Robust Governance and Centralization
Speaker
Yotam Gafni
Date
03/11/2026 - 12:30 - 11:15Add to Calendar
2026-11-03 11:15:00
2026-11-03 12:30:00
Robust Governance and Centralization
We consider systems of voting, where the rules of voting themselves are subject to change. If voters deem a voting rule as misaligned with their goals, they vote to change it to a rule that better serves them. We seek stable voting rules that are robust to such changes. To study such questions, one must consider many different aspects: Do voters tie-break in favor of status-quo or of change? Are we satisfied with rules that are stable for some belief (rationalizable) or that are stable for any belief (robust)? Which domain of voting rules to consider: Weighted Majority Games (WMGs), where every voter is assigned a weight and the sum of supportive votes must pass a certain threshold, or a more general class of voting rules? We develop a framework that answers the question across all the above modeling decisions, and find that robust rules are highly centralized.To alleviate the prohibitive centralization results, we consider three mitigations: (i) forward-looking voters, who can plan to avoid the authoritarian trap, (ii) bi-level constitutions, (iii) A model with "Condorcet Jury Theorem" effects, which incorporate the fact that diverse voter participation matters for the epistemic and operational success of the system.We demonstrate how each of the mitigations indeed helps, to some extent, to circumvent centralization.Link to the paper: https://arxiv.org/abs/2512.22051 Yotam Gafni's homepage: https://www.yotamgafni.com
BIU Economics common room
אוניברסיטת בר-אילן - Department of Economics
Economics.Dept@mail.biu.ac.il
Asia/Jerusalem
public
Place
BIU Economics common room
Affiliation
Weizmann Institute
Abstract
We consider systems of voting, where the rules of voting themselves are subject to change. If voters deem a voting rule as misaligned with their goals, they vote to change it to a rule that better serves them. We seek stable voting rules that are robust to such changes. To study such questions, one must consider many different aspects: Do voters tie-break in favor of status-quo or of change? Are we satisfied with rules that are stable for some belief (rationalizable) or that are stable for any belief (robust)? Which domain of voting rules to consider: Weighted Majority Games (WMGs), where every voter is assigned a weight and the sum of supportive votes must pass a certain threshold, or a more general class of voting rules?
We develop a framework that answers the question across all the above modeling decisions, and find that robust rules are highly centralized.
To alleviate the prohibitive centralization results, we consider three mitigations: (i) forward-looking voters, who can plan to avoid the authoritarian trap,
To alleviate the prohibitive centralization results, we consider three mitigations: (i) forward-looking voters, who can plan to avoid the authoritarian trap,
(ii) bi-level constitutions,
(iii) A model with "Condorcet Jury Theorem" effects, which incorporate the fact that diverse voter participation matters for the epistemic and operational success of the system.
We demonstrate how each of the mitigations indeed helps, to some extent, to circumvent centralization.
Link to the paper: https://arxiv.org/abs/2512.22051
We demonstrate how each of the mitigations indeed helps, to some extent, to circumvent centralization.
Link to the paper: https://arxiv.org/abs/2512.22051
Yotam Gafni's homepage: https://www.yotamgafni.com
Last Updated Date : 29/10/2026
