Date post: | 27-Nov-2023 |
Category: |
Documents |
Upload: | independent |
View: | 0 times |
Download: | 0 times |
1
Scarred from the past or afraid of the future?
Unemployment and Job Satisfaction across European Labour Markets
Thomas Lange
Abstract
Previous research has shown that both, past unemployment and anticipated future unemployment
have a detrimental impact on employees’ attitudes and behaviours, which may affect organizational
performance. Surprisingly, however, very little is known about the relative impact of past
unemployment compared with current job insecurity. Although it is possible that both effects operate
simultaneously, the analysis in this paper, focused on employees’ job satisfaction and utilising a set of
cross-sectional data derived from the European Social Survey 2006-07, reports on a strongly
pronounced insecurity effect: anticipated unemployment substantially reduces employees’ job
satisfaction. Interestingly, including the perceived risk of future unemployment as a separate
predictor variable in ordered probit regressions relegates the experience of past unemployment to a
statistically insignificant coefficient and thus weakens the ‘scarring’ hypothesis. These results hold
true even when several socio-demographic characteristics and proxies for individual personality
traits are controlled for. Implications for organisations and human resource practitioners and scope
for future research endeavours conclude the analysis.
1. Introduction
Organisational change and restructuring, the growing implementation of downsizing practices and
respective unemployment risks at the level of the employee have all become commonplace
phenomena in an increasingly competitive business environment. In this context, it is perhaps
unremarkable to note that there have been major changes to the employee-employer relationship. It is
now argued that a new psychological contract has emerged, which no longer promotes lifetime
employment and loyalty to the firm. Instead, the focus has seemingly shifted towards the paradigm of
‘employability’ (Martin, Staines and Pate 1998; Forrier and Sels 2003) in response to the fear that
jobs may be dramatically altered or eliminated altogether. Unemployment and feelings of job
insecurity have certainly become an all too common experience in many employees’ working lives.
Whilst the effects of job loss on those who are currently and directly affected by the layoff
(Leana and Feldman 1994) as well as the negative effects on the firm’s survivors (e.g. Kalimo, Taris
and Schaufeli 2003; Sahdev 2003) have figured prominently in the organisational research literature,
there has also been the development of a substantial literature devoted to the inter-temporal effects of
unemployment - the impact of past unemployment as well as the effects of perceived future
unemployment risk.
By reference to the experience of previous unemployment spells, the economics discourse in
the subjective well-being arena, itself an extension of earlier practice in psychological scholarship, has
demonstrated empirically that past unemployment can lead to a marked deterioration in current life
satisfaction. This ‘scarring effect’ of past unemployment (Clark, Georgellis and Sanfey 2001),
2
suggesting that unemployment experienced in the past inflicts permanent damage on the human
psyche and thus leaves a psychological scar, has also been noted in assessments of several satisfaction
domains, including the job satisfaction of workers who have become re-employed since their
unemployment experience.
Job insecurity, as the perceived risk or fear of future unemployment, has also been shown to
lead to adverse attitudinal reactions in the workplace - intentions to quit, reduced commitment, and
reduced satisfaction (e.g. Chirumbalo and Hellgren 2003; Laine, van der Heijden, Wickström,
Hasselhorn and Tackenberg 2009). Insecure employees frequently experience increased psychological
distress and channel their anxiety through various coping strategies, which may lead to detrimental
outcomes for the organisation. Comparing the impact of job insecurity and short-term unemployment
spells on individuals’ psychological well-being, De Witte (1999) notes that both experiences appear to
be equally harmful. This finding builds on earlier results in the stress literature, suggesting that the
fear of losing one’s job is as detrimental to individuals’ health and well-being as losing the job itself
(Lazarus and Folkman 1984).
The unemployment – subjective wellbeing relationship is clearly of interest from an HR-
practitioner perspective since the widely reported negative impact of past as well as potential future
unemployment manifests itself in detrimental effects in workplace behaviour. Studying the impact of
past unemployment and future unemployment risk on employees’ job satisfaction is thus particularly
relevant, especially since the latter construct has been shown to serve as a powerful predictor for such
employee behaviours as commitment, motivation, absenteeism, quitting intentions, and other affective
responses to aspects of the job or the employer. Surprisingly, however, very little is known about the
relative impact of past unemployment compared with current job insecurity on employees’ job
satisfaction. Is it primarily past unemployment or the perceived risk of future unemployment that
serves as the driving force behind the detrimental impact on employees’ reduced happiness at work?1
Recently, the economics discourse attempted to disentangle the impact of past unemployment
and future unemployment risk on individuals’ generic well-being (life satisfaction). In their analysis,
using longitudinal German data and correcting for potential causality problems via fixed-effect
models, Knabe and Rätzel (2011: 283) suggest that “the scar from past unemployment operates via
worsened expectations of becoming unemployed in the future, and that it is future insecurity that
makes people unhappy”. Although constrained by data on individuals in Germany only and a sole
focus on life satisfaction, their results nevertheless raise further questions about the relative impact of
past unemployment and perceived future unemployment risk on other life domains, including job
satisfaction.
Against this background, the intention of the present paper is to adduce additional results to
the unemployment – subjective wellbeing relationship and perform a complementary analysis focused
on employees’ job satisfaction across several European countries. To this end, the present study
utilises data derived from the 2006-07 European Social Survey.2 The general survey captures
3
observations for 25 European countries and provides a rich, cross-sectional source of respondents’
economic, demographic and personal characteristics, together with information on respondents’ job
satisfaction, personality disposition, past unemployment experience during the last 5 years and an
individual’s perception of future unemployment risk. The latter is differentiated by the degree of this
risk, ranging from ‘very likely’ to ‘not at all likely’. The data set at our disposal is thus particularly
suited for the purpose of this analysis.
The remainder of this paper is organised as follows: a brief summary of the reported impact
of past unemployment and the perceived risk of future unemployment on several well-being indicators
is provided in Section 2. In Section 3, details of the data, econometric framework and data limitations
are introduced, followed by the results and respective interpretations in Section 4. Section 5 provides
some concluding remarks.
2. Unemployment, Job Insecurity and Well-being
Spanning across several academic disciplines, the impact of both, past unemployment and job
insecurity on individual well-being in general, and workplace behaviour in particular, has attracted an
enormous, nuanced and often controversial literature. In this study, only a ‘snapshot’ summary of the
latter is attempted.
