Introduction
A loss of work productivity can be a result of health impairments and arise from absenteeism (being away from work due to illness or disability) and presenteeism (being present at work but constrained in certain aspects of job performance by health problems) [
1]. Maintaining a healthy and productive workforce is increasingly challenging due to the continuing structural changes in the working environment, an aging workforce and an increasing number of employees affected by stress at work [
2]. Gaining better knowledge of the stress-related causes of absenteeism and presenteeism is therefore of high social and economic importance. A detailed analysis of the drivers of work stress-related productivity losses may be particularly useful to understand which employees are most at risk of incurring stress-related productivity losses and to identify those who might profit the most from interventions that improve work conditions.
Productivity losses are determined by multiple factors [
3], but work-related factors have often been proposed as especially important [
4]. According to the models developed in occupational health psychology, such as the Job-Demands Control model (JDC) [
5,
6] and the Job-Demands Resources model (JDR) [
7], unfavorable job conditions are associated with high levels of job stressors and a lack of job resources. Exposure to such job conditions can lead to stress among employees, resulting in decreased performance and motivation and, over time, in serious health problems [
8]. However, only a handful of empirical studies have analyzed these propositions in relation to productivity losses caused by absenteeism and presenteeism (e.g., [
4,
9,
10]). For example, a lack of job control, which is a well-established work resource [
11] defined as the ability to determine when and where work is done, has been shown to increase the risk of presenteeism [
12]. Another study found a similar relation with sickness absence but only for women [
13]. Additionally, high time demands and physical demands at work have been shown to be associated with presenteeism and absenteeism [
14,
15].
In this study, we estimate the effects of work stressors and resources on health-related productivity losses caused by absenteeism and presenteeism, and add to the current literature in three ways.
First, we estimate the effects of task-related and social job stressors and resources on health-related productivity losses, whereas the current literature mainly focused on task-related factors. Empirical evidence suggests that social stressors may be especially harmful to employee health and well-being, even more so than other job stressors [
16]. Furthermore, the recent “Stress-as-Offense-to-Self” model underlines the relevance of social job resources, such as appreciation at work, and highlights its absence as particularly stressful for employees [
17]. Absenteeism and presenteeism can also be explained by the social exchange perspective [
18]. According to this approach, the employee–organization relationship is a trade of effort and loyalty for benefits such as pay, social support and recognition [
19]. When employees are satisfied with this mutual exchange, they will be engaged in their jobs. However, when employees perceive the benefits received as too low compared to their contribution, they may withdraw from the relationship. Absenteeism and presenteeism can thus be seen as a method of restoring equity in the employee–organization relationship [
20]. Some previous studies support these assumptions, although evidence is still scarce. Injustice at work [
20], low organizational support [
18] and low workgroup cohesiveness [
21] have, for example, been shown to increase the risk of absenteeism. Similarly, negative relationships with colleagues [
22], role ambiguity [
23], and workplace bullying [
24] have been shown to increase the risk of presenteeism. A few studies also provide evidence of the relevance of positive social aspects at work. Employees working under a supportive supervisor [
25] who demonstrated strong integrity [
26] showed less presenteeism and absenteeism.
Second, in addition to social and task-related stressors and resources at work, we consider personal resources. According to the JDR model, job and personal resources can affect health and organizational outcomes both directly and indirectly; personal resources might enable employees not only to deal with job demands in a resilient way but also to make better use of available job resources [
7]. Previous studies have shown that personal resources are related to absenteeism; however, studies on presenteeism rarely consider personal factors [
27]. We included occupational self-efficacy as a relevant personal resource for individuals in organizations [
28,
29], expecting self-efficacy to act as a buffer for the negative effects of job stressors [
7]. Occupational self-efficacy is defined as the belief or confidence in one’s ability to successfully fulfill a task or cope with difficult tasks or problems [
30]. Previous research has shown direct beneficial effects of occupational self-efficacy on productivity [
28,
29], work-related behavior [
31] and job attitudes [
32] and demonstrated its moderating effects in the stressors–strain relationship [
33,
34]. However, to date, no studies have explored its effects on health-related productivity losses. Based on the conservation of resources theory (COR, [
35]), according to which individuals who lack resources are more vulnerable to resource loss and less capable of resource gain (negative spiral), we expected that employees lacking both job and personal resources are most at risk of experiencing health-related productivity losses due to absenteeism and presenteeism when job stressors increase. Furthermore, in line with the “gain paradox principle” [
35] according to which resource gains become more important when resources are loss is high, we expect that an increase in job resources is especially important for employees with low self-efficacy and high stressors.
