Volume 48 - Article 26 | Pages 775–808

A test of the predictive validity of relative versus absolute income for self-reported health and well-being in the United States

By David Brady, Michaela Curran, Richard Carpiano

Print this page  Facebook  Twitter


Date received:24 Jun 2022
Date published:16 May 2023
Word count:4526
Keywords:absolute income, health, inequality, measurement, relative income, well-being


Background: A classic debate concerns whether absolute or relative income is more salient. Absolute values resources as constant across time and place while relative contextualizes one’s hierarchical location in the distribution of a time and place.

Objective: This study investigates specifically whether absolute income or relative income matters more for health and well-being.

Methods: We exploit within-person, within-age, and within-time variation with higher-quality income measures and multiple health and well-being outcomes in the United States. Using the Panel Study of Income Dynamics and the Cross-National Equivalent File, we estimate three-way fixed effects models of self-rated health, poor health, psychological distress, and life satisfaction.

Results: For all four outcomes, relative income has much larger standardized coefficients than absolute income. Robustly, the confidence intervals for relative income do not overlap with zero. By contrast, absolute income mostly has confidence intervals that overlap with zero, and its coefficient is occasionally signed in the wrong direction. A variety of robustness checks support these results.

Conclusions: Relative income has far greater predictive validity than absolute income for self-reported health and well-being.

Contribution: Compared to earlier studies, this study provides a more rigorous comparison and test of the predictive validity of absolute and relative income that is uniquely conducted with data on the United States. This informs debates on income measurement, the sources of health and well-being, and inequalities generally. Plausibly, these results can guide any analysis that includes income in models.

Author's Affiliation

David Brady - University of California, Riverside, United States of America [Email]
Michaela Curran - University of Iowa, United States of America [Email]
Richard Carpiano - University of California, Riverside, United States of America [Email]

Most recent similar articles in Demographic Research

» A behaviorally-based approach to measuring inequality
Volume 19 - Article 49    | Keywords: inequality, measurement

» Introduction to the Special Collection on The new roles of women and men and implications for families and societies
Volume 48 - Article 29    | Keywords: well-being

» Gender and educational inequalities in disability-free life expectancy among older adults living in Italian regions
Volume 47 - Article 29    | Keywords: health

» Female sterilization in the life course: Understanding trends and differentials in early sterilization
Volume 47 - Article 18    | Keywords: inequality

» Measuring contraceptive use in India: Implications of recent fieldwork design and implementation of the National Family Health Survey
Volume 47 - Article 4    | Keywords: measurement