At the outset, it is worth noting that unemployment related effects on well-being can be
positioned in a historical context. To the extent that the ‘value of unemployment’ can be linked to a
utility trade-off between time spent in leisure and time spent in paid work, the relationship between
unemployment and well- being can be traced back to simple labour supply models, basic utility and
indifference curve analyses in the economics discourse. Indifference curve analysis, developed by
Edgeworth, Pareto and others in the first part of the 20th century and derived from ordinal utility
theory (Blaug 1985), posits that individuals can always rank any consumption bundles by order of
preference. Strictly speaking, therefore, it would be reasonable to expect that the unemployed report
higher levels of well-being than the employed, holding income constant. Put differently, a
substantial degree of disutility is gained from time spent in paid work, because this time involves
effort and clearly distracts from the ability to contribute to activities that individuals would
otherwise enjoy. On this basis, it is reasonable to argue that people should engage in work solely
because their wage payments compensate for the opportunity costs associated with foregone
leisure.
Beyond economists’ simple labour supply model, however, numerous examples in the
psychology literature serve as testimony to support the claim that there are substantial, non-
pecuniary costs associated with unemployment, just as they are substantial non-pecuniary
4
benefits associated with working. Eisenberg and Lazarsfeld (1938) were amongst the first
psychologists to demonstrate, albeit with entirely descriptive methods, that unemployment can lead to
emotional destruction. What is more, extending their analysis to unemployment risk and job
insecurity, they note that “just having a job itself is not as important as having a feeling of economic
security. Those who are economically insecure, employed or unemployed, have a low morale.”
(Eisenberg and Lazarsfeld 1938: 361). These early observations are consistent with subsequent
empirical findings, which suggest that unemployment can cause numerous detrimental psychological
effects, including anxiety, lack of self-esteem, limited locus of control, limited regularity associated
with a working day, limited social contact and a weakened definition of personal status, identity and
self-worth (Jahoda 1981, 1988; Judge, Locke and Durham 1997; McKee-Ryan, Song, Wanberg and
Kinicki 2005; Paul and Moser 2009). It follows that individuals who experience unemployment, or
are at risk of becoming unemployed in the future, worry about more than labour income. Rather, the
withdrawal of a variety of latent functions in life appears to highlight the true destructive nature of
unemployment.
In the literature on workplace behaviour, these detrimental effects continue to be reaffirmed.
Previous studies in the psychological discourse have argued, although often on the basis of very small
data samples, that past unemployment causes long-lasting damage even when individuals have
become re-employed (Mallinckrodt 1990). Utilising larger data sets and building on the influential
work by Clark, Georgellis and Sanfey (2001), the impact of past unemployment on job satisfaction is
now firmly established in the empirical literature. Based on data from the European Household
Community Panel, for example, Kaiser reminds us that “the experience of short or long term spells of
unemployment in the past causes a consistent dissatisfaction with the current job, regardless of the
considered job satisfaction category. This, in turn, clearly supports the hypotheses of so-called
‘scarring effects of unemployment’ […]” (Kaiser 2007: 87). As mentioned previously, these findings
are particularly relevant in a workplace behavioural context since a large number of studies have
demonstrated that job satisfaction serves as a strong empirical predictor for employees’ commitment,
motivation, quitting intentions and even changes in culturally motivated attitudes (e.g. Cotton and
Tuttle 1986; Judge, Thoresen, Bono and Patton 2001; Saari and Judge 2004; Lange 2009, 2010).
In a similar vein, the impact of job insecurity on individual well-being measures has attracted
a substantial literature, with contributions across several academic disciplines. A number of studies
have assessed the health of unemployed individuals and compared it with symptoms of insecurely
employed workers. Although it is often assumed that the stress levels of the unemployed are
particularly heightened, similar levels of distress have also been reported for individuals who face job
insecurity (e.g. De Witte 1999). Interestingly, it has been noted that the level of distress fell for those
who were ultimately laid off, whereas levels of distress and burnout remained high for a comparison
group of employed workers whose job futures continued to be uncertain (Dekker and Schaufeli 1995).
The literature on impaired well-being in the health arena has also provided insights on the causality
5
direction of the relationship between job insecurity and mental distress. So observe Hellgren and
Sverke (2003), using longitudinal questionnaire data that job insecurity does indeed lead to mental
health complaints, whereas the reversed effects were shown to be statistically insignificant.
In the specific context of workplace behaviour, job insecurity was found to be strongly and
negatively associated with organisational commitment, job satisfaction and other organisationally
oriented attitudes and behaviours (e.g. Lim 1997), whilst it was reported to be strongly and positively
associated with the inclination to leave the organisation (e.g. Cavanaugh and Noe 1999). More
recently, it has been noted that job insecurity is strongly linked with various forms of workplace
aggression (Appelbaum, Deguire and Lay 2005) and that alcohol consumption also interacts with job
insecurity to predict aggression against subordinates (LeBlanc and Barling 2004). Insecure employees
have been shown to display lower creativity scores than their securely employed counterparts (Probst,
Stewart, Gruys and Tierney 2007), and a recent meta-analytical analysis also reported on a strong and
negative association between job insecurity and job performance (Cheng and Chan 2008).
Finally, in one of the few studies that assess the impact of both, a past experience and an
anticipated future reality in a work context Kalimo, Taris and Schaufeli (2003) examined workers’
well-being in Finland as a function of surviving previous downsizing and expectations concerning
future downsizing. The past experience of downsizing survival as well as the anticipation of future
downsizing was associated with elevated levels of psychological strain, cynicism, and absenteeism
behaviour. Feelings of inequity were also noted, and workers with elevated levels of such feelings
also considered early retirement more often than others. The statistical effects, although overall
relatively small, were reported to be stronger when downsizing was thought to lead to layoffs.
The findings on the impact of past unemployment and future unemployment risk clearly have
important implications from both, the individual and the organisational perspective. Surprisingly,
however, despite a plethora of scholarly contributions, we have yet to uncover if these experiences
display significantly different or broadly equal effects on the reduction of employees’ reported
satisfaction scores. By specific reference to job satisfaction, the present study aims to close this gap in
the literature. The data, respective limitations and empirical specifications in support of this
endeavour are brought forward in the following pages.
