Int. Journal of Business Science and Applied Management, Volume 16, Issue 2, 2021
Developmental Proactivity and Professional Ability as Older
Workers’ Employability Resources: A Longitudinal Study
Explaining Career Events
Conny Roobol
Erasmus School of Social and Behavioural Sciences, Erasmus University Rotterdam
Burgemeester Oudlaan 50, Rotterdam, 3062PA, the Netherlands
Tel: +31 10 408 2071
Email: roobol@essb.eur.nl
Hans Pruijt*
Erasmus School of Social and Behavioural Sciences, Erasmus University Rotterdam
Burgemeester Oudlaan 50, Rotterdam, 3062PA, the Netherlands
Tel: +31 10 408 2071
Email: pruijt@essb.eur.nl
Ferry Koster
Erasmus School of Social and Behavioural Sciences, Erasmus University Rotterdam
Burgemeester Oudlaan 50, Rotterdam, 3062PA, the Netherlands
Tel: +31 10 408 2231
Email: koster@essb.eur.nl
Fenna Ruby Marie Leijten
Erasmus School of Health Policy & Management, Erasmus University Rotterdam
Burgemeester Oudlaan 50, Rotterdam, 3062PA, the Netherlands
Tel: +31 10 408 8555
Email: fenna.leijten@gmail.com
Abstract
This four-year longitudinal study examines how two facets of employability—professional ability and developmental
proactivity—are linked to career events among workers ages 45 years and older. We construe employability as a
personal resource that predicts a higher likelihood of experiencing positive career events and a lower likelihood of
experiencing negative ones. Results reveal that developmental proactivity leads to a higher probability of internal
promotion, while professional ability leads to a lower probability of salary loss, demotion, and unemployment. The
findings indicate that these two facets of employability can offer critical insights for understanding the career events
of older workers.
Keywords: employability, personal resources, career resources, longitudinal design, older workers
Acknowledgements:
This work was financially supported by the Netherlands Organization for Scientific Research,
TOP Grant “Sustaining Employability” (NWO 407–13-021).
Int. Journal of Business Science and Applied Management/ Business-and-Management.org
65
1. INTRODUCTION
With the ageing of the workforce it is becoming ever more crucial that older workers can sustain their careers
and keep them exciting, or at least satisfying. As such, these workers have good reason to focus on their job tenure
and internal opportunities within their current workplaces because of their precarious position in the external labour
market (Van der Heijden et al., 2009; Van Harten, 2016; Van Veldhoven & Dorenbosch, 2008).
Ideas for enhancing career sustainability centre on the concept of employability, which HRM gurus and policy
makers tend to view as amenable to substantial improvement through investments (Pruijt & Dérogée, 2010).
Employability is a broad concept that spans a range of resources. One of these resources is proactivity, especially in
the form of pursuing learning opportunities (termed developmental proactivity). An additional resource is the level of
the ease and confidence with which workers perform their day-to-day work (professional ability). In this study, we
focus on these two resources and assess their impact on career development, while considering both positive and
negative career outcomes.
The gap in the employability literature that this paper addresses concerns problematic employability. The
existing literature is mainly brimming with positivity and frames employability in the light of opportunities that
employees need to grab. Thus, substantial research exists that shows how promotions and salary gains depend on
conforming to an ideal of proactive attitude and behaviour (Ng et al., 2005; Raemdonck et al., 2012; Van der Heijden
et al., 2009; Volmer & Spurk, 2011). A small number of studies provide a more balanced picture by including
negative career events among the possible outcomes, specifically the loss of job security or becoming unemployed.
However, these studies also choose the predictors from the realm of ideal attitudes and behaviours, especially
proactivity and personal adaptability (De Cuyper, Mäkikangas et al., 2012; Van der Heijde &Van der Heijden, 2006).
These studies do not consider that this may be a luxury, and that workers’ employability could depend solely on
being able to confidently keep up in the job.
This study followed the career development of a large group of older workers (ages 45 to 64) during a four-year
period (2010–2013). The longitudinal design allowed for more insight into time-effects and causality, and thus adds
to previous research, which is mostly cross-sectional (e.g. De Vos et al., 2011; Van der Heijde & Van der Heijden,
2006; Van der Heijden et al., 2009; Volmer & Spurk, 2011).
2. THEORETICAL BACKGROUND—CONSERVATION OF RESOURCES THEORY
Employability is empowerment in matters of career development. This definition follows Gazier (2001, p. 23),
who sees employability as what makes ‘each worker a more aware and a more independent organizer of the
succession of activities and commitments that, combined, constitute his/her working life.’
In this study, we focus on the supply side of employability—the dispositions, competencies, and attitudes that
workers bring to their organizations and the labour market (De Cuyper, Mäkikangas et al., 2012; Forrier & Sels,
2003; Fugate et al., 2004; Hillage & Pollard, 1998; Thijssen et al., 2008; Van der Heijde & Van der Heijden, 2006).
