Does proportional representation (PR) lead to broader public goods spending? Past literature has overwhelming shown that it does, but the empirics underlying these findings have mostly relied on ethnically homogenous Western countries. Categorizing countries along three dimensions of ethnic structure—ethnic fractionalization, ethno-income crosscuttingness and ethno-geographic dispersion—I argue that in some types of societies, PR has the positive effect on fiscal spending type predicted by past models, but not in others. Specifically, in countries with high ethnic salience (ethnically heterogeneous, low crosscutting) where ethnic groups are geographically intermixed, PR leads to narrower fiscal spending; in high ethnic salience societies where ethnic groups are geographically isolated, neither PR nor majoritarian electoral rules lead to broader fiscal spending. I test this socio-institutional theory in a sample of 70 developing democracies using life expectancy and illiteracy as proxies for public goods provision.

比例代表制（PR）是否会导致更广泛的公共物品支出？过去的文献不计其数地证明了这一点，但是这些发现背后的实证研究大多依赖于种族同源的西方国家。按照种族结构的三个维度对国家进行分类：种族划分，种族收入交叉和种族地理分散性。我认为，在某些类型的社会中，公关对过去模型所预测的财政支出类型具有积极影响，而在其他模型中则没有。具体而言，在种族显着性高（种族异质性，跨领域性低）的国家中，种族在地理上是混杂的，PR导致财政支出收窄。在种族群体在地理上相互隔离的高级种族显着社会中，公关或多数派选举规则都不会导致更广泛的财政支出。我在70个发展中民主国家的样本中测试了这种社会制度理论，这些样本使用预期寿命和文盲作为提供公共物品的代理。