1 Introduction

The growing number of workers who feel insecure about the future of their jobs has produced concern among researchers about likely unfavourable consequences of job insecurity for employees’ well-being. This concern is prompted by fundamental changes in the labour market, such as the growth in temporary employment, globalization and new technologies, along with organizational restructuring and downsizing (De Witte 1999; Quinlan et al. 2001). These changes reflect a new reality in working life (Sullivan 1999) that requires changes in employment strategies.

Employability has become a key element in this debate, and it has captured both researchers’ and practitioners’ attention. Following the employability discourse, employees are expected to seek employment security rather than job stability, meaning that employees establish employment stability by moving across organizational boundaries and from one job to the next, rather than staying within a single organization or job (Forrier and Sels 2003). In this regard, employment security is conditioned by employees’ employability, defined as the available alternatives in the labour market. This situation marks a shift away from the traditional model of employment based on job security and social protection. Thus, there is a need for research on how people experience this new scenario. First, research is needed on individuals’ perceptions of employability and their implications for employees’ well-being. Perceived employability is expected to be associated with employees’ well-being in much the same way that perceived job security is. Second, the role of perceived employability in the relationship between perceived job insecurity and employees’ well-being should be explored. Perceived employability may be helpful in dealing with perceived job insecurity.

However, unlike job insecurity research, employability research is still scarce. The present study’s aims are twofold: first, we investigate the association between both perceived job insecurity and perceived employability and employees’ affective well-being. Second, we investigate perceived employability as a moderator of the relationship between job insecurity and affective well-being. Thus, on the one hand, this study replicates previous findings addressing perceived job insecurity correlates. On the other hand, it extends previous research in the field of perceived employability.

Well-being has been defined as a global assessment of life (Kim-Prieto et al. 2005), a mental state, the absence of negative feelings, a personal state brought about by the pursuit of a meaningful life for oneself, or a state resulting from the attainment of goals and rewards for efforts (Lu and Gilmour 2004). This study focuses on employees’ affective well-being, specifically general distress and life satisfaction. These variables were selected to measure context-free affective well-being, as suggested by Warr’s (1987) formulation of the content-discontent axis.

1.1 Job Insecurity and Well-Being

Job insecurity refers to employees’ perceptions and concerns about potential involuntary job loss (e.g., De Witte 1999; Greenhalgh and Rosenblatt 1984; Heaney et al. 1994; Van Vuuren 1990). It is mostly interpreted as a work stressor with probable unfavourable consequences for employees (Cheng et al. 2005; Jacobson 1991). For example, job insecurity implies uncontrollability and feelings of powerlessness, which are known to be related to poor well-being (e.g., De Witte 1999).

This hypothesis on the negative association between job insecurity and well-being has found support in many empirical studies (e.g., De Witte 1999; Hellgren et al. 1999; Hellgren and Sverke 2003; Cheng et al. 2005; Ferrie et al. 2005). In congruence with this evidence, we hypothesize the following:

H1:

Job insecurity is positively related to psychological distress (H1a), and it is negatively related to life satisfaction (H1b).

1.2 Perceived Employability and Well-Being

Unlike the case of job insecurity, there is no consensus on how to define employability (Fugate et al. 2004; Kanter 1989; Kluytmans and Ott 1999). Most authors, however, agree that employability refers to the employee’s chance of finding alternative employment, either on the internal or the external labour market (Forrier and Sels 2003). This concept has been assessed using both objective and subjective indicators. Some authors have used objective indicators related to human capital or career indicators to measure employability. Some examples are education, training, occupational position or the number of job changes (De Jong and Schalk 2005; Elman and O’Rand 2002; Forrier and Sels 2003; Van Dam 2004; Virtanen et al. 2003; Worth 2002). Other authors have defined employability according to subjective indicators, such as ‘employees’ perceptions of the available alternatives in the internal and/or external labour market’ (Berntson et al. 2006; Berntson and Marklund 2006; De Cuyper et al. 2008). We decided to follow this subjective approach for three reasons.

First, the way people interpret reality determines individuals’ feelings and behaviours. This assumption is a basic one in the social sciences, as formulated in Thomas’ theorem: “If men define situations as real, they are real in their consequences” (Thomas and Thomas 1928, pp. 572). For instance, when people are employable according to objective indicators, such as training, experience, etc., but they do not perceive themselves as such, their feelings will be congruent with their interpretation of reality. Second, according to stress theories (e.g. Lazarus and Folkman 1984), individuals’ reactions to stressors are partly dependent on individuals’ cognitive appraisals of the potential stressors. This is particularly true when predicting outcomes such as employee well-being. Third, objective indicators may fail to account for the labour market context when trying to predict actual chances of re-employment (Trevor 2001). In this sense, its usefulness may depend largely on the particular sector of the economy in which the worker is employed (De Grip et al. 2004).

