9AM GMT (6AM Rio de Janeiro, 9AM London, 10AM Paris, 12PM Istanbul, 2:30PM New Delhi, 5PM Shanghai, 6PM Tokyo/Seoul, 8PM Sydney, 10PM Auckland)
Huaxia Zeng (School of Economics, SUFE) "Equity in Strategic Exchange"
Host: Marcus Pivato
Abstract. New fairness notions aligned with the merit principle are proposed for designing exchange rules. We show that for an obviously strategy-proof, efficient and individually rational rule, (i) an agent receives her favorite object when others unanimously perceive her object the best, if and only if preferences are single-peaked, and (ii) an upper bound on fairness attainable is that, if two agents' objects are considered the best by all agents partitioned evenly into two groups, it is guaranteed that one, not both, gets her favorite object. This indicates an unambiguous trade-off between incentives and fairness in the design of exchange rules.
(Joint work with Peng Liu)
2PM GMT (9AM Montréal/Toronto, 11AM Rio de Janeiro, 2PM London, 3PM Graz, 5PM Istanbul, 7:30PM New Delhi, 11PM Tokyo/Seoul)
Steven Kivinen (University of Graz) "Robust Median Voter Rules"
Host: Marcus Pivato
Abstract. Generalized median voter (GMV) rules on the single-peaked preference domain are group strategy-proof. We show that if incomplete information coexists with the ability to commit to coalitional agreements, then GMV rules can be susceptible to insincere voting by groups with heterogeneous beliefs. We identify strategic compromise as a novel source of insincere voting in this environment. Our two main results characterize the set of fair, efficient, and robust voting rules: those that ensure sincere voting under asymmetric information and coalition formation. Each result uses a different notion of robustness, and both give (at most) two alternatives special treatment, with the remaining alternatives chosen according to a type of consensus.
(Joint work with Norovsambuu Tumennasan)
8PM GMT (12PM Vancouver, 2PM Austin, 3PM Toronto/Montréal, 5PM Rio de Janeiro, 8PM London, 9PM Paris, 11PM Istanbul, 9AM Wednesday in Auckland)
Harvey Lederman (University of Texas, Austin) "Maximal Social Welfare Relations on Infinite Populations Satisfying Permutation Invariance"
Host: Marcus Pivato
Abstract. We study social welfare relations (SWRs) on an infinite population. Our main result is a characterization of the common core shared by prominent utilitarian SWRs over distributions which realize finitely many welfare levels on this population. We characterize them as the largest SWR (with respect to set-inclusion when the weak relation is viewed as a set of pairs) which satisfies Strong Pareto, Permutation Invariance (elsewhere called ``Relative Anonymity'' and ``Isomorphism Invariance''), and a further ``Pointwise Independence'' axiom.
(Joint work with Jeremy Goodman)
2PM GMT (9AM College Park, 11AM Rio de Janeiro, 2PM London, 3PM Paris, 5PM Istanbul, 7:30PM New Delhi, 11PM Tokyo/Seoul)
Eric Pacuit (University of Maryland) "Characterizations of voting rules based on majority margins"
Host: Marcus Pivato
Abstract. In the context of voting with ranked ballots, an important class of voting rules is the class of margin-based rules (also called pairwise rules). A voting rule is margin-based if whenever two elections generate the same head-to-head margins of victory or loss between candidates, then the voting rule yields the same outcome in both elections. Although this is a mathematically natural invariance property to consider, whether it should be regarded as a normative axiom on voting rules is less clear. In this paper, we address this question for voting rules with any kind of output, whether a set of candidates, a ranking, a probability distribution, etc. We prove that a voting rule is margin-based if and only if it satisfies some axioms with clearer normative content. A key axiom is what we call Preferential Equality, stating that if two voters both rank a candidate x immediately above a candidate y, then either voter switching to rank y immediately above x will have the same effect on the election outcome as if the other voter made the switch, so each voter's preference for y over x is treated equally.
This is joint work with Yifeng Ding and Wes Holliday
5PM GMT (9AM Berkeley, 12PM Montréal/Toronto, 2PM Rio de Janeiro, 5PM London, 6PM Paris, 8PM Istanbul, 10:30PM New Delhi)
Snow Zhang (University of California, Berkeley) "Coherent combinations of experts' opinions"
Host: Marcus Pivato
Abstract. One popular rule for aggregating multiple experts' point estimates is linear averaging. However, this rule is incompatible with Bayesian conditionalization except in trivial cases of expert consensus (Dawid et al., 1995; Ranjan & Gneiting, 2010; Bradley, 2018; Gallow, 2018). This paper proves a generalization of this impossibility result. If time permits, we'll also discuss how to extend the analysis to the imprecise setting.