In contrast, the demand side of employability includes such factors as sectoral or governmental arrangements (Pruijt,
2013; Pruijt & Dérogée, 2010) or organizational policies to accommodate the work-life balance (Xanthopoulou et al.,
2009).
Conservation of resources (COR) theory distinguishes between resource conservation and resource acquisition.
The key scenario in resource conservation is resources circumvent resource losses (Hobfoll, 1989; Hobfoll, 2002).
The individual with resources is less prone to resource losses and is able to invest in the conservation of those
resources (Hobfoll et al., 2018). Conversely, when individuals engage in resource acquisition, the scenario is
resources beget resources. The individual is thriving and able to invest in the acquisition of even more resources
(Hobfoll, 2002).
A central employability resource which makes the acquisition of more resources possible is proactivity (Eby et
al., 2003; Fugate et al., 2004; Van der Heijde & Van der Heijden, 2006), more specifically developmental
proactivity. This is the worker’s motivation to learn new things, build relationships with colleagues, explore ways to
improve the work (Dorenbosch, 2014), and assess future skill requirements (Van Veldhoven & Dorenbosch, 2008).
Developmental proactivity overlaps with several other psychological and social resources, such as the propensity to
learn (Fugate et al., 2004), openness to change and curiosity or “career adaptability” (Hirschi, 2012), the inclination
to build a (developmental) network (Arthur & Rousseau, 1995; Hirschi, 2012), and to some extent, self-management
orientation and ability (Savickas, 2000). Overall, proactivity does not seem to be related to age (Zacher & Kooi,
2017).
We view professional ability as an employability resource that workers are motivated to conserve (De Cuyper,
Raeder et al., 2012; Xanthopoulou et al., 2012). Crucially important is whether workers operate close to or beyond
Conny Roobol, Hans Pruijt, Ferry Koster and Fenna Ruby Marie Leijten
66
their abilities, or whether they have headroom to draw upon their intellectual, physical, and emotional resources to
meet their work demands (Hirschi, 2012). With such headroom, older workers can invest in the conservation of their
resource pools. Conversely, workers who struggle in their jobs may lose personal resources and fail to replenish
them. Then, any endeavour becomes an uphill battle (Trougakos et al., 2015). Workers’ professional ability can
decline with age, to the extent that their abilities depend on physical, mental, and sensory capacities rather than on
experience, expertise, and soft skills; however, this decline is subject to individual differences, of course (Hardy &
Reyes, 2015).
Because resource conservation and resource acquisition are separate processes (Halbesleben et al., 2014),
professional ability and developmental proactivity can be viewed as separate employability resources. Another
reason for making a sharp distinction between professional ability and developmental proactivity is that we sought to
avoid the implicit assumption that proactivity is the key to being able to thrive sustainably in the world of work. Such
an underlying implicit assumption seems to exist in much of the employability literature, which promotes a variety of
proactive behaviours: assessing skill sets, acting on assessments by honing, redirecting, and expanding skills
(Waterman et al., 1994), becoming more aware of the marketability of one’s skills and presenting them in an
effective way (Ghoshal et al., 1999), network building (Arthur & Rousseau, 1995), creating one’s own opportunities
in an entrepreneurial fashion (Kanter, 1993), and job crafting (Dorenbosch, 2014).
The concept of a career event refers not only to “career” in the sense of a possible progression (Hirschi, 2012),
but also to the effects of employment shocks, offshoring, and downsizing (Bosch & Ter Weel, 2013; Ng & Feldman,
2014). This view deviates from the “boundaryless” career perspective, which emphasizes mobility between
companies while the workers’ personal development takes centre stage and employers become contingent and
replaceable (Littleton et al., 2000). The boundaryless career perspective is less applicable to older workers (for
empirical evidence, see Appendix 1).
We focus entirely on what Spurk et al. (2019) called ‘objective career success’. Thus, this study covers positive
and negative career events or outcomes. Promotion is a positive career event, and the study covers four possible
negative career events: salary loss, demotion, salary loss combined with demotion, and transition into
unemployment.
The linkage between employability resources and career events consists of several pathways. Employability
resources can affect performance (De Cuyper et al., 2014). Moreover, the sheer amount of resources in one’s
possession affects others’ judgments (Hobfoll, 2002). For (older) workers, the judgments of their immediate
managers play a major role, since monitoring employees is part of a manager’s official duties. Managers’ judgments,
in turn, inform decisions that determine whether or not various career events occur (Van der Heijden et al., 2009). A
third mechanism is the resource crossover between interacting individuals (Hobfoll et al., 2018), which implies that a
worker’s resources affect those enjoyed by co-workers, and in turn can influence personnel management decisions.
3. HYPOTHESIS DEVELOPMENT
Hypothesis 1 concerns the likelihood that older workers will experience an internal promotion. Performance
appraisals may be pervasive in managers’ decisions to grant their subordinates (new) career entitlements (Van der
Heijden et al., 2009).