3. Data and Empirical Specifications
The data for this analysis originates from the 2006-07 European Social Survey (hereafter ESS). The
general survey is a cross-sectional account and covers responses from 25 European countries. ESS
contains detailed information on job satisfaction and various socio-demographic characteristics,
including information on age, gender, marital status, health, income and educational attainment. A
number of studies also uncovered the positive impact of opportunities for ongoing education and
6
training on workers’ job satisfaction (Gazioglu and Tansel 2006). Since information on participating
in training courses, lectures or conferences during the past 12 months is available in the ESS in the
form of a dummy variable, this information is also included in the empirical analysis.
Crucial to the aim of the present study, ESS also allows for access to information on
respondents’ experience of unemployment in the past 5 years and reports on the anticipated risk of
future unemployment in the next 12 months. Observations for the past unemployment variable
constitute a sub-section of those respondents who answered positively to the question: “Have you ever
been unemployed and seeking work for more than 3 months?” Limiting the analysis to an
unemployment experience within the past 5 years naturally results in a loss of usable information,
since individuals with an unemployment experience of more than 5 years ago are excluded from the
assessment. However, a time constraint on the unemployment experience was nevertheless deemed
appropriate, especially since set point theory (Headey and Wearing 1992) would predict that
unemployment experienced a very long time ago – say some 20 or 30 years ago - may no longer affect
an individual’s well-being today. Notwithstanding these remarks, recent research on set-point altering
consequences of unemployment needs to be acknowledged. In this context, it is argued that on
average, previously unemployed individuals do not completely return to their former levels of
satisfaction, even after they become reemployed. It is thus suggested that certain life events, such as
unemployment, can have a strong influence on long-term levels of subjective well-being (e.g. Clark,
Georgellis, Lucas and Diener 2004).3 However, these arguments may not hold for single
unemployment spells (Headey 2008), and may apply more strongly to generic unemployment rather
than a previous experience of layoffs (Clark, Diener, Georgellis and Lucas 2008). Of particular
relevance to the chosen time constraint in the present analysis, set-point altering consequences have
not been assessed over a very prolonged period of time, but appear mostly restricted to lag effects of
up to 5 years. On this basis, the present study proceeds with caution and focuses on a 5-year
timeframe for the experience of previous unemployment.
Proxy variables for levels of self-esteem, generalised self-efficacy, locus of control, and
neuroticism are also available. Accounts of external core evaluations are also added. Whereas the
former represent self-appraisals, the latter represent appraisals individuals make of their external
environment, pertaining to other people and the outside world in general. These variables capture
values which are broadly consistent with personality traits that are said to serve as strong dispositional
predictors of job satisfaction (Judge, Locke, Durham and Kluger 1998). Table 1 lists the chosen
variables for each dispositional category.
[Table 1 near here]
The inclusion of these variables is deemed important, since our cross-sectional data does not
allow us to control for time-invariant personality traits. Ignoring these traits would thus raise doubts
7
about the causality of the relationship between unemployment and job satisfaction. For example,
anxious employees or employees with low levels of self-esteem may experience more frequent
unemployment spells, or may be more pessimistic about their future job prospects. Past
unemployment or the perception of future unemployment risk could thus be correlated with lower
levels of job satisfaction, yet the relationship could be simultaneous rather than causal. Although the
cross-sectional data at our disposal cannot completely overcome these potential causality problems,
by controlling for several personality traits the risk is reduced.
By excluding countries with missing or incomplete information on the main variables of
interest, the analysis is restricted to the following 17 countries: Austria, Belgium, Bulgaria,
Switzerland, Germany, Denmark, Spain, Finland, France, United Kingdom, Netherlands, Norway,
Poland, Russian Federation, Sweden, Slovakia and Slovenia. The focus of the analysis is on
employees in full-time employment.4 Workers in agriculture, the fishing industry and the armed
forces are excluded. Further excluding observations with missing or inconsistent values and restricting
our sample to individuals between 18 and 65 years of age yields an effective sample of 3411
observations - 1714 for male and 1697 for female employees. In order to avoid relatively small
samples at country level and respective estimation biases, a pooled data sample for the chosen
countries was created. As such, the present study focuses on Europe-wide data covering responses
from a number of countries, rather than data for single economies. The downside of this treatment is
that potentially important contrasts between more narrowly defined groups may moderate the results.
For example, a potential risk may originate from the possibility that insecure individuals may have
self-selected into more secure jobs because of a specific country’s employment protection legislation.
However, in controlling for country-specific effects, the residual risk is mitigated by including
country dummies in the regressions.
Job satisfaction in the ESS is an ordinal categorical variable on a scale of 0–10, where 0
represents ‘‘extremely dissatisfied’’ and 10 represents ‘‘extremely satisfied’’. Specifically, the survey
question on job satisfaction asks: ‘‘All things considered, how satisfied are you with your present
job?’’ The experience of unemployment during the past 5 years is measured by a dummy variable.
The variable on the perception of future unemployment risk is ordinal-categorical in nature, ranging
from ‘very likely’ to ‘not at all likely’, in response to the question: “How likely would you say it is
that you will become unemployed in the next 12 months?”
The definitions and sample means of all variables used in the empirical analysis are presented
in the Appendix.
Based on the above sample, job satisfaction regressions are performed, using an ordered
probit model. In the context of the present study, it is assumed that the subjective measurement of job
satisfaction is determined by a transformation of employees’ characteristics into a cardinal latent
index, Si*, which is interpreted as a proxy for unobserved utility. The index follows the normal
distribution with mean µi and unit variance [Si* ~ N (µi , 1)]. In formal terms, Si
* is given by:
8
iii ezS * , (1)
where iz is a vector of explanatory variables describing employees’ characteristics, represents a
vector of parameters to be estimated and ie is a random error term, which is assumed to be normally
distributed. The observation mechanism is
Si = j if τj-1 ≤ Si* ≤ τj (2)
for j = 1, …., J, where J is the total number of categories.
Given the constraints τl < τm for all l < m and τ0 = -∞ and τJ = ∞, the cardinal index of
unobserved utility is subsequently mapped into observed subjective ordinal evaluations of job
satisfaction, iS , which are determined from the model as follows:
.10
.
.
.
2
1
0
*
10
3
*
2
2
*
1
1
*
i
i
i
i
i
Sif
Sif
Sif
Sif
S
(3)
i , with i = 1, …, J, denotes thresholds to be estimated along with the parameter vector .
