Replication Data for: Electoral Systems, Partisan Politics, and Income Redistribution: A Critical Quasi-Experiment

Iversen and Soskice’s (2006) notion that electoral rules affect democracies’ propensity for income redistribution is one of the political economy’s most discussed concepts. Yet, it comes with a number of caveats. Most importantly, it is not clear whether electoral rules indeed affect states’ propensity for redistribution or vice versa and thus whether or not Iversen and Soskice’s findings are endogenous and spurious. In this article, we focus on the critical case of New Zealand’s electoral reform of the 1990s and offer a comprehensive test of Iversen and Soskice’s concept. We employ the recently developed dynamic multilevel latent factor model, a Bayesian alternative to synthetic controls (Pang et al. 2021), and compare the relevant dynamics for New Zealand to those of six majoritarian democracies. Our test largely supports Iversen and Soskice’s claims; due to the lower prevalence of right (center-right) governments, proportional representation democracies tend to redistribute more than majoritarian ones.

艾弗森与索斯凯斯（Iversen and Soskice，2006）提出的“选举规则影响民主国家收入再分配倾向”的理论，是政治经济学领域被学界广泛讨论的核心概念之一。然而该理论仍存在若干需审慎考量的限定条件：最为关键的是，目前尚无明确定论——究竟是选举规则切实影响一国的收入再分配倾向，还是二者因果方向恰好相反，进而无法判定艾弗森与索斯凯斯的研究结论是否存在内生性与伪回归问题。

本文以1990年代新西兰选举改革这一关键案例为研究对象，对艾弗森与索斯凯斯的理论展开全面检验。本文采用最新提出的动态多层潜因子模型（dynamic multilevel latent factor model）——一种可替代合成控制法（synthetic controls）的贝叶斯分析框架（Pang等人，2021），并将新西兰的相关动态变化与六个多数制民主国家进行对比分析。

本次检验在很大程度上支持了艾弗森与索斯凯斯的理论主张：由于右翼（中右翼）政府执政的概率更低，比例代表制民主国家的收入再分配规模往往高于多数制民主国家。