We reason that resource-rich workers, those who favourably assess their professional ability and developmental
proactivity, will project a corresponding image. Managers value resource-rich workers. Promotions are widely
recognized as proximal indicators of workers’ assessed net worth or typically classified as tangible rewards for
valuable organizational assets (Ng et al., 2005).
The idea that employability affects employees’ promotion rates is well-accepted. For instance, Van der Heijde
and Van der Heijden (2006) along with Van der Heijden et al.
(2009) and Volmer and Spurk (2011) posited that
employability predicts a higher likelihood of experiencing an (internal) promotion. Likewise, Raemdonck et al.
(2012) related self-directedness in learning (which displays overlap with developmental proactivity) to low-qualified
employees’ chances for a promotion. Proceeding from these considerations, we hypothesize:
Hypothesis 1: Self-perceived employability (professional ability and developmental proactivity) is positively
related to the likelihood of experiencing a future internal promotion.
Hypotheses 2 through 5 concern negative career events in the form of a salary loss, demotion, and transition
into unemployment.
Workers who confidently meet work demands (who possess high levels of professional ability) are less likely to
be judged as underperforming employees, and, based on performance at least, there is no reason for a salary loss to
occur. Likewise, workers who exhibit proactive learning behaviour should be able to minimize the risk of a salary
Int. Journal of Business Science and Applied Management/ Business-and-Management.org
67
loss because they can anticipate conditions in which their activities may be less valuable to the organization and,
therefore, can prepare for new, more important activities.
Contemporary research has stringently focused on the salary level or gain. For instance, Van der Heijden et al.
(2009) along with Van der Heijde and Van der Heijden (2006) and Volmer and Spurk (2011) posited that being
employable is associated with higher salaries. The outcome of salary loss is relatively absent from the literature.
Based on these considerations, we propose:
Hypothesis 2: Self-perceived employability (professional ability and developmental proactivity) is negatively
related to the likelihood of experiencing a future salary loss.
Workers with ample resources—even resources to spare—to perform their work adequately and who learn
proactively are likely to win managers’ confidence (Van der Heijden et al., 2009). Managers are not likely to
perceive such workers as bad decision makers or inadequate performers, which could be reasons for a forced
demotion. Also, workers who are in control and thriving seem to have little reason to consider relinquishing their job
responsibilities (voluntary demotion). Also relevant is the notion that the leadership exercised by a resourceful
individual can engender resource increases in colleagues (Hobfoll et al., 2018; Ma et al., 2019). Thus, demoting such
a person would entail a disadvantage for the organization in the form of a decline in resources or potential new
resources among a group of workers. Hence, we hypothesize:
Hypothesis 3: Self-perceived employability (professional ability and developmental proactivity) is negatively
related to the likelihood of experiencing a future demotion.
A loss in salary and a loss of job responsibility (demotion) can occur together. In the Dutch HRM literature
there are indications that demotion combined with a reduction in salary tends to be a management intervention
directed at a structurally underperforming employee. Such action is the outcome of a process triggered when a
worker is no longer able to meet the minimally required performance level and can be seen as an alternative to
dismissal (De Rooij, 2014; Jongsma, 2014). A vignette study by Van Dalen and Henkens (2015) suggested that
managers employ this procedure reluctantly, because workers across the organization will see it as the employer’s
breaking the agreed-upon rules and acting untrustworthily. Thus, we assume that workers need only relatively
modest employability resources in order to prevent the procedure from being triggered. The corollary is that workers
who experience both salary loss and demotion have lower professional ability and developmental proactivity than
those who experience only one of these negative career events. Accordingly, we predict:
Hypothesis 4: The negative association between self-perceived employability (professional ability and
developmental proactivity) and future negative career events is stronger when salary loss and demotion occur
concurrently.
Finally, we posit that self-perceived employability (professional ability and developmental proactivity) affects
the likelihood of becoming unemployed. Employable workers possess qualities that managers seek; therefore,
resource-rich workers have a higher chance of retaining employment compared with resource-poor workers. A
number of studies show that a crossover of resources can occur in the course of interpersonal exchange in teams
(Bakker et al., 2006; Hobfoll et al., 2018). In this way, resource-rich employees create extra value for the
organization, which could imply some protection against dismissal.
Van der Heijde and Van der Heijden (2006) posited that employable employees are less likely to be confronted
with periods of unemployment. De Cuyper, Mäkikangas et al., (2012) proposed that perceived employability reduces
job insecurity. Accordingly, we posit:
Hypothesis 5: Self-perceived employability (professional ability and developmental proactivity) is negatively
related to the likelihood of experiencing future unemployment.
Figure 1 displays our theoretical framework.
Conny Roobol, Hans Pruijt, Ferry Koster and Fenna Ruby Marie Leijten
68
Figure 1 The proposed theoretical framework
4. METHODOLOGY
4.1 Sample
We based this study on the Dutch Study on Transitions in Employment, Ability and Motivation (STREAM), a
prospective cohort study with a four-year time span conducted by The Netherlands Organisation for Applied
Scientific Research or TNO (Ybema et al., 2014). The sample is stratified by age (45 to 64 years at baseline) and
employment status (employed, self-employed and non-employed or those inactive in the labour market).