9AM GMT (6AM Rio de Janeiro, 9AM London, 10AM Paris, 12PM Istanbul, 2:30PM New Delhi, 6PM Tokyo/Seoul, 8PM Sydney, 10PM Auckland)
Vassili Vergopoulos (Université Paris II Panthéon-Assas) "Egalitarianism in Preference Aggregation under Uncertainty"
Host: Marcus Pivato
Abstract. This paper puts forward a theory of preference aggregation under uncertainty that incorporates egalitarian concerns both ex ante and ex post and accommodates heterogeneity in individual utilities and beliefs. Existing approaches often appeal to a compromise between an ex ante perspective committing to the Pareto Condition (PC) and an ex post perspective committing to Subjective Expected Utility (SEU). In contrast, we maintain each of PC and SEU in their full force on adequately restricted domains. Key to our approach is Choquet integration and its Fubini-like properties.
(Joint work with Federica Ceron)
5PM GMT (10AM Vancouver, 1PM Toronto/Montréal, 2PM Rio de Janeiro, 6PM London, 7PM Saint Etienne, 8PM Istanbul, 10:30PM New Delhi)
Antoinette Baujard (Université Jean Monnet Saint-Etienne) "How people understand novel voting rules: approval voting, evaluative voting and majority judgment"
Host: Marcus Pivato
Abstract. A primary condition of individuals' empowerment in their selection or use of a voting rule is that they understand it. Two papers analyze people's understanding of three voting rules: approval voting, evaluative voting and majority judgment. We draw on results from two data sets, first a lab experiment on incentivized voting where participants are exogenously assigned single-peaked preferences (first paper) and, second, a survey based on two representative samples of 1000 French voters, organized during the first round of the French presidential election (second paper). We distinguish three components of understanding of voting rules: how to fill in the ballot; how votes are aggregated; and how to vote strategically. We scrutinize each component by observing the respondents' voting behaviors and their answers to comprehension questions on the rules. First, from both data sets, we find that most participants understand how to fill in the ballot with the three novel voting rules. Second, in the lab experiment, participants' understanding of vote aggregation under majority judgment is lower and, crucially, more heterogeneous than with evaluative voting and approval voting. The fact that majority judgment is poorly understood is particularly striking in the representative survey. Third, participants' voting behavior is oddly similar between evaluative voting and majority judgment. Data from the lab experiment confirm the theoretical prediction that under evaluative voting there will be a high incidence of strategic voting through the use of extreme grades, but contradict the prediction that under majority judgment voters will vote less strategically. We also find that with majority judgment, the better voters understand how votes are aggregated, the more they use extreme grades. Fourth, the responses from the representative survey revealed genuine confusion among participants testing majority judgment, who conflated its vote aggregation process with the one used in evaluative voting. These results complement social choice theory by enabling an analysis of the properties of voting rules in practice.
(Joint work with Roberto Brunetti and Isabelle Lebon)
2PM GMT (10AM Toronto/Montréal, 11AM San Luis, 3PM London, 4PM Saint Etienne, 5PM Istanbul, 7:30PM New Delhi, 11PM Tokyo/Seoul)
Agustín Bonifacio (GATE Saint-Étienne and Instituto de Matemática Aplicada San Luis) "Obvious Manipulations by Groups"
Host: Marcus Pivato
Abstract. We introduce the notion of obvious manipulation by groups, extending obvious manipulability from individual to coalitional deviations. While existing work focuses on individual incentives, coordinated deviations are often natural, calling for a notion of obviousness at the group level.
We propose a definition based on simple best- and worst-case reasoning and study its implications for voting rules. In tops-only domains, ruling out obvious group manipulations imposes strong restrictions: it implies efficiency, monotonicity, and an almost-unanimity property.
Our main result shows that any tops-only, monotonic voting rule that is not obviously manipulable by groups must be dictatorial. This is somewhat surprising in light of the positive results for individual non-obvious manipulability, where rich classes of rules can be sustained. By contrast, robustness to obvious group deviations sharply limits the design of voting rules.
We also show that many standard rules satisfying the majority criterion are vulnerable to obvious group manipulations. Overall, our results provide a first step toward a theory of group obviousness and uncover new tensions between collective incentives and classical voting principles.
2PM GMT (10AM Toronto/Montréal, 11AM Rio de Janeiro, 3PM London, 4PM Paris, 5PM Istanbul, 7:30PM New Delhi, 11PM Tokyo/Seoul)
Kai Spiekermann (London School of Economics and Political Science) "What are Social Norms?"
Host: Marcus Pivato
Abstract. Many theorists tie social norms to attitudes, such as expectations towards others, perhaps along with conforming practices. Challenging this view, we instead ground social norms in a social norming process, an often non-verbal social communication process that ‘makes’ the norm through mutual expressions of support. We present the process-based account of social norms and social normativity, and distinguish social norms from social pressures, social practices, and Lewisian conventions. The process-based view brings social norms closer to legal norms, by tying them to ‘expressive acts’, just as laws and contracts arise through acts of voting or signing, not through mere attitudes.