The interpretation of this regression model is based on coefficients and, therefore, accounts for the
sign and statistical significance. It should be noted that the parameters do not calculate marginal
effects on job satisfaction. Positive signs for the estimated parameters indicate higher levels of job
satisfaction as the value of the associated variable increases. In other words, a positive value for
reveals that the entire distribution of Si* moves to the right. With a decreasing value of the associated
variable, negative signs for suggest the converse.
4. Empirical Results and Interpretations
Before the above mentioned regressions are performed, it is worth noting that the experience of past
unemployment may be an indicator of an employee’s perceived risk of future unemployment. This
9
scenario is certainly corroborated in Table 2. Amongst the employees in our sample who experienced
unemployment during the last 5 years, over 30 per cent reported that becoming unemployed again in
the next 12 months is likely or very likely. This compares with less than 15 per cent of respondents
who did not report on past unemployment. Similarly, over 86 per cent of those without a previous
unemployment experience reported that future unemployment is not very likely or not at all likely.
The respective proportion reduces to just over 67 per cent for those employees who experienced
unemployment in the past. Building on these rudimentary findings, it is thus possible to conceptualise
that past unemployment affects an employee’s job satisfaction directly via the past experience of
unemployment, and indirectly via the psychological impact on the perceived future risk of becoming
unemployed.
[Table 2 near here]
Based broadly on the methodological specifications by Knabe and Rätzel (2011), the effect of past
and potential future unemployment can be disentangled by introducing separate predictor variables
and running cumulative, sequential regression models, which allow us to control for these variables in
a step-wise manner. This is the approach taken in the present study. In the first sequence, the ordered
probit model is executed with standard socio-demographic background variables and a variable
representing the experience of past unemployment only. The results, shown in column (1) of Table 3,
are consistent with previous cross-sectional research on job satisfaction.
Specifically, the impact of age on employees’ job satisfaction displays the frequently
observed U-shape, which Clark, Oswald and Warr (1996) attribute to lifecycle, non-job related
aspects of individuals’ circumstances. The results also indicate that men are, on average, less satisfied
with their job than their female counterparts and that being married exhibit a positive impact on job
satisfaction, although neither effect is statistically significant at the conventional levels. Self-reported
health ratings display a positive and statistically significant impact, reflecting perhaps the fact that
those employees who benefit from good health tend to report higher levels of satisfaction with all
aspects of their life (which may spill-over into the job domain) or because they are generally more
likely to be in satisfying jobs, compared with those who are limited by poor health or disability
(Georgellis and Lange 2007). Again broadly in line with previous cross-sectional research in the
subjective well-being arena, middle and upper level income has a positive effect on employees’ job
satisfaction, compared with lower level income as the reference category. Educational attainments, on
the other hand, display a negative impact, although only upper-level education is statistically
significant. In contrast, the effect of ongoing education and training opportunities during the past 12
months is strongly and positively associated with employees’ job satisfaction. The negative impact of
educational attainments may be explained by reference to discrepancy theory (Locke 1976). The
result may thus be linked to employees’ perception of the correspondence between what they expect
10
and what they receive at work, in the sense that higher educational attainments may have resulted in
expected pecuniary or non-pecuniary benefits, which have failed to materialise. This is not an
uncommon interpretation in the literature (e.g. Robinson, Murrells and Clinton 2006). Certainly in
routine or less demanding jobs, it is now widely accepted that well educated employees often feel
more frustrated and less satisfied than their less educated counterparts (e.g., Vila and Garcia-Mora
2005). The strong and positive effect of ongoing training opportunities, to take a very different
example, may be understood on the basis of social exchange theory (Blau 1964). The motivational
processes of the theory incentivise workers who believe that committed organisations provide
education and training opportunities for the benefit of the worker to reciprocate by way of attitudinal
and behavioural commitments that are of benefit to the firm. Empirical support for such a process is
also shown by Georgellis and Lange (2007) for the German labour market.
Turning our attention to the first inter-temporal unemployment effect in the regression, the
experience of past unemployment – in line with the ‘scarring hypothesis’ – exhibits a strong and
negative impact on how employees rate their current job satisfaction. The result thus adds credence to
the assertion that, net of various socio-demographic characteristics, unemployment experienced in the
past makes an individual’s current subjective well-being less satisfying, even if the individual has
become re-employed in the meantime (Clark, Georgellis and Sanfey 2001).
In the second sequence of the regression approach, the ordered probit model is repeated with
the same socio-demographic background variables, but this time with both of the inter-temporal
predictor variables included: the experience of past unemployment and the perceived risk of future
unemployment. The inclusion of both variables will allow us to examine if the negative impact of
unemployment experienced in the past continues to hold once the perceived risk of future
unemployment has been controlled for. The respective results are shown in column (2) of Table 3.
A brief glance at the coefficients associated with the socio-demographic variables
demonstrates that controlling for the risk of future unemployment has not altered their effects on
employees’ job satisfaction in a substantial manner. Age, gender, marital status, health, educational
attainment, ongoing training, and income all still display statistical effects with the expected signs. In
terms of their statistical significance, gender and marital status remain statistically insignificant, and
the statistical impact of income has reduced marginally to a 5 per cent level of significance.
However, there is now strong evidence in support of the argument that the perceived risk of
future unemployment substantially reduces an employee’s job satisfaction. In absolute value terms,
the size of the different coefficients associated with future unemployment risk is noteworthy.
Compared with a negligible risk of future unemployment (‘not at all likely’), the job satisfaction of an
employee who perceives the risk of future unemployment to be ‘likely’ or ‘very likely’ falls
significantly below that of someone who has no concerns about being unemployed again in the future.
Compared with the latter, even someone who believes that the chance of future unemployment within
the next 12 months is ‘not very likely’ still displays lower job satisfaction ratings. Most interestingly,
11
controlling for future unemployment risk relegates the impact of unemployment experienced in the
past to a still negative, but statistically insignificant coefficient, thus weakening the claims of the
‘scarring hypothesis’. It is entirely possible that the experience of past unemployment influences the
propensity to worry about the risk of future unemployment. However, net of several socio-
demographic variables, it is not unemployment experienced in the past, but the fear of future
unemployment, which accounts for the reduction in employees’ job satisfaction.