For the baseline measurement, the Dutch panel agency, Intomart Gfk, invited 26,601 individuals in 2010 to
complete an online questionnaire on topics such as proactivity, health, work, and labour market transitions.
Respondents received a small financial incentive, which increased slightly each year the respondent completed the
questionnaire, depending on the time spent. The final sample at baseline included 15,118 respondents, a response
rate of 71%, which is above average for survey research conducted among individuals (Baruch & Holtom, 2008).
Respondents who participated in all follow-up waves (conducted in 2011, 2012, and 2013) numbered 9,639, or 64%
of the baseline number.
Respondents in STREAM do not constitute a representative sample of the Dutch population; however, we do
not consider this problematic because our aim is to draw causal conclusions about the relationship between
employability and career events rather than to determine the proportion of older workers in the Dutch population
who have experienced particular career events. Therefore, for our purpose, heterogeneity is preferable to
representativity. A flowchart about the responses to the STREAM questionnaire is found in Appendix II.
To test hypotheses 1 through 4, we used a sub-dataset that enabled us to assess the likelihood of experiencing a
career event (promotion, salary loss, demotion or combination of salary loss and demotion) rather than sustained
“stability” (remaining employed throughout the study without experiencing any of these career events). Since we
were interested in respondents’ career development at a later point in time, we focused on respondents who
participated at the baseline and at least one follow-up wave (N=13,781).
Next, we excluded self-employed persons, since they may have different career trajectories and thus score
differently on employability resources (remaining N=12,573). We also excluded non-employed persons, because we
wanted to concentrate on career events occurring while working relative to no event occurring (remaining N=8,666).
Finally, we omitted those respondents who had experienced a career event in the 12 months prior to the baseline
measurement to ensure a stable and equal starting position for all. The final sample included 7,751 respondents.
To test hypothesis 5, we relied upon a second sub-dataset. This sub-dataset is comparable to the first in that all
self-employed persons were removed and only those who participated in at least two waves were included. As a
result, only respondents classified as ‘employed’ or ‘non-employed’ who participated at the follow-up were part of
Employability:
Developmental Proactivity
Professional ability
Promotion
Circumventing salary loss
Circumventing demotion
Circumventing salary
loss + demotion
Circumventing
unemployment
H2
H3
H4
H5
H1
Int. Journal of Business Science and Applied Management/ Business-and-Management.org
69
our sample (N=12,573). Next, we removed respondents with a mixed employed/non-employed status at the same
point in time to avoid biases in our estimates. We wanted to compare those who had remained employed with those
who had made the transition to unemployment at follow-up (remaining N=10,027). In addition, we removed
participants who were non-employed at the baseline, so the starting position was similar for all (i.e., everyone was
employed; remaining N=8,335). Finally, we excluded participants who had an inactive labour market status other
than unemployment at the follow-up (such as work disability). The final sample included 7,241 respondents.
Appendix III provides details of the participation pattern of respondents in both samples.
We performed drop-out analyses to compare the scores on our predictor variables of respondents who
participated only at the baseline with scores of those who participated in at least two measurement waves. Because
we employed two sub-datasets, we performed two drop-out analyses using T-tests and χ²-tests. We compared each
study sample (N=7,751 and N=7,241) to a sample of workers who took part only in the baseline questionnaire
(N=912 for sub-dataset 1; N=929 for sub-dataset 2
1
). Except for gender (only in sub-dataset 2), age, relaxation and
recovery, and professional ability, we did not find any serious loss to follow-up biases. Although significant, mean
differences remained rather small, in part due to the large sample sizes. In essence, the Cohen’s d effect size measure
was |.20| or lower in all cases
2
. Also, Φ was |.03| for gender. Based on these small to negligible effect sizes, we do
not consider loss to follow-up a problem when interpreting the results from both sub-datasets.
4.2 Measurements
For the dependent variables, we used factual questions, which seem less susceptible to self-selection and
common-method biases than perceptual measures.
Participants belonging to sub-dataset 1 were asked whether they had received a promotion in their company in
the past 12 months. The question was included in each measurement wave. Scale anchors were 1 (yes) and 2 (no).
Workers who answered ‘no’ were asked if they had moved from a higher to a lower position or from a higher to a
lower salary in the past 12 months. Respondents could choose one of four answers: 1 (no), 2 (yes, lower position and
lower salary), 3 (yes, lower position), 4 (yes, lower salary). These questions were collapsed into one variable with
five mutually exclusive categories at each follow-up wave: promotion, salary loss, demotion, salary loss and
demotion (combined), and sustained “stability” (those experiencing no career event).
