Underemployment and mental health: comparing fixed-effects and random-effects regression approaches in an Australian working population cohort

Objectives Underemployment occurs when workers are available for more hours of work than offered. It is a serious problem in many Organisation for Economic Co-operation and Development (OECD) countries, and particularly in Australia, where it affects about 8% of the employed population. This paper seeks to answer the question: does an increase in underemployment have an influence on mental health? Methods The current paper uses data from an Australian cohort of working people (2001-2013) to investigate both within-person and between-person differences in mental health associated with being underemployed compared with being fully employed. The main exposure was underemployment (not underemployed, underemployed 1-5, 6-10, 11-20 and over 21 hours), and the outcome was the five-item Mental Health Inventory. Results Results suggest that stepwise declines in mental health are associated with an increasing number of hours underemployed. Results were stronger in the random-effects (11-20 hours =-1.53, 95% CI -2.03 to -1.03, p<0.001; 21 hours and over -2.24, 95% CI -3.06 to -1.43, p<0.001) than fixed-effects models (11-20 hours =-1.11, 95% CI -1.63 to -0.58, p<0.001; 21 hours and over -1.19, 95% CI -2.06 to -0.32, p=0.008). This likely reflects the fact that certain workers were more likely to suffer the negative effects of underemployment than others (eg, women, younger workers, workers in lower-skilled jobs and who were casually employed). Conclusions We suggest underemployment to be a target of future workplace prevention strategies.