This paper provides new evidence on educational assortative mating (PAM) and household income inequality in Brazil, Indonesia, Mexico, and South Africa, and introduces a methodological approach that isolates pure sorting from educational expansion . PAM exists across all education levels, with the strongest sorting among the highly educated; after controlling for shifts in education, overall PAM declines over time, not primarily driven by expansion . The observed minimal impact of changes in sorting on inequality arises because decreases among the highly educated offset increases among the least educated, yielding only small changes in the aggregate inequality . While the findings largely reassure about the effect of sorting on inequality, they signal rising marginalization at the bottom of the educational distribution, underscoring the need to address opportunities for the least educated .
速读内容
Educational PAM is present across all education levels
PAM observed in Brazil, Indonesia, Mexico, and South Africa; strongest among the highly educated.
Overall PAM declines after controlling for education expansion
Controlling marginals reveals a downward trend in PAM across all four countries.
Changes in sorting have limited impact on inequality
Counterfactuals show small differences (1–3%) in Gini if sorting degree differed over time.
Group dynamics explain the small inequality impact
High-education sorting decreases offset increases from low-education sorting; aggregate effect is muted.
Policy implications: attention to marginalization at the bottom
While PAM may not drastically widen household inequality, the least educated face growing isolation and limited mobility.
思维导图
深度阅读
以下为对“Educational Assortative Mating and Household Income Inequality: Evidence from Brazil, Indonesia, Mexico, and South Africa”这篇工作论文的极为详细解读与剖析。分析遵循原文结构,逐节梳理核心论点、数据、方法、图表与结论,并结合图表/表格的关键点进行逐项解读与溯源标注。文末的溯源均以原文页码呈现,便于追溯。
一、元数据与概览(引言与研究概览)
标题、作者与机构与时间(主题概要)
标题:Educational Assortative Mating and Household Income Inequality: Evidence from Brazil, Indonesia, Mexico, and South Africa。作者为 Ana Kujundzic,来自欧洲体系外部研究环境的研究工作,发表于 Eesti Pank 的 Working Papers 系列(2026 年首发,No 1)。
4.3 Comparison with two methodologically related studies(与两项相关研究的对比)
论文指出教育排序趋势在不同研究中存在显著差异,部分源于边际依赖性问题及排序测量的比较方式差异。 Greenwood et al.(2014, 2015)与 Eika et al.(2019)的结论在方法上具有可比性,但在趋势解释上存在分歧;修正版本的结果与本文一致:排序对不平等的影响存在但随时间变化的影响有限。