The effect of mental health on early retirement decisions: Evidence from Australia

Abstract

Health and labour supply are interconnected; However, research has predominantly
focused on the impact of physical health, leaving a gap in understanding the role of
mental health problems. This study addresses this gap by examining the effect of
mental health on early retirement decisions using data from the Household, Income,
and Labour Dynamics in Australia (HILDA) Survey. We use both linear probability models
and a discrete-time hazard approach. While linear models estimate the average effect,
the discrete-time hazard model tracks initially employed individuals aged 50 to 64 over
time until they retire early or reach retirement age. To mitigate potential bias arising from
the timing of reporting of mental health and retirement decisions, lagged measures of
mental health are applied, with respect to the temporal sequence of events. To address
measurement bias, the association between our derived mental health variable and other
objective psychiatric measures is examined. Furthermore, we include the death of a close
friend as an instrument for mental health status, helping us validate and strengthen
causal findings of our study. Lastly, we examine whether unobserved heterogeneity
poses a problem in our analysis by estimating models with and without unobserved
heterogeneity. Our findings indicate a significant and positive causal impact of poor
mental health on early retirement decisions, which is also supported by the nonlinear
analysis. To explore potential gender heterogeneity, separate analyses are conducted
for males and females. The observed differences in the results between the two groups
support the assumption of gender-specific effects. These findings suggest that poor
mental health has a significant and potentially causal impact on premature exit from the
labour market, particularly among men. The results highlight the importance of effective
mental health management in supporting longer working lives.