness (OSP). OSP, introduced by Li (2017), requires that truthful reporting be an obviously dominant strategy, meaning that even the worst possible outcome of following the strategy is at least as good as the best possible outcome of any deviation. The central question is whether the flexibility permitted by standard strategy-proofness survives when agents are required to "obviously" understand the optimality of truth-telling.

Methodology and Model
The authors adopt the formal model of Bahel and Sprumont (2020), defined by:

A finite set of agents $N$ , a finite set of states of nature $\Omega$ , and a set of social outcomes $X$ .
Agents hold strict SEU preferences, represented by injective valuation functions $v_i$ and injective probability measures $p_i$ .
Social choice functions map preference profiles to acts ( $\phi: D^n \to X^\Omega$ ).

The paper utilizes the definition of OSP mechanisms as extensive-form games (trees) where agents report preferences at internal nodes. A mechanism is OSP if, at every decision node for an agent, the utility of the worst-case outcome from following the truthful path is greater than or equal to the utility of the best-case outcome from any deviation.

The proof strategy relies on the characterization of unanimous and strategy-proof SCFs provided by Bahel and Sprumont (2020), which are locally bilateral top-selection rules. These rules partition the state space $\Omega$ into events, where the assignment of states to agents is determined by rules that are either constant, bilaterally dictatorial, or bilaterally consensual. The authors then analyze the structural constraints imposed by OSP on these specific assignment rules.

A key technical tool is Lemma 1, which establishes that for any belief $p_i$ , one can construct a valuation function $v_i$ with a specific geometric progression of values such that the resulting preference is strict (no indifference between distinct acts) for all but finitely many parameter values. This allows the authors to construct specific preference profiles where agents have strict preferences over the outcomes generated by different assignment rules, facilitating a contradiction argument.

Key Results
The paper's main result, Theorem 1, provides a complete characterization:

An SCF $\phi$ is unanimous and OSP implementable if and only if it is dictatorial.

The proof proceeds by contradiction, assuming the existence of a unanimous, OSP implementable SCF that is not dictatorial. Based on the Bahel and Sprumont characterization, such a function would correspond to a locally bilateral assignment rule. The authors demonstrate that if more than one agent is assigned a non-empty set of states (which is necessary for a non-dictatorial rule), one can construct two preference profiles that diverge at a specific node in the OSP mechanism. By carefully choosing valuations (using Lemma 1), they show that at the divergence node, the OSP condition would require an agent to prefer an outcome that contradicts their constructed preferences.

Specifically, the proof shows:

Claim 1: For any agent, the set of states assigned to them is either empty for all belief profiles or non-empty for all belief profiles.
Claim 2: At least one agent must be assigned a non-empty set of states for all profiles (due to the unanimity and partition properties).
Claim 3: It is impossible for two distinct agents to both be assigned non-empty sets of states for all profiles. If two agents were assigned states, the OSP condition would be violated by constructing a scenario where one agent prefers the outcome associated with the other agent's assigned states, creating a cycle of dominance that OSP cannot resolve.

Consequently, only one agent can be assigned the entire state space $\Omega$ across all profiles, rendering the SCF dictatorial. Conversely, the authors note that dictatorial SCFs are trivially OSP implementable.

Significance and Contribution
The paper contributes to the literature on obvious strategy-proofness by demonstrating that the implications of OSP are highly sensitive to the economic environment. While previous studies in voting environments with single-peaked domains (Bade and Gonczarowski, 2016; Arribillaga et al., 2020) or division problems (Arribillaga et al., 2023) have found non-dictatorial OSP implementable rules, this paper shows that in the context of social choice over acts with subjective uncertainty, OSP is extremely restrictive.

The authors claim that their result highlights a "strong tension between incentive simplicity and non-dictatorial collective choice" in environments where agents hold subjective expected utility preferences. While Bahel and Sprumont (2020) found that strategy-proofness allows for a rich class of locally bilateral rules, the imposition of the "obvious" requirement collapses this flexibility entirely, leaving only dictatorial rules as viable candidates for OSP implementation. This suggests that in complex environments involving subjective uncertainty, mechanisms that are easy for agents to understand (OSP) may be incompatible with non-dictatorial, unanimous social choice.