[Table 3 near here]
In the final sequence of the regression approach and in an attempt to control, at least partially,
for individual heterogeneity in a cross-sectional data environment, several proxies for different
personality traits are added to the ordered probit model. This is based on the presumption that
individuals who are predisposed to hold predominantly positive (negative) assumptions about their
world – in terms of both, external and self-evaluations - are more likely to report that they find their
jobs more (less) satisfying. Broadly consistent with the suggestions by Judge, Locke, Durham and
Kluger (1998), a number of these proxy variables have indeed a strong impact on an employee’s job
satisfaction. The final regression results are shown in column (3) of Table 3.
Whilst the frequency of feeling anxious as a measure for neuroticism, and thus as a negative
lens through which an individual’s environment is interpreted, exhibits a predictably negative impact
(e.g. Clark and Watson 1991), the effect is not statistically significant at the conventional levels.
However, all of the other personality traits display a statistically strong impact on employees’ job
satisfaction. It is clear that the way in which people see themselves strongly affects how they
experience their jobs. Specifically, the regression results highlight the strong and positive impact of
self-esteem, measured by the perception of the value of an individual’s activities in life and frequently
displayed as a strong association with and as an important source of positive affectivity (e.g. Wood,
Heimpel and Michela 2003). The result is also consistent with many previous findings on the strong
association between self-esteem and job satisfaction (e.g. Locke, McClear and Knight 1996). Whilst
previous work by Judge and his collaborators (e.g. Judge, Thoresen, Pucik and Welbourne 1999) has
collapsed self-esteem and generalised self-efficacy into a single-core self-evaluation construct to
predict managerial coping mechanisms and job satisfaction, there is also strong counter-evidence to
suggest that these constructs should be kept separate (Chen, Gully and Eden 2004). The latter
approach is followed in the present study. To this end, the regression results also report that job
satisfaction is strongly and positively related to the degree of freedom associated with how to live
one’s live, interpreted as a proxy for someone’s fundamental ability to exercise general control over
life events and thus cope with life’s exigencies. Similarly, locus of control, understood broadly in the
present study as the degree to which individuals believe that they control specific events in their lives,
exhibits a strong and positive effect. The effect of decision-making freedom at work, as the proxy
12
variable, has been shown to meet employees’ ‘higher-order needs’ (e.g. self-expression, autonomy
and independence) and ultimately enhance job satisfaction (e.g. Gagné and Bhave 2011). By reference
to external core evaluations, the regression results emphasise the importance of interpersonal
trust/distrust and optimism/pessimism with regard to the future of the world. As the findings suggest,
those who believe that they live in a trusting world with a promising future ahead also rate the
satisfaction with their job more highly than those who hold opposite views. As such, it confirms the
assertion by Judge, Locke, Durham and Kluger (1998: 18) who contend that “people who consider
other people to be fundamentally untrustworthy or the world to be a dangerous place will view their
jobs in a much less benign way than those with the opposite premises”.
Personality traits thus clearly serve as powerful predictors of an employee’s job satisfaction.
However, whilst their overall influence is evident, their mitigating effect on other socio-demographic
control variables is limited to the impact of income, which is no longer statistically significant at the
conventional levels. Above all, the inclusion of personality traits in the regression model does not
alter the impact of past unemployment vs. future unemployment risk. The effect of unemployment
experienced in the past remains statistically insignificant, although it is interesting to note that the
coefficient is now positive.5 The perceived risk of becoming unemployed during the next 12 months
continues to exhibit a strong and negative influence on how satisfied employees are with their job.
It follows that it is the perceived risk of future unemployment, rather than the experience of
unemployment in the past, which serves as a key driver behind the reduction in employees’ job
satisfaction. What is more, the results hold true even when several socio-demographic characteristics
and proxies for individual personality traits are controlled for.
5. Concluding Remarks
For several decades both, the experience of past unemployment and the risk of future unemployment
have been associated with detrimental effects on individuals’ psychological well-being and workplace
behaviour. Although it is possible that both effects operate simultaneously, the analysis in this paper,
focused on employees’ job satisfaction and utilising a set of cross-sectional data derived from the
European Social Survey 2006-07, reports on a strongly pronounced insecurity effect: anticipated
unemployment substantially reduces employees’ job satisfaction. Interestingly, including the
perceived risk of future unemployment as a separate predictor variable in ordinal probit regressions
relegates the experience of past unemployment to a statistically insignificant coefficient and thus
weakens the ‘scarring’ hypothesis. Net of socio-demographic background variables and several
proxies for individual personality traits, it is not unemployment experienced in the past, but the fear of
future unemployment, which accounts for the reduction in employees’ job satisfaction.
13
These findings have important implications for organisations and respective human resource
practitioners who are charged with minimising the detrimental impact on organisational performance.
This matters even more since it has been reported that the feeling of job insecurity may even arise in
environments that are objectively non-threatening (Rosenblatt and Ruvio 1996). Moreover, perceived
job insecurity develops expectations about an uncertain future, which may influence otherwise
affective responses to an employee’s work environment and generates requirements for assurances
and additional information. So note Laine et al (2009) that “the necessity of actively searching out
information in order to reduce uncertainty and to be able to control future possibilities places
additional demands on the employee, implying that social support is all the more important” (Laine,
van der Heijden, Wickström, Hasselhorn and Tackenberg 2009: 435). To this end, the present study
provides additional evidence in support of endeavours aimed at assisting those who believe that their
job may be under threat, including counselling, confidence-building exercises, workforce
consultations, advice on alternative employment opportunities within the firm, and informal support
from co-workers and supervisors. The latter activities should certainly be embraced, especially since
previous research has shown that support derived from others in the workplace environment can
contribute significantly in buffering individuals against job dissatisfaction and non-compliant job
behaviours when employees’ job security is at stake (e.g. Lim 1997).