Since we restricted our sample to those who had not experienced a career event in the 12 months prior to the
baseline measurement, four transition categories could be distinguished from baseline stability to promotion, salary
loss, demotion, and salary loss and demotion (combined) at follow-up, as well as a category encompassing those who
had not experienced a career event at follow-up. Based on these transitions, we created a final variable for sub-
dataset 1 that prioritized promotion over combined salary loss and demotion; combined salary loss and demotion
over salary loss (singular); and salary loss (singular) over demotion (singular). Thus, respondents were categorized as
“promoted” if they had experienced a promotion at least once at follow-up. We consider this hierarchical
classification valid, because older workers’ career patterns were quite consistent over time. In fact, only 10% had a
mixed participation pattern in which they had experienced two or more different career events at follow-up.
We classified participants in sub-dataset 2 as employed or non-employed (unemployment) based on the question
‘Are you currently’…, after which respondents could choose between nine possible choices of status. These choices
were not mutually exclusive. Participants were categorized as employed if they held at least one job with one or
multiple employers or combined a career as an employee with a career as an entrepreneur (self-employment), but
spent most hours as an employee. The non-employed were those who had lost their jobs and were thus fully
unemployed or were combining unemployment with another inactive status, such as studying or running the
household.
Our rationale for classifying the non-employed in the aforementioned way is that we were interested in whether
respondents became unemployed at follow-up (and hence, lost the jobs they had) rather than in how they spent their
time as an unemployed person. Nor could we not rule out the possibility that respondents who classify themselves as
‘unemployed’ also spent time on other activities outside the work domain.
As for the independent variables: for both datasets, we measured employability based on whether workers had
two resources—professional ability and developmental proactivity—which we assessed at the baseline.
1
Due to differences in the composition of our study samples (see heading ‘Datasets and participants’), the sizes of
the benchmark samples differed slightly.
2
According to Cohen (1977), values for Cohen’s d between .20 and .35 are indicative of a small effect size, while
values below .20 indicate effect sizes that can be considered negligible; however, Wolf (1986) considers an
effect meaningful when Cohen’s d is .25 or above.
Conny Roobol, Hans Pruijt, Ferry Koster and Fenna Ruby Marie Leijten
70
We measured professional ability using a scale developed by TNO which consists of five items that reflect
workers’ assessment of their reactive adjustment to future work-related changes. Example items are ‘Able to perform
the work in the coming 12 months’ and ‘Able to continue working if work becomes physically more demanding’.
Answers ranged from 1=certainly not to 5=certainly. Only respondents with a valid score on at least four items
contributed to the mean scale. Cronbach’s alpha was .85 in both samples.
We measured developmental proactivity using four items to assess workers’ judgement of their willingness to
learn new skills and proactive adjustment to future job requirements. The scale was originally developed by Van
Veldhoven & Dorenbosch (2008) and includes items such as ‘In my work, I keep trying to learn new things’ and
‘With regard to my skills and knowledge, I see to it that I can cope with changes in my work’. Response anchors
ranged from 1=strongly disagree to 5=strongly agree. Only respondents with a valid score on at least three items
contributed to the mean scale. Cronbach’s alpha was .81 in both samples, which is comparable to the coefficient
found by Van Veldhoven & Dorenbosch (2008).
Appendix IV lists the items used to operationalize professional ability. For an overview of the items used to
operationalize developmental proactivity, we refer the reader to the technical report on STREAM, which can be
found in the reference list (Van den Heuvel et al., 2014).
4.3 Analysis
The aforementioned employability resources have not yet been studied simultaneously; therefore, we first
performed an exploratory factor analysis on one random half of the samples for sub-dataset 1 (N=3,910) and sub-
dataset 2 (N=3,655) using baseline data. In both datasets, two factors with an eigenvalue greater than 1 emerged from
the principal component analysis with oblimin rotation. All cross-loadings fell below .32, while intended loadings
exceed .50 (mean λ value=.79). The average variance explained (AVE) amounted to .63 for both professional ability
and developmental proactivity (.64 for sub-dataset 2). The composite reliability (CR) was .90 for professional ability
in both datasets, and .87 and .88 respectively for developmental proactivity.
All these values lie above the recommended threshold of .50 for the AVE and .70 for the CR (Hair et al., 2010).
As a result, our factor analysis provides sufficient evidence for the convergent validity of our latent constructs.
Moreover, the square root of the AVE belonging to a particular construct was higher than the correlation between
that construct and the other employability resource, confirming the divergent validity of our measurement model.
To validate our measurement model, we performed a confirmatory factor analysis (CFA) on a second random
half of the samples for sub-dataset 1 (N=3,841) and sub-dataset 2 (N=3,586) using data assessed at baseline.
Specifically, we compared a hypothesized second-order factor model, in which professional ability and
developmental proactivity together with their corresponding items load onto a second-order factor, ‘employability
resources’, with two alternatives. The first alternative concerned a two-factor model in which the items
accompanying professional ability and developmental proactivity load onto their corresponding first-order factor.
