Working Paper: NBER ID: w22817
Authors: Margherita Borella; Mariacristina De Nardi; Fang Yang
Abstract: Wages, labor market participation, hours worked, and savings differ by gender and marital status. In addition, women and married people make up for a large fraction of the population and of labor market participants, total hours worked, and total earnings. For the most part, macroeconomists have been ignoring women and marriage in setting up structural models and by calibrating them using data on males only. In this paper we ask whether ignoring gender and marriage in both models and data implies that the resulting calibration matches well the key economic aggregates. We find that it does not and we ask whether there are other calibration strategies or relatively simple models of marriage that can improve the fit of the model to aggregate data.
Keywords: gender; marriage; labor market; economic aggregates
JEL Codes: D1; E1; E21
Edges that are evidenced by causal inference methods are in orange, and the rest are in light blue.
| Cause | Effect |
|---|---|
| One-gender model using only male data (C29) | Significant overestimations in labor participation, hours worked, and earnings (J39) |
| Combined male and female data (C39) | Missed observed patterns of participation and hours (J22) |
| Household-level calibration for couples (D19) | Improved fit for labor earnings but missed participation accurately (J49) |
| Modeling both married and single individuals (C20) | Better match to actual economic aggregates (E19) |
| Ignoring gender and marriage (J12) | Significant discrepancies in model outputs (C52) |