Perceived employability is likely to be positively related to employee well-being. Unlike job insecurity, perceived employability may enhance employees’ sense of “self-governing”. According to self-determination theory (Ryan and Deci 2001) and eudaimonic approaches to wellness (Ryan et al. 2006), the experience of self-regulation is associated with greater wellness. In much the same way, several authors suggest that perceived employability is likely to be accompanied by a sense of control over one’s career, and this feeling, in turn, may be associated with employees’ well-being (Fugate et al. 2004; Marler et al. 2002).

A similar line of reasoning is expressed in recent speculations about the emergence of boundaryless careers, which develop in multiple employment settings and across organizational boundaries rather than within the context of a single organization, as in the case of traditional career models (Defillipi and Arthur 1994; Hall and Chandler 2005; Mirvis and Hall 1994). Whereas job security is assumed to be important in predicting employees’ well-being under the traditional career model, perceived employability may predict employees’ well-being in the case of the boundaryless career (Forrier and Sels 2003; Kluytmans and Ott 1999; Rajan 1997).

In this respect, De Cuyper et al. (2008) established a positive relationship between perceived employability and employees’ general and work-related well-being. Berntson and Marklund (2006), in a two-wave study among Swedish employees, found perceived employability to be associated with general health and mental well-being when baseline health was controlled for. Thus, our second hypothesis reads as follows:

H2:

Perceived employability is negatively related to psychological distress (H2a), and it is positively related to life satisfaction (H2b).

1.3 Job Insecurity and Perceived Employability

Some authors have argued that employability may not replace job security (Baruch 2001), as predicted by advocates of the boundaryless career concept. Instead, these authors suggest that employability may reduce the likely unfavourable consequences of job insecurity (Greenhalgh and Rosenblatt 1984; Fugate et al. 2004; Näswall 2004, 2005; Sverke and Hellgren 2002). Specifically, they suggest that the relationship between job insecurity and well-being is less negative when employees perceive many rather than few alternative employment opportunities. This possible interaction effect between perceived employability and job insecurity can be understood by referring to the Job Demand Resources Model (Bakker 2001; Demerouti et al. 2001).

According to the JD-R model, job demands refer to aspects of the job that employees consider burdensome (e.g., job insecurityFootnote 1). In contrast, resources have to do with all the aspects related to work that promote control, personal growth, development and learning (Demerouti et al. 2001; Hakanen et al. 2005), all of which are evident in the case of employability. In particular, as discussed earlier, employability has been highlighted as an important tool in increasing employees’ control over their working life (Berntson and Marklund 2006; De Grip et al. 2004). Furthermore, policy makers have identified job changes as a possible avenue to stimulate growth and learning. A core assumption of the Job Demand Resources model is that resources are likely to buffer the likely unfavourable effects of demands.

A similar idea, although based on Karasek’s Job Demand Control Model, is expressed by the notion of employment strain, which was proposed by Lewchuk et al. (2005), following suggestions by Cooper (2002), Quinlan et al. (2001) and Le Blanc et al. (2000). Employment strain results when employees face high demands together with low control, with this combination being shaped by the employment relationship and by employment conditions rather than by the job. Following this model, the combination of high job insecurity, which involves high demands, and low perceived employability leads to employment strain. Low perceived employability is expected to be accompanied by a low sense of control, since it implies that the employee sees his/her ability to secure future employment as limited.