Although the present study provides robust empirical evidence on the impact of experienced
and anticipated unemployment on employees’ job satisfaction, it is nevertheless important to bring
some data limitations to the reader’s attention. As mentioned earlier and without recapitulating at
length, the study is constrained by the absence of longitudinal data, which imposes design limitations
to following trends and changes over time. It follows that the analysis of cross-sectional data cannot
rely on fixed-effect estimations to control for time-invariant factors. Moreover, although several
personality traits have been controlled for, the variables used represent only broad proxies for these
traits, are thus admittedly imperfect and consist of single-item variables rather than multi-item
constructs. It is also noteworthy that self-reported data have been analysed, which have been shown to
suffer from several biases. These constraints need to be unreservedly acknowledged, but we can take
comfort from observations by Schimmack and Oishi (2005) who note that self-reported measures of
wellbeing possess adequate validity and reliability. Setting aside the vexed issue of self-reported data,
the more general issue of the use of single-item measures of complex attitude structures, such as job
satisfaction, remains a controversial one, as such measures tend to have only marginally acceptable
internal consistency. On a positive note, the meta-analysis of US data sets by Wanous, Reichert and
Hudy (1997) gives the use of single–item measures a cautious thumps-up. Rose (2005) raises similar
concerns on the use of single-item measures, but he also adopts a more pragmatic attitude towards the
use of such measures and proceeds with his analysis of employee despondency in the UK.
In terms of scope for future research and assuming the availability of detailed information, an
analysis of data sub-divided by different unemployment frequencies and different unemployment
14
spells, unavailable in the current data set, may provide useful insights. Replicating the study with
longitudinal data sets and multi-item personality constructs, or with different job satisfaction domains
and extending the analysis to include a larger set of countries or regions with more salient personal
trait variations would also provide interesting avenues for future scholarly endeavours. Finally, as
speculated by reference to the results reported in Table 2, it is conceivable that the observed effects of
past unemployment are transmitted through perceived job insecurity. Put differently, the perception of
future unemployment may represent an intervening variable in the relationship between past
unemployment and employees’ job satisfaction. In such a scenario, future unemployment would serve
as a mediator. This arguably deserves separate examinations, and in an extension of the present
assessment future studies may wish to formally explore this possibility.
Acknowledgement
The analysis in this study is based on unit record data, derived from the third wave of the European
Social Survey (ESS) and collected during the period 2006–07. The usual disclaimer applies.
Notes
1. On this occasion, the terms ‘job satisfaction’ and ‘happiness at work’ are used interchangeably.
This is merely an editorial simplification.
2. Data from the 2006-07 European Social Survey (ESS) were publicly released in April 2008. For a
description of the sampling design, see Lynn, Hader, Gable and Laaksonen (2004). For further
information, including questionnaire design details see www.europeansocialsurvey.org.
3. For a brief survey of this research, see Lucas (2007).
4. In the present study, full-time employees are defined as respondents in 30 hours or more of paid
employment per week for their main job.
5. Knabe and Rätzel (2011), using German longitudinal data to explore the impact of unemployment
on life satisfaction, uncover similar results when they control for individual heterogeneity.
Speculatively, they suggest that this may occur because finding a job after a long spell of
unemployment could be surprising and may have a particularly favourable impact on satisfaction
ratings.
15
References
Appelbaum, S.H., Deguire, K.J. and Lay, M. 2005), “The relationship of ethical climate to deviant
workplace behaviour”, Corporate Governance, 5(4): 43-55.
Blau, P. (1964), Exchange and power in social life, New York: Wiley.
Blaug, M. (1985), Economic Theory in Retrospect, Cambridge: Cambridge University Press.
Cavanaugh, M.A. and Noe, R.A (1999), “Antecedents and Consequences of Relational Components
of the New Psychological Contract”, Journal of Organizational Behavior, 20: 323-340.
Chen, G., Gully, S.M. and Eden, D. (2004), “General self-efficacy and self-esteem: toward theoretical
and empirical distinction between correlated self-evaluations”, Journal of Organizational Behavior,
25: 375-395.
Cheng, G. H.-L. and Chan, D. K.-S. (2008), “Who Suffers More from Job Insecurity? A Meta-analytic
Review”, Applied Psychology, 57(2): 272-303.
Chirumbalo, A. and Hellgren, J. (2003), “Individual and Organizational Consequences of Job
Insecurity: A European Study”, Economic and Industrial Democracy, 24(2): 217-240.
Clark, A.E. (1997), “Job satisfaction and gender: why are women so happy at work?”, Labour
Economics, 4: 341-372.
Clark, A.E., Diener, E., Georgellis, Y. and Lucas, R.E. (2008), “Lags and Leads in Life Satisfaction: a
Test of the Baseline Hypothesis”, Economic Journal, 118(529): F222-F243.
Clark, A.E., Georgellis, Y., Lucas, R.E., and Diener, E. (2004), “Unemployment alters the set point
for life satisfaction”, Psychological Science, 15: 8–13.
Clark, A., Georgellis, Y. and Sanfey, P. (2001), "Scarring: The psychological impact of past
unemployment", Economica, 68(2): 221-241.
Clark, A.E., Oswald, A. and Warr, P. (1996), “Is job satisfaction u-shaped in age?”, Journal of
Occupational and Organizational Psychology, 69(1): 57-81.
Clark, L.A. and Watson, D. (1991), “General affective dispositions in physical and psychological
health”, in C.R. Snyder and D.R. Forsyth (eds.), Handbook of clinical and social psychology, New
York: Pergamon Press, pp. 221-245.
Cotton, J. and Tuttle, J. (1986), “Employee turnover: A metaanalysis and review with implications for
research”, Academy of Management Review, 39: 949–969.
De Witte, H. (1999), “Job Insecurity and Psychological Well-being: Review of the Literature and
Exploration of Some Unresolved Issues”, European Journal of Work and Organizational Psychology,
8(2): 155-177.
Dekker, S. and Schaufeli, W. (1995), “The effects of job insecurity on psychological health and
withdrawal: A longitudinal study”, Australian Psychologist, 30: 57-63.
Eisenberg, P. and Lazarsfeld, P. (1938), The Psychological Effects of Unemployment”, Psychological
Bulletin, 35: 358-390.
16
Forrier, A. and Sels, L. (2003), “The concept employability: a complex mosaic”, International
Journal of Human Resources Development and Management, 3(2): 102-124.
Gagné, M. and Bhave, D. (2011), “Autonomy in the workplace: an essential ingredient to employee
engagement and well-being in every culture”, in V.I Chirkov, R.M. Ryan and K.M. Sheldon (eds.),
Human Autonomy in Cross-cultural Context, Dordrecht: Springer, pp. 163-187.
Gazioglu, S. and Tansel, A. (2006), “Job satisfaction in Britain: Individual and job related factors”,
Applied Economics, 38: 1163–1171.