Third, we contribute to the literature on the economic burden of work stress. Although work stress and its consequences for employees and employers are high on the political agendas of European institutions and policy-makers [
36], evidence of the economic burden of work stress is scant, especially regarding stress-related presenteeism [
37]. The few available studies suggest that the costs of work stress are substantial [
38]. We add to previous studies on the productivity losses caused by work stress by estimating the cost of employees’ health-related productivity loss due to presenteeism and absenteeism of being exposed to an imbalance between job stressors and job resources. Such an imbalance, according to occupational stress models (e.g., JDC, JDCR [
39]), results in work stress and has a high probability of leading to serious health problems. We calculated the total health-related productivity loss due working under unfavorable job conditions per employee and month, considering both absenteeism and presenteeism.
The aim of this study was threefold. First, we estimated the effects of task-related and social stressors and resources on health-related productivity losses due to absenteeism and presenteeism. We assessed stressors and resources at work based on six indices measuring (1) task-related work stressors (time pressure, task uncertainty, performance constraints, and mental and qualitative overload), (2) social work stressors (social stressors from supervisor and co-workers), (3) task-related work resources (job control and task significance), (4) social work resources (social support from supervisor, appreciation at work), as well as (5) overall work stressors and (6) overall work resources. We controlled for a wide range of confounding factors, such as socio-economic characteristics, job characteristics, private demands, and personal characteristics (self-efficacy). Second, we explored the interaction effects between job stressors, job resources, and personal resources. We aimed to understand which employees are most at risk if job stress increases and which employees would benefit the most from interventions improving the balance between job stressors and resources. Third, we built an economic model estimating the productivity losses caused by employee exposure to an imbalance between job stressors and resources.
Discussion
We estimated the impact of job stressors and job resources on productivity losses due to sickness absenteeism and presenteeism based on a representative survey of Swiss employees conducted in 2014 and 2015. First, we found that health-related productivity losses increase with an increase in job stressors and decrease with an increase in job resources, with social and task-related stressors and resources being equally important determinants. Second, the analysis of heterogeneous effects revealed that an increase in job stressors is especially harmful if job resources are low. These effects are even more pronounced if occupational self-efficacy is low as well. On the other hand, an increase in job resources is most effective in reducing health-related productivity losses if job stressors are high and occupational self-efficacy is low. Third, the results of a counterfactual analysis suggest that job stress (defined as job stressors exceeding job resources) accounts for 23% of the total health-related productivity losses due to absenteeism and presenteeism. This corresponds to CHF 195 per person and month.
Our findings contribute to studies on the effects of positive and negative social aspects of work on presenteeism and absenteeism. In line with research showing that social aspects of work may be especially relevant to employee health and organizational behavior [
25,
46], we found that social stressors and resources at work are important determinants of health-related productivity losses due to absenteeism and presenteeism in addition to task-related job stressors and resources. Moreover, we found that social and task-related stressors have direct and equal effects on health-related productivity losses, and while social resources remain a significant predictor, task-related resources do not. If employees work under unfavorable work conditions characterized by high levels of job demands, do not feel appreciated or respectfully treated at work, or lack social support, health-related productivity losses due to absenteeism and presenteeism might increase. This behavior can be seen as a method of employees restoring equity in the employee-organization relationship, as proposed by the social exchange perspective [
18]. These results have scientific and practical implications. Our findings suggest that both social and task-related factors should be considered in future studies and in planning interventions aiming to reduce health-related productivity losses by improving workplace conditions.
