This theoretical paper develops a multi-district model that endogenizes voter pivotality to link redistricting, candidate selection, and the equilibrium allocation of group-targeted transfers; it shows two channels — a competition channel (how district composition alters targeting incentives) and a selection channel (how it alters who is nominated/elected) — whose curvature properties determine whether packing or dispersion best serves minority welfare, and demonstrates that endogenous feedback in competitive (tipping) districts can produce nonmonotonic welfare in minority concentration .
速读内容
Core thesis: two channels and endogenous pivotality
Districting affects minority welfare through (i) a competition channel (changes in marginal responsiveness and targeting incentives) and (ii) a selection channel (changes in candidate matchups and who wins). The interaction and endogenous feedback can align, diverge, or overturn benchmark predictions about packing vs. cracking .
With the general-election matchup held fixed, minority distributive share b_m is proportional to π_m·N_m,k divided by aggregate power Π_k, so increasing minority voters in a district always raises their local share, but curvature matters: if π_m is lowest, T_{mD,k} is convex (concentrate/minority-packing optimal); if π_m is highest, T_{mD,k} is concave (dispersion optimal). Thus packing helps when minorities have low distributive leverage; spreading helps when they are pivotal across districts .
Increasing minority share weakly raises the probability a minority candidate wins the seat; convexity of winning probability implies packing boosts descriptive representation. Convexity depends on primary type and crossover/support parameters (thresholds β^O and β^C determine when concentration increases ideological welfare) .
Under open primaries, replacing Republicans by nonminority Democrats generally increases minority nominee success; under closed primaries the effect of adding nonminority Democrats is ambiguous due to primary vs. general-election trade-offs .
Interaction and general-equilibrium feedback (tipping vs safe districts)
Full welfare W(D) = selection term + competition term + interaction I_k; I_k captures reweighting of outcome probabilities and endogenous changes in pivotality via matchup shifts. In safe districts I_k is second-order and isolated-channel predictions hold; in tipping districts I_k can be first-order and can reinforce or overturn the fixed-matchup benchmark, producing nonmonotonic welfare in concentration .
Policy takeaways and testable predictions
Whether majority–minority districts increase minority welfare depends on: minority electoral leverage (π_i, κ_i), voters’ weight on descriptive vs partisan representation (β), primary rules (open vs closed), and electoral competitiveness (whether districts are tipping). The model predicts systematic variation in packing/cracking welfare by competitiveness, crossover voting, and responsiveness .
Example illustrations and constructive results
Geometric characterizations on the simplex show: (i) optimal district plans typically have at most one interior district and pair minority-rich districts with the less powerful nonminority group; (ii) concave/convex surfaces explain when marginal returns to concentration rise or fall .
思维导图
深度阅读
以下为对所提供研究报告“Distributive Politics, Representation, and Redistricting” 的逐字剖析与综合评估,按照开发者指示分节详尽解读,并对文中所有重要论点、数据表与图表进行逐一说明与批判性评价。为便于溯源,我在引用或从原文推断的句子末尾均标注了页码标识 [page::页码](同一处引用不超过两页)。
一、元数据与概览(引言与报告概览)
标题与作者:论文题为 "Distributive Politics, Representation, and Redistricting",作者为 Thomas Groll(Columbia University)与 Sharyn O’Halloran(Columbia University / Trinity College Dublin)。
发布日期与主题:提交日期为 March 31, 2026,研究主题为:如何在有初选与两阶段选举(primary + general)的多选区体系下,将配给政治(distributive politics)与选区划分(redistricting)结合起来,通过内生化“选民边际决定性(voter pivotality)”来解释选举结果与政策利益在群体间的分配。作者以少数族裔(minority)代表性为主要应用情境,但方法具有更广泛的制度设计适用性。
逻辑与假设:引入初选与两阶段选举结构,主张选区构成不仅改变谁有可能胜出(描述性),还改变在不同候选人配对下候选人為不同群体提供轉移的动力(實質性)。该观点依赖于:候选人以争取票数最大化为目标、选民对候选人既有描述性偏好也对实质转移有消费效用、以及转移的边际效用(marginal utility)影响其可被“购买”的廉价度。作者明确把Voting Rights Act的制度逻辑作为研究背景之一。
(2)模型设定(Section 2)
人物设定与分组:总人口 N 被划为三类(mD = minority Democrats, nD = nonminority Democrats, R = Republicans),全州被划为 K 个等人口限制的选区,每一选区以人数向量 N_{i,k} 表示其组分布。作者用三角单纯形(simplex S^2)表示任一区选区成分并以图示说明可行位置与样例方案。
核心结果:在固定普选配对(例如 mD vs R 或 nD vs R)下,少数族裔在某一选区中分得的群体总转移比例可写为 f(N_{mD,k}, N_{nD,k}) = π_{mD} N_{mD,k} / Π_k(Π_k 为加权总群体力量),并给出对 π 与 N 的偏导结构说明:增加少数人数总是提高少数分配份额,而增加其他组的力量(π)会降低其份额。作者证明在最优化方案中,至多只有一个选区会落在三角形内部(interior),其余要么边界化,且少数密集区倾向与较弱的非少数群体配对(即“与较不强势的非少数配对”以提升少数得益)。
曲率判据(Lemma 2、图3):表面 f 在单纯形内部的曲率由 π 排位决定:若 π_{mD} 为最大,曲面为凹(concave),即边际收益递减;若 π_{mD} 为最小,则为凸(convex),即边际收益递增。基于该几何性质,作者给出在竞争通道下的分配策略:当少数“政治杠杆”最低时应集中(packing);当杠杆最高时应分散(cracking)。
(5)选择通道(Section 5)——固定配给下的描述性代表性
核心逻辑:在只考虑谁会被提名与当选(不考虑竞选时的承诺调整)情况下,增加少数在区内比例 s_k 一般会弱化或增强少数候选人获胜概率,具体符号由初选制度(closed vs open)、非少数选民的构成(t_k)与候选人相对的跨越性支持(crossover)决定。Proposition 3 给出偏导与二阶偏导的分析,揭示在 open primary 下若把 R 替换成 nD 会弱化或提升少数候选胜率(通常为弱非负),而 closed primary 情形下增加 nD 对少数候选的影响既有来自 primary 的负面效应又有来自 general 的可能正效应,因此效果是模糊的。作者推导出少数当选概率在 s_k 上往往是凸的(packing 导致收益递增),但凸性阈值受参数(尤其是跨越投票强度与候选人可竞争性 β)决定。该部分同时给出预期意识形态效用与 β(少数选民对非少数民主党候选人之相对偏好权重)阈值的解析字符化(Lemma 3、Proposition 4)。