The second alternative pertained to a one-factor model in which the items belonging to professional ability and
developmental proactivity load onto one first-order factor.
A CFA model is considered a good fit to the data if the CFI and TLI are .90 or higher, the RMSEA falls below
.08 and the SRMR is .10 or lower (Hox, 2010; Van den Broeck et al., 2010). Our results in R lavaan (Rosseel, 2012)
indicated that the hypothesized second-order factor model provided a good fit to the data with χ²(26) =352.77,
CFI=.98, TLI=.97, RMSEA=.06 and SRMR=.04 (sub-dataset 1) and χ²(26) =298.58, CFI=.98, TLI=.97,
RMSEA=.05 and SRMR=.04 (sub-dataset 2). However, the two-factor model also provided an acceptable fit. In
contrast, the one-factor model provided a very poor fit to the data (see Table 1). These results justify the conclusion
that professional ability and developmental proactivity should be treated as two conceptually distinct constructs,
which are apparently an integral part of a second-order factor, ‘employability resources’.
Int. Journal of Business Science and Applied Management/ Business-and-Management.org
71
Table 1: Confirmatory factor analysis: fit indices for the hypothesized model and two alternative models
χ² df CFI TLI RMSEA SRMR
Model 1: Second-
order factor model (hypothesized
model)
Sub-dataset 1 352.77 26 0.98 0.97 0.06 0.04
Sub-dataset 2 298.58 26 0.98 0.97 0.05 0.04
Model 2: 2 factor model
Sub-dataset 1 286.68 25 0.98 0.98 0.05 0.04
Sub-dataset 2 277.27 25 0.98 0.97 0.05 0.04
Model 3: 1 factor model (Harman's model)
Sub-dataset 1 5044.95 26 0.66 0.53 0.23 0.18
Sub-dataset 2 4319.59 26 0.68 0.55 0.22 0.18
Notes: N=3,801 for sub-dataset 1 / N=3,539 for sub-dataset 2.
χ²=Maximum Likelihood chi-square.
CFI=Comparative Fit Index; TLI=Tucker-Lewis Index; RMSEA=Root Mean Square Error of Approximation; SRMR=Standardized Root Mean Square Residual.
¹Since different techniques have been used to identify the models, the fit of the different measurement models could not be compared.
Conny Roobol, Hans Pruijt, Ferry Koster and Fenna Ruby Marie Leijten
72
Previous research indicates that demographic, well-being and work-related variables affect workers’
employment position and loss of paid employment (e.g. Leijten et al., 2015; Seibert et al., 1999; Van der Heijden et
al., 2009; Volmer & Spurk, 2011; Yang & Chau, 2016). Therefore, we included these potential confounders in our
analyses. Demographic variables included workers’ age (in years), gender (0=male; 1=female), and educational level
(dummies for low and middle; reference category is high). In addition, we used workers’ years of job tenure to
represent work-related aspects. Finally, as a proxy for well-being, we used relaxation and recovery—the degree to
which employees could detach from work after a working day (three items, α=.71 in both samples).
We performed two logistic regression analyses, one to address hypotheses 1 through 4 (using sub-dataset 1) and
one to address hypothesis 5 (using sub-dataset 2). In the first analysis, we used one outcome variable with multiple
categories in a multinomial logistic regression analysis. In the second analysis, we conducted a binary logistic
regression analysis of whether paid employment was lost or not. For both analyses, we report the odds ratios or
Exp(B) and the statistical significance levels.
For both logistic regression analyses, we estimated three models. In the first, we included the employability
resources. In the second, we added the control variables to the equation to address potential confounding
relationships. In the third, we included the variable ‘participation’ as part of a panel attrition analysis. Specifically,
we added a main term for participation as well as interaction terms with our employability resources to the equation
(see Model 3 in Tables 2c and 3c). In this way, we were able to assess how the frequency of participation affected
the outcome variables of interest and the assumed relationships between these outcomes and employability.
The aforementioned panel attrition analysis is analogous to the pattern mixture model frequently used in linear
mixed effect models (Son et al., 2012). We used the likelihood ratio test statistic (denoted by χ²) to assess the
improvement in model fit and relied on the AIC to adjust for model complexity. In addition, we used the Nagelkerke
R² as an analogue to the R² commonly used in linear regression analyses.
5. RESULTS
Tables 2a and 2b display the descriptive statistics for sub-dataset 1. The reported mean score for professional
ability was 4.20 and 3.91 for developmental proactivity, with little variation among the respondents. Our first sample
consisted of slightly more males than females, with an average age of 53, and mainly with medium and higher
education levels. Multi-collinearity was not a concern because correlations between the study variables (i.e.,
independent and control variables) fell below the threshold of .80 (Field, 2009). There was a small and positive
correlation between professional ability and developmental proactivity (r=.16, p<.01). Although these two resources
are conceptually distinct, this coefficient indicates that these resources are related, which aligns with their
classification as ‘employability resources’.