As expected, our results confirm that job resources buffer the negative effects of job stressors on productivity losses. These findings are in line with the buffering hypothesis of the JDC model as well as with the postulation that high-strain jobs, characterized by a combination of high job demands and low resources, should see the most harmful effects, while the combination of high demands and high resources is considered to be the most beneficial (active job) [
8]. Moreover, our results show that an increase of 1% in job stressors results in a larger effect on health-related productivity losses than a decrease of 1% in job resources. These results are in line with those of previous studies showing that negative conditions and events typically have stronger effects than good conditions [
71]. This implies that an increase in demands at work should always be accompanied by an even larger increase in job resources in order to prevent the negative consequences regarding health-related productivity impairments.
Our results also show that not only job resources but also occupational self-efficacy buffer the negative effects of job stressors on health-related productivity losses. Furthermore, we find that employees with a simultaneous lack of personal and job resources are the most vulnerable with respect to an increase in job stressors. This finding is in line with the vicious cycle postulated by the COR model: individuals who lack resources are more vulnerable to resource loss and less capable of resource gain. We also find that employees with low personal resources facing high job stressors are the ones who would profit the most from an increase in job resources. This is in line with the “gain paradox principle” of the COR model [
35], stating that resources are even more important when resource losses are high.
We do not find a significant productivity effect of an increase in job resources for employees with high personal resources and low level of job stressors. Therefore, an increase in job resources without a reduction in job stressors may not always be sufficient to reduce health-related productivity losses.
We add to the economic literature by estimating the total health-related productivity loss due to unfavorable job conditions. Our estimated productivity loss of CHF 195 per person and month may seem modest at first. However, extrapolation indicates that job stress may have cost Swiss companies up to CHF 10 billion in 2014, corresponding to 1.7% of the gross domestic product. This emphasizes the economic importance of interventions aiming to improve work conditions in general and the balance between work demand and resources.
Our study has several methodological and theoretical strengths. First, the cross-sectional data were representative of Swiss employees with respect to gender, age, region, and industry branch. Second, we tested the robustness of our cross-sectional results using longitudinal data, as this allowed the application of methodologically superior panel-data estimation methods. Third, we included several task-related and social work conditions and explored the relevance of positive and negative social aspects at work beyond the task-related aspect. Fourth, in addition to job resources, we considered personal resources—occupational self-efficacy—and explored interaction effects with job stressors and job resources.
Several limitations need to be taken into account.
First, self-reported measures such as the WPAI-GH may suffer from social desirability and recall bias. While a recall bias is unlikely, given the 1-week recall period of WPAI-GH, a social desirability bias is likely to be present. Studies comparing self-reported with company-registered absenteeism show that employees tend to underreport absenteeism [
22]. If this were due to social desirability, we would also expect employees to underreport presenteeism. While this would lead to an underestimate of the magnitude of health-related productivity losses, it would not necessarily bias the validity of the associations between workplace conditions and health-related productivity losses. A
second shortcoming related to the WPAI-GH is that it has not (yet) been validated against objective work productivity data. We thus do not know whether an employee-reported productivity impairment of 10% translates into a 10% loss of an employee’s value to the employer. A study comparing self-reported measures from the Work Limitations Questionnaire (WLQ) with objective productivity outcomes found that a self-reported 10% health-related limitation at work translated into a 4–5% reduction in work output. However, the generalizability of these results is unclear because the study was carried out in a single work setting and did not consider quality of work [
14]. If this overestimation in self-reporting of productivity losses applied to our data, it would imply an overestimation in our job stress-induced productivity losses. There is a clear need for more research on the extent to which employee-reported productivity measures translate into production losses for employers. A
third limitation relates to the high dropout rate in the second wave of the survey. Although we show that the inverse-probability-of-attrition weights are capable of correcting for selective attrition to a large extent, we cannot rule out the possibility that our panel data estimations are still biased.
Our results suggest that improvements in work conditions could help organizations to reduce previously undetected productivity losses by implementing programs targeting an improved balance between job stressors and job resources. We also show that an increase in job demands affects employees to different degrees depending on their levels of job and personal resources and that not everyone benefits from increased job resources. This finding highlights the need for organizations to take a tailored approach by providing additional attention to the most vulnerable employees. Moreover, our data suggest that job stressors and resources as well as health-related productivity losses vary greatly across occupations. Our sample size prevents the estimation of occupation specific effects though, offering an opportunity for future research.
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.