Georgellis, Y. and Lange, T. (2007), “Participation in continous, on-the-job training and the impact on
job satisfaction: longitudinal evidence from the German labour market”, International Journal of Human
Resource Management, 18(6), 969-85
Headey, B. (2008), “Life Goals Matter to Happiness: A Revision of Set-Point Theory”, Social
Indicators Research, 86: 213–231.
Headey, B., & Wearing, A. (1992), Understanding happiness: A theory of subjective well-being,
Melbourne: Longman Cheshire.
Hellgren, J. and Sverke, M. (2003), “Does job insecurity lead to impaired well-being or vice versa?
Estimation of cross-lagged effects using latent variable modelling”, Journal of Organizational
Behavior, 24(2): 215-236.
Jahoda, M. (1981), “Work, employment and unemployment: Values, theories and approaches in social
research”, American Psychologist, 36: 184–191.
Jahoda, M. (1988), “Economic recession and mental health: some conceptual issues”, Journal of
Social Issues, 44: 13-23.
Judge, T.A., Locke, E.A. and Durham, C.C. (1997), “The dispositional causes of job satisfaction: A
core evaluations approach”, Research in Organizational Behavior, 19: 151-188.
Judge, T.A., Locke, E.A., Durham, C.C. and Kluger, A.N. (1998), “Dispositional Effects on Job and
Life Satisfaction: The Role of Core Evaluations”, Journal of Applied Psychology, 83(1): 17-34.
Judge, T. A., Thoresen, C., Bono, J. and Patton, G. (2001), “The job satisfaction–job performance
relationship: A qualitative and quantitative review”, Psychological Bulletin, 127: 376–407.
Judge, T.A., Thoresen, C.J., Pucik, V. and Welbourne, T.M. (1999), “Managerial coping with
organizational change: a dispositional perspective”, Journal of Applied Psychology, 84: 107–122.
Kaiser, L.C. (2007), “Gender-job satisfaction differences across Europe: An indicator for labour
market modernization”, International Journal of Manpower, 28(1): 75-94.
Kalimo, R., Taris, T.W. and Schaufeli, W.B. (2003), “The Effects of Past and Anticipated Future
Downsizing on Survivor Well-Being: An Equity Perspective”, Journal of Occupational Health
Psychology, 8(2): 91-109.
Knabe, A. and Rätzel, S. (2011), “Scarring or Scaring? The Psychological Impact of Past
Unemployment and Future Unemployment Risk”, Economica, 78: 283-293.
Laine, M. , van der Heijden, B.I.J.M. ,Wickström, G., Hasselhorn, H.-M. and Tackenberg, P. (2009),
“Job insecurity and intent to leave the nursing profession in Europe”, International Journal
of Human Resource Management, 20(2):420-438.
17
Lange, T. (2009), “Attitudes, Attributes and Institutions: Determining Job Satisfaction in Central and
Eastern Europe”, Employee Relations, 31(1): 81-97.
Lange, T., Pacheco, G. and Shrotryia, V.K. (2010), “Culture, Industrialisation and Multiple Domains
of Employees’ Job Satisfaction: A Case for HR Strategy Redesign in India”, International Journal of
Human Resource Management, 21(13): 2438-2451.
Lazarus, R.S and Folkman, S. (1984), Stress Appraisal and Coping. New York: Springer.
Leana, C.R. and Feldman, D.C. (1994), “The psychology of job loss”, in G.R. Ferris (Ed.),
Research in Personnel and Human Resources Management, Vol. 12, Greenwich, CT: JAI Press, pp.
271-302.
LeBlanc, M.M. and Barling, J. (2004), “Workplace Aggression”, Current Directions in Psychological
Science, 13(1): 9-12.
Lim, V.K.G. (1997), “Moderating Effects of Work-Based Support on the Relationship between Job
Insecurity and its Consequences”, Work and Stress, 11: 252-266.
Locke, E.A. (1976), “The Nature and Causes of Job Satisfaction”, in M.D. Dunnnette (Ed.),
Handbook of Industrial and Organizational Psychology, Chicago, IL: Rand McNally, pp. 1297–1349.
Locke, E.A., McClear, K. and Knight, D. (1996), “Self-esteem and work”, International Review of
Industrial and Organizational Psychology, 11: 1-32.
Lucas, R.E. (2007), “Adaptation and the Set-Point Model of Subjective Well-Being: Does Happiness
Change After Major Life Events?”, Current Directions in Psychological Science, 16(2): 75-79.
Lynn, P., Hader, S., Gable, S., and Laaksonen, S. (2004), Methods for achieving equivalence of
samples in crossnational surveys: The European Social Survey Experience, ISER working paper
2004–09, University of Essex, Colchester.
Mallinckrodt, B. (1990), “Satisfaction with a new job after unemployment: Consequences of job loss
for older professionals”, Journal of Counseling Psychology, 37(2): 149-152.
Martin, G., Staines, H. and Pate, J. (1998), “Linking job security and career development in a new
psychological contract”, Human Resource Management Journal, 8: 20–40
McKee-Ryan, F.M., Song, Z., Wanberg, C.R. and Kinicki, A.J. (2005), “Psychological and physical
well-being during unemployment: A meta-analytic study”, Journal of Applied Psychology, 90: 53-76.
Paul, K.I. and Moser, K. (2009), “Unemployment impairs mental health: Meta-analyses”, Journal of
Vocational Behavior, 74: 264-282.
Probst, T.M., Stewart, S.M, Gruys, M.L. and Tierney, B.W. (2007), “Productivity, counterproductivity
and creativity: The ups and downs of job insecurity”, Journal of Occupational and Organizational
Psychology, 80(3): 479-497.
Robinson, S., Murrells, T. and Clinton, M. (2006), “Highly qualified and highly ambitious:
implications for workforce retention of realising the career expectations of graduate nurses in
England”, Human Resource Management Journal, 16(3): 287–312.
Rose, M. (2005), “Job Satisfaction in Britain: Coping with Complexity”, British Journal of Industrial
Relations, 43(3): 455-467.
18
Rosenblatt, Z. and Ruvio, A. (1996), “A Test of a Multidimensional Model of Job Insecurity: The
Case of Israeli Teachers”, Journal of Organizational Behavior, 17: 587-605.
Saari, L. and Judge, T. (2004), “Employee attitudes and job satisfaction”, Human Resource
Management, 43(4): 395-407.
