Voter Priming Campaigns: Strategies, Equilibria, and Algorithms
Jonathan Shaki, Yonatan Aumann, Sarit Kraus · 2024
Read Voter Priming Campaigns: Strategies, Equilibria, and Algorithms by Jonathan Shaki, Yonatan Aumann, Sarit Kraus (2024) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2412-13380.txt.
Introduction
Jonathan Shaki, Yonatan Aumann, and Sarit Kraus model how parties spend to make their best issues loom large in voters' minds, in a 2024 paper. Voters weigh issues by salience — how top-of-mind each issue feels — and campaigns can shift that ranking with money. This steering is called priming.
The team study priming contests with many issues and many parties. In parliamentary races, where parties chase vote share, a stable spending balance always exists and can be computed fast. With two parties, each side in balance pours everything into a single issue. In most winner-take-all presidential settings, by contrast, no stable balance exists at all. The results map when campaign spending wars settle and when they spiral.
Catalog record
| CREATOR | Jonathan Shaki, Yonatan Aumann, Sarit Kraus |
|---|---|
| YEAR | 2024 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 334 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2412-13380
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Voter Priming Campaigns: Strategies, Equilibria, and Algorithms Jonathan Shaki, Yonatan Aumann, Sarit Kraus Abstract Issue salience is a major determinant in voters’ decisions. Candidates and political parties campaign to shift salience to their advantage - a process termed priming . We study the dynamics, strategies and equilibria of campaign spending for voter priming in multi-issue multi-party settings. We consider both parliamentary elections, where parties aim to maximize their share of votes, and various settings for presidential elections, where the winner takes all. For parliamentary elections, we show that pure equilibrium spending always exists and can be computed in time linear in the number of voters. For two parties and all settings, a spending equilibrium exists such that each party invests only in a single issue, and an equilibrium can be computed in time that is polynomial in the number of issues and linear in the number of voters. We also show that in most presidential settings no equilibrium exists. Additional properties of optimal campaign strategies are also studied. 1 Bar-Ilan University [email protected], [email protected], [email protected] 1 Introduction Political parties and candidates invest substantial resources in campaigns aimed at winning over voters. Interestingly, research reveals that these campaigns often do not directly alter voters’ views or attitudes toward candidates and issues. Rather, campaigns tend to influence the relative salience of various topics, thereby shaping the importance voters assign to these issues ( Moniz and Wlezien 2020 ; Macdonald 2023 ; Druckman 2004 ) . Consequently, an essential element in political campaigns involves strategically guiding the salience of issues to the candidates’ advantage, a process known in political science as priming ( Bartels 2006 ) . Political campaigns are costly, and candidates have limited budgets. Thus, from a strategic perspective, the question is how to best invest the available campaign budget in order to obtain the most beneficial priming for a given candidate, and whether an equilibrium investment profile exists. From a computational perspective, the question is how to compute the optimal and equilibrium investments, if they exist. These are the topics of this paper. The importance of issue salience in voting has recently gained attention in the context of election control and manipulation ( Lu et al. 2019 ; Estornell et al. 2020 ) . In these works, priming is viewed in the context of election manipulation and hence assumed to be limited to a single malicious actor. We view priming in the context of campaigning and study equilibria thereof. 1.1 Summary of Results We consider two core settings, which we term parliamentary and presidential . In the parliamentary setting, parties aim to maximize their share of votes. In the presidential case, the primary goal is to be ranked first. For the presidential setting, we consider several variants, which relate to a possible secondary goal of increasing the share of votes. 1 1 1 without such a secondary goal, unnatural behavior can emerge wherein candidates invest against their own good. The different variants, formally defined in detail in Section 2 , are: (i) u_{ind} : the sole goal is to be ranked first, with no secondary goal, (ii) u_{plus} : a weighted average the goal of being ranked first and increasing the share of votes, with the weight highly skewed towards the former, (iii) u_{max} : the primary goal is to be ranked first, only if this cannot be achieved then increasing the share is considered as a goal, with significantly lesser utility. Throughout, we consider plurality ballot form, wherein each voter casts one vote for its most favorable candidate/party ( Meir et al. 2010 ) . For each setting and variant, we consider the question of whether a Nash equilibrium necessarily exists and whether, knowing the strategies of the other players, a best-response necessarily exists. Our main results are summarised in Table 1 . For the parliamentary setting, both a Nash equilibrium (in pure strategies) and best-response always exist. When there are only two candidates, it is further the case that in equilibrium each candidate invests its entire budget in priming of a single issue. For the presidential setting, for both variants that include a secondary goal, a Nash equilibrium need not exist in the general case, and not even a best-response. For the variant where the sole goal is winning, with no other considerations, a best response does exists, and we do not know if a Nash equilibrium necessarily exists. We note, however, that the Nash equilibria in this variant may be unnatural, in the sense that they require candidates to invest not for their own good, but rather for the benefit of others. In all presidential variants, if there are only two candidates, an equilibrium exists wherein each candidate invests its entire budget in priming of a single issue. Whenever a Nash equilibrium is guaranteed to exist, we provide an algorithm that is linear in the number of voters to compute it. For the two candidates case, the algorithms are also polynomial in the number of issues and candidates. For the general parliamentarian case, the algorithm may be exponential in the number of issues and candidates. Due to limited space, the full proofs appear in appendix. In the main body we mostly provide limited proofs and outlines. setting variant best response Nash equilibrium Nash equilibrium 2 candidates parliamentary u_{frac} ✓exists ( Theorem 2 ) ✓exists (in pure strategies) ( Theorem 4 ) ✓exists (with pure strategies, single issue investments) ( Theorem 5 ) presidential u_{plus} ✗ need not exist ( Theorem 9 ) ✗ need not exist ( Corollary 2 ) ✓exists (with pure strategies, single issue investments) ( Theorem 8 ) u_{max} ✗ need not exist ( Theorem 11 ) ✗ need not exist ( Theorem 12 ) ✓exists (with pure strategies, single issue investments) ( Section 4.3 ) u_{ind} ✓exists ( Theorem 7 ) \bm{?} open ✓exists (with pure strategies, single issue investments) ( Theorem 6 ) Table 1: Summary of results. Whenever Nash equilibrium is guaranteed to exist, it can be computed in time linear in the number of voters. For two candidates, the time is also polynomial in the the number of issues. 1.2 Related Work Campaign management and bribery have been researched extensively in the literature (see ( Schlotter, Faliszewski, and Elkind 2011 ; Faliszewski et al. 2015 ; Knop, Kouteckỳ, and Mnich 2020 ; Zagoury et al. 2021 ) . And for extensive review see ( Simpser 2013 ; Islam, Mohajan, and Moolio 2010 ) ). Both these terms refer to a setting where individual voters are influenced to alter their vote, and the main question is how to choose these voters. We consider situations where changing voters’ decision-making is only possible by general campaign messages, which alter the importance of different issues. Ad delivery algorithms also aim to change voters’ actions. They promise to lower the cost of advertising and increase the efficiency of campaigns through detailed targeting, where advertisers can specify the users they would like to reach using attributes ( Danaher 2023 ; Kreiss 2011 ) . However, it was shown (e.g., ( Ali et al. 2021 ; Cotter et al. 2021 ) ) that targeting may not work as intended, and may shape the political ad delivery in ways