Hogyan szálazzuk szét a megfigyelhető változások okait? [Decomposing observable changes into their causes]

This paper presents a decomposition scheme that can be applied to decompose inter-temporal changes into the ceteris paribus effects of certain factors and their interactions. This decomposition scheme was proposed by Biewen (2014). In his view, the significance of the scheme is the following: while the interaction effects can be quantified using his approach, these effects are often disregarded in certain strands of the empirical decomposition literature. For instance, decomposition schemes with no interaction terms are commonly applied in the literature analyzing marital matching at the aggregate level. This fact motivates analyzing a problem in the assortative mating literature to illustrate the application of the Biewen's formula in this paper. I perform the analysis on data on the final educational attainments of couples in five countries (France, Hungary, Portugal, Romania, and the US) using census data. This analysis makes it possible to quantitatively characterize a phenomenon that cannot be directly observed, namely how educational homophily (i.e., aggregate preferences for educationally homogamous matches) has changed from one generation to another in the 1980s, 1990s, and 2000s. The outcome of the decompositions in this paper confirm the findings in Naszódi and Mendonça (2021) obtained with a single-country analysis: after the Second World War, there was a period when homophily weakened, i.e., marriages between people with different education levels became more accepted socially. This period was followed by another period characterized by the opposite trend. These findings can only be obtained in a robust form with a carefully selected decomposition formula.

NASZODI Anna; 

2025-01-20

Economic Review Foundation

JRC140856

http://dx.doi.org/10.18414/KSZ.2022.11.1407,    https://publications.jrc.ec.europa.eu/repository/handle/JRC140856,   

10.18414/KSZ.2022.11.1407 (online),