To date, empirical research on employability as a moderator of the relationship between job insecurity and well-being is scarce. Some studies have considered variables that are closely related to perceived employability, such as employment security (Kuhnert and Vance 1993), perceived alternatives in the labour market (Büssing 1999) and chances of getting a new job (Mohr 2000). Along these lines, Kuhnert and Vance (1993) establish that employment security moderates the relationship between job insecurity and depression. However, no such moderation by perceived employability is found for the relationship between job insecurity and anxiety. Similarly, Büssing (1999) shows that the interaction between job insecurity and perceived alternatives in the labour market adds to predicting irritation/strain and psychosomatic complaints. Finally, Mohr (2000) finds that the relationship between job insecurity and psychosomatic complaints is stronger for those with few rather than many chances on the labour market. This finding was not replicated for anxiety, depression and irritation. However, these studies mostly recruited companies undergoing restructurings, which may have influenced the results, due to range restriction (Rousseau and Fried 2001; Hunter and Schmidt 1996). Organizational restructuring is perhaps the most important antecedent of job insecurity (Mauno et al. 2001), which means that variability in job insecurity might be considerably reduced in these contexts. This lack of variability may, in turn, result in non-significant results. In the present study, we investigate perceived employability as a potential moderator of the relationship between job insecurity and well-being, using a sample recruited from different organizational contexts, as follows:

H3:

Perceived employability moderates the relationship between job insecurity and well-being, (1) so that the positive relationship between job insecurity and psychological distress is weaker under the condition of high versus low perceived employability (H3a), and (2) so that the negative relationship between job insecurity and life satisfaction is weaker under the condition of high versus low perceived employability (H3b).

2 Method

2.1 Data Collection and Respondents

Data were collected among 639 Belgian employees from six organizations: an industrial setting (N = 144; 22.5%), three temporary help agency offices (N = 39, 6.1%), a research and development organization (N = 56, 8.8%), a research institute (N = 46, 7.2%), a shop (N = 38, 5.9%), and a hospital (N = 316, 49.5%). The organizations were recruited based on assumed variability in job insecurity and employability, and based on possibilities for generalizing findings. These organizations represent different sectors, and they differ with regard to employment stability, educational level and occupational status. Thus, even though not representative of the Belgian working population, the participants were selected taking into account the aims of the present study.

The researchers informed all the participants about the aims of the study and its relevance for all the parties involved, by means of group sessions. Additionally, the researchers underlined the practical implications and benefits for participant organizations, in order to promote participation and participants’ candidness. Furthermore, the questionnaire was accompanied by a letter from the research organization in charge of the project, stressing that confidentiality was guaranteed and that participation was voluntary. Participants completed the questionnaires during working time. This procedure was followed in every organization, except in the temporary help agencies, where the questionnaire was handed to the participants to fill out at home. Surveys were administered in Dutch. When published Dutch versions of the scales were not available, the instrument was translated following the translation/back translation technique (Behling and Law 2000). The response rates were high (over 50%) in all the organizations, except in the shop (22%).

Regarding the characteristics of the sample, more women (N = 395; 61.8%) than men (N = 243; 38.1%) participated in the study. Most of them were employed as white collar workers (N = 470; 74.8%) on a permanent contract (N = 458; 71.7%). About half of the respondents had not pursued education beyond the high school level (N = 299; 47.6%). Mean age of the sample was 36 years (SD = 10.3), and respondents worked an average of 36 hours per week (SD = 7.1).

2.2 Measures

For all scales reported in this section, Confirmatory Factor Analysis supported single factor structures with an acceptable fit.

Control variables: Type of contract (0 = permanent; 1 = temporary), gender (0 = female; 1 = male) and education (0 = no academic degree; 1 = academic degree) were dichotomous variables, while age (in years) and weekly working hours were continuous variables. We did not control for occupational position because of its high association with education (χ² (1, N = 617) = 175.6, p < 0.001). Similarly, we did not control for organization due to substantial overlaps between all the background variables (contract: χ² (1, N = 634) = 118.21, p < 0.001; gender: χ² (1, N = 638) = 94.3, p < 0.001; education: χ² (1, N = 628) = 187.5, p < 0.001; weekly working hours: F(5, 628) = 22.53, p < 0.001; age: F(5, 623) = 16.11, p < 0.001).

Job insecurity was measured with four items developed by De Witte (2000). A sample item was ‘I think I might lose my job in the near future’. The scale ranged from 1 (strongly disagree) to 5 (strongly agree). Cronbach’s alpha was .89 (χ² = 8.15, df = 2, p < 0.05; GFI = 0.99; AGFI = 0.97; RMSEA = 0.07).

Perceived employability was measured with five items developed by De Witte (1992) based on a literature review. This scale has been extensively used in different European countries, obtaining a good fit for its one-dimensional structure (Isaksson et al. 2003; PSYCONES 2003). A sample item is ‘I am optimistic that I will find another job, if I look for one’. The scale ranged from 1 (strongly disagree) to 5 (strongly agree). Cronbach’s alpha was 0.79 (χ² = 30.37, df = 5, p < 0.001; GFI = 0.98; AGFI = 0.94; RMSEA = 0.09). As job insecurity and perceived employability are potentially overlapping, we calculated their discriminant validity, χ² = 83.44, df = 26, p < 0.001; GFI = 0.97; AGFI = 0.95; RMSEA = 0.06.