Table 2a: Descriptive results for sub-dataset 1: means, standard deviations and ranges for the independent
and control variables at baseline
Mean/proportion SD Range
Employability resources
Developmental proactivity
3.91
0.57
1-5
Professional ability
4.20
0.62
1-5
Demographic characteristics
Age (years)
53.31
5.04
45-64
Gender (ref=male)
0.44
0/1
Education (ref=high)¹
low
0.26
0/1
middle
0.39
0/1
Work-related characteristic
Job tenure (years)
11.19
9.51
0-45
Int. Journal of Business Science and Applied Management/ Business-and-Management.org
73
Well-being related characteristic
Relaxation and recovery
2.91
0.71
1-5
Notes: Due to missing values (< 1% of total), the N ranges from 7,705 to 7,751. ¹low: (not) finished
primary school or completed lower vocational education; middle: finished secondary school or higher
vocational training; high: obtained a bachelor, master or PhD degree.
Table 2b: Zero-order (Pearson r) correlation coefficients for the independent and control variables at baseline
(sub-dataset 1)
1. 2. 3. 4. 5. 6.
1. Developmental proactivity
-
2. Professional ability
.156**
-
3. Age (years)
-.035**
-.150**
-
4. Gender (ref=male)
.033**
-.033**
-.026*
-
5. Education¹
.186**
.053**
-.030**
-.056**
-
6. Job tenure (years)
-.059**
-.075**
.194**
-.061**
-.044**
-
7. Relaxation and recovery
-.123**
.081**
.017
-.072**
-.130**
.006
Notes: *p<0.05; **p<0.01. ¹Spearman correlation used.
Due to missing values (< 1% of total), the N fluctuates between 7,705 and 7,751.
Table 2c shows the results from the sequential multinomial logistic regression analyses for sub-dataset 1. Model
1, which contains only employability resources, provides a statistically significant improvement in fit over the
intercept-only model, with χ² (8)=113.1, p<.001 and AIC is 1754.3 (R²=.020). Model 2, which displays the full
model, shows a statistically significant better fit than Model 1, with χ² (24)=174.6, p<.001 and AIC is 9323.4
(R²=.052). Model 3, which contains the panel attrition analysis, shows negligible differences in comparison to our
full model (Model 2), indicating that respondents’ participation patterns did not bias our results.
Hypothesis 1 predicted that workers with a favourable self-assessment of their employability (professional
ability and developmental proactivity) would have a higher probability of experiencing a future internal promotion.
The estimated results shown in Table 2c support this hypothesis for developmental proactivity: respondents who
assessed their developmental proactivity more favourably at baseline were more likely to have experienced an
internal promotion at follow-up (Exp(B)=1.58, p<.001). However, no significant results were found for professional
ability. Therefore, hypothesis 1 is partially supported.
Hypotheses 2–4 posited that positive self-rated employability (professional ability and developmental
proactivity) predicts a lower probability of experiencing a future salary loss (H2), demotion (H3), or demotion and
salary loss (H4). Across these outcomes, no statistically significant findings were found for developmental
proactivity. However, professional ability was significantly related to having experienced a salary loss at follow-up.
The odds ratio fell below 1, indicating that respondents who assessed their professional ability positively at baseline
had a lower likelihood of having experienced a salary loss at follow-up (Exp(B)=0.68, p<.001). Therefore,
hypothesis 2 is partially supported.
The analyses also partially support hypotheses 3 and 4. Professional ability reliably distinguished between
workers who had experienced a demotion and those who had experienced no career transition at follow-up.
Specifically, we found that workers who rated their professional ability favourably at baseline had a lower likelihood
of having experienced a demotion at follow-up (Exp(B)=0.65, p<.001). In addition, we found a significant
relationship between self-assessed professional ability at baseline and having experienced demotion and salary loss
at follow-up. Specifically, workers who rated their professional ability positively at baseline were less likely to have
experienced a simultaneous demotion and salary loss at follow-up (Exp(B)=0.47, p<.001). This negative relation is
stronger than those separate relationships between professional ability and salary loss (H2) and demotion (H3),
which confirms hypothesis 4.