Sahdev, K. (2003), “Survivors’ reaction to downsizing: the importance of contextual factors”, Human
Resource Management Review, 13(4): 56-74.
Schimmack, U. and Oishi, S. (2005), “The influence of chronically and temporarily accessible
information on life satisfaction judgments”, Journal of Personality and Social Psychology, 89: 395-
406.
Vila, L.E. and Garcia–Mora, B. (2005), “Education and the Determinants of Job Satisfaction”,
Education Economics, 13(4): 409-425.
Wanous, J. P., Reichers, A. E. and Hudy, M. J. (1997), “Overall job satisfaction: How good are
single-item measures?”, Journal of Applied Psychology, 82: 247–52.
Wood, J.V., Heimpel, S.A. and Michela, J.L. (2003), “Savoring Versus Dampening: Self-Esteem
Differences in Regulating Positive Affect”, Journal of Personality and Social Psychology, 85(3): 566-
580.
19
Table 1: Dispositional Variables
Dispositional Category Proxy Variables
Self-Esteem
‘I feel what I do in life is valuable’
Generalized Self-Efficacy
‘I am free to decide how to live my life’
Locus of Control
‘I feel that I am allowed to decide how daily work is organised’
Neuroticism
‘I felt anxious’ (how often past week?)
External Core Evaluations
- ‘Most people can be trusted’ vs. ‘can’t be too careful when trusting
people’
- ‘It is hard to be hopeful about the future of the world’
Note: The choice of dispositional categories is based on the ‘core evaluation’ concept by Judge,
Locke, Durham and Kluger (1998).
Table 2: Past Unemployment and Perceived Risk of Future Unemployment
Unemployment and work seeking (min. spell: 3 months)
during last 5 years
Perceived risk of becoming
unemployed in next 12 months
Yes
No
Very likely
12.7%
4%
Likely
20%
9.3%
Not very likely
41%
41%
Not at all likely
26.3%
45.6%
Observations 1412 1999
Source: Samples derived from ESS 2006-07; author’s own calculations.
20
Table 3: Job satisfaction – ordered probit regressions
(1)
Past unemployment only
(2)
Past unemployment & future unemployment risk
(3)
Inter-temporal unemployment & personality traits
Past Unemployment -.106*** (0.038) -.005 (0.039) .042 (0.040)
Future Unemployment Risk
- ‘Very likely’ - -.738*** (0.073) -.603*** (0.074)
- ‘Likely’ - -.575*** (0.058) -.464*** (0.060)
- ‘Not very likely’ - -.271*** (0.041) -.229*** (0.041)
Background Controls
Male -.029 (0.034) -0.025 (0.035) -.0.008 (0.035)
Age -.025** (0.012) -.029** (0.013) -.035*** (0.013)
Age2 .000*** (0.0002) .000*** (0.001) .000*** (0.001)
Married .017 (0.037) .018 (0.038) .016 (0.039)
Health .183*** (0.023) .162*** (0.024) .110*** (0.025)
Education
- ‘Middle’ -.044 (0.048) -.058 (0.049) -.092* (0.050)
- ‘Upper’ -.097** (0.053) -.127** (0.054) -.229*** (0.056)
Ongoing education/training opportunities past 12 months
.175*** (0.038) .143*** (0.038) .072** (0.039)
Household income
- ‘Middle’ .201*** (0.064) .136** (0.065) .086 (0.066)
- ‘Upper’ .268*** (0.075) .171** (0.076) .108 (0.077)
Personality Trait Proxies
‘I felt anxious’ (how often past week?)
- - -.005 (0.065)
‘I feel what I do in life is valuable’
- - .529*** (0.050)
‘I am free to decide how to live my life’
- - .186*** (0.044)
‘I feel I am allowed to decide how daily work is organised’
- - .057*** (0.006)
‘Most people can be trusted vs. can’t be too careful’
- - .039*** (0.009)
‘It is hard to be hopeful about the future of the world’
- - -.097*** (0.038)
Country dummies Yes Yes Yes
Observations 3,411 3,411 3,411
Log Likelihood 14779.763 14113.255 13479.895
Note: (***) = significant at the 1 percent level, (**) = significant at the 5 percent level, and (*) = significant at the 10 percent level; robust standard errors in parentheses. Reference categories: educational attainment (lower); household income (lower); future unemployment risk (not at all likely).
21
Appendix: Variable definitions and sample means
Variable Definition
Means
Job satisfaction Ordinal categorical variable on a scale 0 to 10 (0=extremely dissatisfied, 10=extremely satisfied) 6.82
Socio-demographic Background
Male Dummy variable: 1 for male; 0 otherwise 0.50
Age Age in years 40.87
Married Dummy variable: 1 for married; 0 otherwise 0.49
Education Ordinal categorical variable on a scale of 1 to 3 (1 = lower; 3 = upper) 2.07
Ongoing education/training opportunities in past 12 months Dummy variable: 1 for having opportunity in past 12 months; 0 otherwise 0.52
Income Ordinal categorical variable on a scale of 1 to 3 (1 = lower; 3 = upper) 2.05
Health Dummy variable: 1 for health self-rated as good or very good; 0 otherwise 0.74
Inter-temporal Unemployment
Past Unemployment Dummy variable: 1 if unemployed within past 5 years; 0 otherwise 0.41
Future Unemployment Risk Ordinal categorical variable on a scale of 1 to 4 (1 = not likely at all; 4 = very likely) 1.90
Personality Traits
‘Felt anxious’ (how often past week?) Dummy variable: 1 for ‘Most of the time’, or ‘All or almost all of the time’; 0 otherwise 0.09
‘Feel what I do in life is valuable’ Dummy variable: 1 for ‘Agree’ or ‘Strongly agree’; 0 otherwise 0.83
‘Free to decide how to live my life’ Dummy variable: 1 for ‘Agree’ or ‘Strongly agree’; 0 otherwise 0.78
‘It is hard to be hopeful about the future of the world’ Dummy variable: 1 for ‘Agree’ or ‘Strongly agree’; 0 otherwise 0.47
‘Allowed to decide how daily work is organised’ Ordinal categorical variable on a scale of 1 to 10 (1 = I have no influence; 10 = I have complete control) 6.23
‘Most people can be trusted vs. can’t be too careful’ Ordinal categorical variable on a scale of 0 to 10 (0 = can’t be too careful; 10 = most people can be trusted) 5.25
Observations
3,411