Justification bias in self-reported disability: new evidence from panel data

Abstract: Understanding the relationship between health and work is central to labor and health economics research, and crucial for the design of health policies, social welfare systems, and strategies for productivity and growth. This relationship is often investigated using self-assessments of health and disability from social surveys; however, there exists a legitimate concern that individuals without a job may overstate their disability or health problems in order to justify their non-employment status. Such justification implies that estimated population rates of disability and the estimated effects of poor health and disability on labor force withdrawal may be severely biased. In this study we exploit a unique feature of an Australian longitudinal survey to: provide new estimates of the magnitude of justification bias; characterise the types of individuals for whom justification bias is most common; and identify aspects of survey design that exacerbate justification bias. Our individual-level fixed-effects estimates suggest that justification bias is largest for men who are detached from the labor market and for men who receive the Disability Support Pension. We also find significant heterogeneity in the magnitude of justification by gender, marital status, educational attainment and attitudes towards work. Results from robustness and placebo tests support our interpretations.