Preferences over Relative Income within the Household
Project period: April 2022 - June 2023
Research Areas:
Tasks
What preferences do partners hold over relative income within the household? This paper takes relative income concerns to the household by investigating the existence and form of relative income preferences between partners. The motives for relative income concerns within the household are ambiguous. Partners may prefer to earn a similar amount, implying inequality aversion. Alternatively, they may reflect a preference for a male primary earner to adhere to a male breadwinner norm. Understanding preferences over spouses’ relative income has important theoretical implications for household models that commonly neglect relative income concerns, but also empirical implications as it can help explain (income) decision-making of spouses. This is particularly interesting as spouses’ income and labour patterns often follow a traditional role division where the female takes over a larger share of the household work and earns less than her husband. Relative income preferences potentially explain these patterns.
Methods
The paper first develops a theoretical framework of preferences over relative income that captures various motives. In this model, spouses gain utility from relative income concerns. We introduce these non-standard preferences in a marriage market matching model with search frictions to obtain empirically testable predictions for the selection and separation of couples.
Using the main model prediction, we test for the presence and structural form of preferences over relative income by analyzing the relative income distribution of German spouses. Further, we design and implement a survey experiment in the United States and Germany that allows us to identify and quantify women’s and men’s preferences separately.
Data and other sources
The main data source is German administrative tax data: A 10% sample of the universe of wage and income tax returns (Lohn- und Einkommenssteuerdaten) from 2001, 2004, 2007, 2010 and 2014. The data is combined with SOEP data and the results from our Survey.