Well-being was measured with two scales. First, to measure psychological distress, respondents were asked how often (0 = less than usual; 3 = much more than usual) they experienced situations described in the 12-item General Health Questionnaire (Golberg 1979, e.g., ‘Have you recently felt constantly under strain?’). This instrument was constructed to identify people suffering from common minor disorders, such as symptoms of anxiety and depression. It is commonly used to assess affective well-being on the contented-discontented axis (Warr 1987). Cronbach’s alpha was 0.82 (χ² = 299.40 df = 40, p < 0.001; GFI = 0.92; AGFI = 0.86; RMSEA = 0.10). Second, the life satisfaction scale was developed for the purpose of this study (Isaksson et al. 2003). Respondents had to indicate how satisfied they felt with each of seven possible domains (e.g., ‘How satisfied do you currently feel about your life in general?’). Responses ranged from 1 (very dissatisfied) to 7 (very satisfied). Cronbach’s alpha equalled 0.84 (χ² = 51.92, df = 11, p < 0.001; GFI = 0.98; AGFI = 0.94; RMSEA = 0.08Footnote 2).

3 Results

Correlations between variables, with their means and standard deviations, are shown in Table 1.

Table 1 Means, standard deviations and correlations
Full size table

Hypotheses were tested using hierarchical regression analyses for each outcome variable. List-wise deletion was applied, resulting in slightly smaller samples. In the first step of the analyses, the control variables (type of contract, gender, educational level, age and weekly working hours) were introduced. In the second step, job insecurity and perceived employability were introduced. Finally, in the third step, the interaction between job insecurity and perceived employability was introduced. In this regard, we followed the procedure recommended by Aiken and West (1991). Centred scores, also known as deviation scores, are recommended to avoid multicollinearity in regression analyses that involve interaction. Thus, we centred job insecurity and perceived employability at their means before calculating the interaction term (Table 2).

Table 2 Summary of hierarchical regression analyses
Full size table

The control variables did not add to explaining psychological distress. However, age was negatively related to life satisfaction, and both males and highly educated employees were less satisfied with their lives compared to females and low-educated employees, respectively. Our first hypothesis was concerned with possible main effects of job insecurity. As expected, job insecurity was positively related to psychological distress, and negatively related to life satisfaction. Contrary to expectations formulated in Hypothesis 2, perceived employability did not add to predicting psychological distress or life satisfaction. Finally, the interaction term between perceived employability and job insecurity predicted life satisfaction, but not psychological distress. Figure 1 shows the nature of the interaction effect between job insecurity and perceived employability on life satisfaction: the relationship between job insecurity and life satisfaction was more negative under the condition of low compared to high perceived employability. Thus, Hypothesis 3 was supported for life satisfaction but not for psychological distress.

Fig. 1
figure 1

The interaction between job insecurity and employability for life satisfaction

Full size image

4 Discussion

In the present study, we investigated the relationship between job insecurity, perceived employability and well-being, as measured by psychological distress and life satisfaction. As far as job insecurity is concerned, we expected a relationship with poor well-being, which was supported for both psychological distress and life satisfaction. These results replicate previous findings in the field of job insecurity (e.g., De Witte 1999; Hellgren et al. 1999). With regard to perceived employability, we hypothesized that perceived employability would be positively associated with employees’ well-being, which aligns with speculations about the growing importance of boundaryless careers. In this view, perceived employability would relate positively to well-being (Hypothesis 2). However, no such associations were found.

One possible explanation for this pattern of results is that employees still largely rely on job security. Perceived employability may be relevant only in buffering the unfavourable consequences of job insecurity, as suggested by various authors (Forrier and Sels 2003; Fugate et al. 2004; Greenhalgh and Rosenblatt 1984; Iles et al. 1996; Kanter 1989; Sverke et al. 2002), which would imply that the relationship between job insecurity and well-being is moderated by perceived employability (Hypothesis 3). Evidence for this was found in the case of life satisfaction, but not in the case of psychological distress. These mixed findings agree with earlier research that shows that employability-like indicators moderate the relationship between job insecurity and some outcomes, but not others. For example, in the study by Kuhnert and Vance (1993), the interaction term between job insecurity and employability was related to depression, but not to anxiety. Also, the interaction term between job insecurity and employability was not predictive of anxiety and depression in the study by Mohr (2000). Employability may be less likely to buffer unfavourable consequences of job insecurity when these involve early stages of anxiety, depression or strain as measured by the GHQ-12. In contrast, employability may reduce the negative relationship between job insecurity and milder forms of strain, such as life dissatisfaction.