(Joint work with Franz Dietrich)
2PM GMT (10AM Atlanta, 11AM Rio de Janeiro, 3PM London, 4PM Paris, 5PM Istanbul, 7:30PM New Delhi, 11PM Tokyo/Seoul)
Hun Chung (Emory University) "A Formal Theory of Robert Nozick's Framework for Utopia"
Host: Marcus Pivato
Abstract. This paper offers the very first formal model of Robert Nozick’s model of possible worlds and his vision of a utopian society, as outlined in Part III of Anarchy, State, and Utopia. Nozick envisioned utopia as a meta-utopia – a collection of self-organized, voluntary sub-communities – arguing that such an institutional framework is equivalent to the minimal state justified in earlier parts of his book. Nozick’s strategy was to define utopia (the best of all possible worlds) in terms of stability achieved in his possible worlds model, where individuals can create and migrate to any world they imagine. However, Nozick left many key components of this model informal and underdeveloped. This paper fills these gaps by providing a rigorous formal model of Nozick’s possible worlds. We introduce a new stability concept, Nozick stability. We demonstrate that Nozick stability imposes stricter requirements than other established solution concepts such as core and Nash stability, making the existence of a stable framework significantly more difficult to achieve. We then identify sufficient conditions for the existence of a Nozick-stable framework. However, these conditions are highly restrictive and unlikely to hold in reality. Furthermore, contrary to Nozick’s conjecture, individuals may receive far less than their marginal contribution within Nozick-stable frameworks, and, in this sense, Nozick-stable frameworks may institutionalize and perpetuate systemic exploitation. These findings cast doubt on whether Nozick’s minimal state can genuinely function as an inspiring utopian ideal, as he claims.
(Joint work with Susumu Cato)
5PM GMT (10AM Vancouver, 1PM Toronto/Montréal, 2PM Rio de Janeiro, 6PM London, 7PM Zürich, 8PM Istanbul, 10:30PM New Delhi)
Gustav Alexandrie (Universität Zürich) "Probability Aggregation Under Equal Expected Accuracy"
Host: Marcus Pivato
Abstract. A (binary) pooling function takes the probabilistic forecasts of two experts and returns an aggregated probability. We introduce the property of Equal-Expected-Accuracy (EEA) explainability, which is satisfied by pooling functions that treat the two experts as having equal expected accuracy under some scoring rule. EEA-explainable pooling functions provide a natural way of (1) formalizing the concept of epistemic peerhood from the philosophical literature on disagreement, and (2) designing scoring rules for paying forecasters fairly given the aggregated probability.
We prove two characterization theorems showing that EEA-explainable pooling functions can be expressed in terms of their associated scoring rules. These theorems imply that, under EEA-explainability, forecast aggregation among epistemic peers reduces to the problem of finding the scoring rules that best capture forecast accuracy in the relevant setting. We also establish a dual result showing how scoring rules can be expressed in terms of their associated pooling functions. This theorem yields a recipe for constructing scoring rules that pay forecasters fairly given the aggregated probability.
Finally, we explore the implications of EEA-explainability for the most commonly used pooling functions. In particular, we show that (a) linear pooling is uniquely EEA-explainable with respect to quadratic scoring rules, (b) geometric pooling is uniquely EEA-explainable with respect to a novel class of scoring rules based on the square root of the odds, and (c) multiplicative pooling is not EEA-explainable. Moreover, we show that EEA-explainability with respect to the standard logarithmic and spherical scoring rules yields novel pooling functions that have not been discussed in prior literature.
2PM GMT (10AM Toronto/Montréal, 11AM Rio de Janeiro, 3PM London, 4PM Paris, 5PM Istanbul, 7:30PM New Delhi, 11PM Tokyo/Seoul)
Maria Polukarov (King's College London) "Alternative Perspectives on Electoral Control"
Host: Lirong Xia
Abstract. This talk integrates two complementary perspectives on candidate control in voting. Traditionally, candidate control is studied as a harmful intervention: an external agent changes the candidate set to manipulate the winner, and the main question is whether such control is computationally easy or hard. We broaden this view by asking when candidate control can also be beneficial, not just for election outcomes but for participation and representation. In particular, we connect affirmative-action style interventions that can improve representativeness with models in which candidate entry changes who turns out to vote.
The first part of the talk introduces endogenous turnout, where each candidate attracts a subset of voters, so adding or deleting candidates affects both vote shares and electorate size. The second part studies a dynamic, spatial view of elections in which voters and candidates evolve over time, and candidate control can be evaluated not only by winner selection but also by its effect on substantive representation in fragmented electorates. Together, these perspectives show that candidate control is not a single phenomenon: depending on the setting, it may be a manipulation tool, a turnout mechanism, or a way to repair structural inequities. The common theme is understanding when procedural interventions harm collective decisions, when they help, and how the computational complexity of those choices shapes what is feasible in practice.
(password: *E%x0bLH)