Diederik Boertien, Centre d’Estudis Demogràfics (CED)
Milan Bouchet-Valat , French Institute for Demographic Studies (INED)
The last few years have seen a dramatic increase in research addressing the question to what extent the association between partners’ earnings matters for inequality between couples. Studies come to an apparent variety of results, ranging from conclusions that earnings similarity barely impacts inequality, to findings that changes in earnings similarity have been responsible for considerable increases in inequality. In this paper, we aim to reconcile these earlier findings through a detailed literature review as well as a comparative empirical analysis. We first show that studies on the topic answer three similar yet distinct questions: How high would inequality be if people partnered at random? Did changes in earnings similarity over time, including changes in employment rates, contribute to inequality? Did changes in the association between partners’ earnings, net of general changes in employment rates, contribute to inequality? Previous research comes to relatively consistent answers once divided according to these three questions. We subsequently use data from the Luxembourg Income Studies on 21 countries from 1974 to 2016 to focus on the last two questions. Using decompositions based on log-linear models, we show that even though the correlation in earnings between partners increased in most countries, this only amplified inequality in a subset of these countries. In other countries, increases in the earnings correlation are primarily driven by general changes in employment rates. Given that these increases in employment equalized earnings across households, the inherently connected increases in the earnings correlation are of less concern from an inequality perspective.
Presented in Session 105. (Re)Production of Inequalities