4.1 Limitations

Perhaps the most important limitation of this study is its cross-sectional design, which hampers causal interpretations. However, our hypotheses aligned with theory and earlier longitudinal research, which may suggest that perceived job insecurity and perceived employability are more likely to cause well-being than the other way around (Hellgren and Sverke 2003; Berntson and Marklund 2007). Still, the relationship between perceived employability and well-being could be reciprocal. For example, relevant factors related to employees’ well-being, such as depression or negativism, may be associated with lower levels of perceived employability.

A second limitation concerns the risk of common method variance due to using self-reported data. Given that the data were gathered with surveys, social desirability, acquiescence and negative affectivity may lead to common method variance. Some authors, however, argue that the magnitude of the possible inflation of relationships due to common method variance should not be overestimated (Crampton and Wagner 1994; Spector 1987). Furthermore, Podsakoff et al. (2003) provide useful suggestions for questionnaire design that may reduce the risk of common method variance, many of which were followed in this study (e.g., changes in the response format, anonymity, instructing the participants that there are no right or wrong answers). Also, Conway and Briner (2002) argue that self-reports are likely to attenuate rather than strengthen interaction, which may suggest that the interaction effect found is quite powerful. Furthermore, we would like to highlight that, as expected, education and temporary employment were significantly correlated with perceived employability and job insecurity, respectively. Education and temporary employment are often used as objective indicators of employability and job insecurity, respectively. Additionally, temporary employment is positively correlated with perceived employability. The fact that temporary employees have much more experience with being re-employed may account for these results. These findings suggest that our measures tapped perceived employability and job insecurity, as intended.

Third, the explained variance was quite low for both well-being indicators and, in particular, for psychological distress. Still, similar results have been found in previous research addressing the stressor-strain relationship (Zapf et al. 1996). Several factors may account for these results (Zapf et al. 1996). First, the multi-causal nature of the stress-strain relationship is well known. Many factors influence physical and mental health; hence, the impact of a single work stressor is expected to be weak (Semmer et al. 1996; Zapf et al. 1996). Second, the way stressors affect psychological and psychosomatic dysfunction over the course of time may differ depending on both the nature of the stressor and the criterion variable (Zapf et al. 1996). In this sense, addressing long-term consequences of stress using a cross-sectional design may result in an underestimation of the stressor-strain relationship (Abelson 1985; Zapf et al. 1996). It is noteworthy that Sverke et al. (2002) distinguish between immediate (e.g., attitudes) and long-term effects (e.g., health manifestations). This distinction is based on Lazarus and Folkman (1984)’s stress model and Zapf et al. (1996)’s literature review on longitudinal studies in the field of stress research. Further support for this argument can be found in the observation that the explained variance is particularly low for psychological distress as measured by the GHQ-12, which may be the ultimate outcome. Finally, this study measures overall job insecurity and employability using a small number of items, which may restrict the sample variation, resulting in a small percentage of explained variance.

4.2 Concluding Remarks

The present study has important theoretical as well as practical implications. Regarding the theoretical implications, our study provided evidence about the role of employability in the current labour market. Specifically, we established that perceived employability may not replace job security as an important predictor of employees’ well-being. Instead, perceived employability may reduce the unfavourable consequences of job insecurity, at least when these consequences do not involve an early stage of depression or anxiety. This evidence is particularly strong, as we sampled many different organizations, in order to overcome any possible problems related to range restriction. With regard to implications for practitioners, the importance of perceived employability was highlighted in this study. Specifically, perceived employability may buffer the negative consequences of feelings of job insecurity, which are sometimes inevitable in the current labour market. Thus, employment policies based on employability should consider the individual’s appraisal of the situation. The usefulness of these policies depends on the degree to which they increase individuals’ likelihood of finding a job, but it also depends on their effectiveness in diminishing the distress that usually accompanies job instability. This study used psychological distress as an example, but our findings may be generalized to other particularly relevant outcomes for organizations, such as organizational commitment or performance.