Is temporary employment a cause or consequence of poor mental health? A panel data analysis

Highlights

• Permanent workers with poor mental health select into temporary employment.

• This selection effect is mediated by greater job dissatisfaction.

• Cross sectional studies may overestimate the impact of employment type on health.

• Panel data fixed effects models are also likely to give biased estimates.

Abstract

Mental health status has an association with labour market outcomes. If people in temporary employment have poorer mental health than those in permanent employment then it is consistent with two mutually inclusive possibilities: temporary employment generates adverse mental health effects and/or individuals with poorer mental health select into temporary from permanent employment. We apply regression analyses to longitudinal data corresponding to about 50,000 observations across 8000 individuals between 1991 and 2008 drawn from the British Household Panel Survey. We find that permanent employees who will be in temporary employment in the future have poorer mental health than those who never become temporarily employed. We also reveal that this relationship is mediated by greater job dissatisfaction. Overall, these results suggest that permanent workers with poor mental health appear to select into temporary employment thus signalling that prior cross section studies may overestimate the influence of employment type on mental health.

Introduction

Health and labour market status are intrinsically linked. Analyses of these links adopt two distinct perspectives: first, health impacts on employment and, second, employment impacts on health. Health status can be separated into two mutually inclusive parts: physical and mental health conditions. Although the exact proportions are controversial, the Mental Health Foundation (2014) argues that a quarter of people will experience a mental health condition at some point in their lives and around one in twelve people are affected by depression. This study assesses the relationship between mental health conditions and labour market transitions between permanent and temporary employment.

Although there are an increasing number of studies that focus on the link between health and employment, such as Pirani and Salvini (2015), dominant explanations of the impacts of health on employment typically focus on health as a medically classified condition (Oliver, 1990) and emphasise the effects of clinical factors on an individual's employment capabilities. When an individual is in employment but has a mental health condition they are known to be at risk of experiencing presenteeism, which is where an employee is unwell and remains in work but is less productive. Presenteeism can occur when people with poor mental health lack obvious outward signs and are reluctant to have to prove they are ill because of the resulting stigma (Lelliott et al., 2008). Mental health stigma includes the perception that individuals with mental health disorders are weak, flawed, dangerous and/or socially incompetent (Wahl, 2003) and the desire not to want to be thought of as having these characteristics can deter people from seeking or obtaining help (Hinshaw and Cicchetti, 2000). Chen et al. (2015) argue that rates of presenteeism vary with the perceived level of workplace support, with those feeling least supported having higher rates of presenteeism. Individuals with poor mental health are also known to be less likely to be in employment: in the UK in 2004, 74 percent of the working age population was in employment but the comparable figure for people considered disabled by a long term mental illness was only 21 percent (Social Exclusion Task Force, 2006).

A distinctly different literature emphasises the existence of the reverse association, i.e. that lower labour market status affects health. For instance, Silla et al. (2005) find that temporary workers experience relatively poor health outcomes and Martens et al. (1999) find that employees on temporary contracts, working irregular hours or working compressed working weeks report up to 40 percent more health complaints than those with non-flexible work schedules. However, Bardasi and Francesconi (2004) find no evidence that atypical employment is associated with adverse health consequences.

Hence the literature is divided on whether poor mental health affects labour market status or whether a poorer labour market status affects mental health. The literature is equally unclear about the links between mental health and changes in employment status. This article fills this gap in the literature by assessing whether deteriorating health status precedes labour market transitions or vice versa. In particular, it presents temporal relationships between poor mental health and transitions between permanent and temporary employment, and thereby assesses if poor mental health affects or is affected by this type of labour market transition. Although our focus is on the transition between permanent and temporary employment, our methodological approach could be applied to other transitions.

This article contributes to the literature in three ways. First, it presents an investigation into the associations between three indicators of mental health (psychological distress, psychological anxiety and life satisfaction), an overall indicator of general health and transitions between temporary and permanent employment. Second, we draw on data from the British Household Panel Survey (BHPS) to understand whether the link between employment type and health status is more of a causal outcome and/or a selection effect. If the temporarily employed are identified as having poorer mental health than those in permanent employment then it is consistent with two mutually inclusive possibilities: (i) temporary employment generates adverse mental health effects and/or (ii) a selection effect whereby individuals with below average mental health are drawn away from permanent and into temporary employment. This is a particularly pertinent issue as Virtanen et al.'s (2005) review of the empirical associations between temporary employment and psychological morbidity suggests that many results may be confounded by selection bias: if the selection effect is discovered to be more prominent relative to a causal effect then cross sectional studies that present estimates of a negative influence of temporary employment on mental health status may be reporting upwardly biased estimates.

A potential confounding issue is that mental health is associated with job satisfaction, with either lower job satisfaction deteriorating mental health or worsening mental health adversely affecting job satisfaction. We extend our analysis to examine the effect of job satisfaction on mental health and in mitigating any effect of employment type on mental health. This extension is conducive to policy recommendations as mental health conditions can rarely be directly affected by managers whereas job satisfaction often can.