Conny Roobol, Hans Pruijt, Ferry Koster and Fenna Ruby Marie Leijten
74
Table 2c: Multinomial logistic regression analyses for sub-dataset 1: positive and negative career events as well as stability at follow-up as a
function of employability resources and control variables assessed at baseline, including panel attrition analyses
Model 1 (M1)¹
Model 2 (M2)^
Model 3 (M3)^^
Exp(B)
Exp(B)
Exp(B)
Category 1: promotion (n=597 (M1); n=593 (M2 and M3))
Employability resources
Developmental proactivity
1.69***
1.58***
1.49***
Professional ability
1.12
1.03
0.98
Control variables
Age (years)
0.93***
0.93***
Gender (ref=male)
0.85
0.85
Education (ref=high)
low
0.86
0.87
middle
0.74**
0.74**
Job tenure (years)
0.97***
0.97***
Relaxation and recovery
0.83**
0.83**
Panel attrition
Participation in two waves (P1)
1.13
Participation in three waves (P2)
0.03**
Developmental proactivity*P1
1.06
Developmental proactivity*P2
1.49
Professional ability*P1
0.77
Professional ability*P2
1.41
Constant
-4.95***
0.58
1.27
Category 2: salary loss (singular; n=273 (M1); n=272 (M2 and M3))
Employability resources
Developmental proactivity
1.07
1.09
1.15
Professional ability
0.67***
0.68***
0.63***
Control variables
Age (years)
1.01
1.01
Gender (ref=male)
0.98
0.98
Education (ref=high)
low
1.08
1.08
middle
0.83
0.83
Job tenure (years)
1.01
1.01
Relaxation and recovery
1.02
1.03
Int. Journal of Business Science and Applied Management/ Business-and-Management.org
75
Panel attrition
Participation in two waves (P1)
0.22
Participation in three waves (P2)
1.09
Developmental proactivity*P1
0.95
Developmental proactivity*P2
0.77
Professional ability*P1
1.21
Professional ability*P2
1.18
Constant
-1.76***
-2.36*
-2.13*
Category 3: demotion (singular; n=216 in all models)
Employability resources
Developmental proactivity
1.30*
1.24
1.27
Professional ability
0.67***
0.65***
0.59***
Control variables
Age (years)
1.00
1.00
Gender (ref=male)
0.92
0.92
Education (ref=high)
low
0.89
0.90
middle
1.15
1.15
Job tenure (years)
0.97***
0.97***
Relaxation and recovery
0.94
0.95
Panel attrition
Participation in two waves (P1)
0.18
Participation in three waves (P2)
0.21
Developmental proactivity*P1
0.79
Developmental proactivity*P2
1.01
Professional ability*P1
1.54
Professional ability*P2
1.36
Constant
-2.77***
-2.18*
-1.62
Category 4: salary loss and demotion (combined; n=113 in all models)
Employability resources
Developmental proactivity
1.15
1.09
1.09
Professional ability
0.49***
0.47***
0.43***
Control variables
Age (years)
0.98
0.98
Gender (ref=male)
1.38
1.39
Education (ref=high)
Conny Roobol, Hans Pruijt, Ferry Koster and Fenna Ruby Marie Leijten
76
low
0.76
0.77
middle
0.82
0.84
Job tenure (years)
0.97*
0.97*
Relaxation and recovery
1.11
1.12
Panel attrition
Participation in two waves (P1)
5.23
Participation in three waves (P2)
0.16
Developmental proactivity*P1
0.45
Developmental proactivity*P2
1.33
Professional ability*P1
1.06
Professional ability*P2
1.21
Constant
-1.65***
-0.28
1.00
-2LL (χ²)
1730.3 (113.1***)
9251.4 (287.7***)
9171.1 (369.5***)
∆χ²(df)
-
174.6 (24)***
81.8 (24)***
AIC
1754.3
9323.4
9291.1
Nagelkerke R² .020 .052 .066
Notes: Reference category: sustained stability (n=6,527 (M1); n=6,487 (M2 and M3)).
*p<0.05; **p<0.01; ***p<0.001.
¹Listwise deletion reduced the number of records to n=7,726 (< 1% of total).
^Listwise deletion reduced the number of records to n=7,681 (< 1% of total).
^^Listwise deletion reduced the number of records to n=7,681 (< 1% of total).
Int. Journal of Business Science and Applied Management/ Business-and-Management.org
77
We performed an additional analysis to correct for any changes in organizational setting and work
conditions that might have affected the findings for sub-dataset 1. Specifically, we re-ran our
multinomial logistic regression analysis on a sample that included participants who had experienced
neither a reduction in working hours nor a switch of employer at follow-up (N=6,233). The differences
in parameter estimates were negligible, except for salary loss. In that case, the coefficient for
professional ability became non-significant, perhaps because of the severe shrinkage in sample size for
this category, leading to deflated correlations and inflated Type II error rates. (Results are available
from the first author upon request.)
Tables 3a and 3b present the descriptive statistics for sub-dataset 2. Respondents had an average
age of 53 years, were predominantly male, and had medium or higher levels of education. Contrary to
sub-dataset 1, which assessed career events among employed persons, sub-dataset 2 compared those
remaining employed to those who became unemployed at follow-up. Overall, multi-collinearity was
not evident, since correlations between the study variables fell below .80 (Field, 2009). As with sub-
dataset 1, we found a positive relation (r=.17, p<.01) between developmental proactivity and
professional ability. Also, the mean scores on the employability resources were quite similar to those of
sub-dataset 1: 3.93 for professional ability and 4.21 for developmental proactivity. In addition, we
found a negative correlation between developmental proactivity and relaxation and recovery (r=-.12,
p<.01), indicating that workers who anticipated knowledge needs in their jobs had more difficulty
turning off from work.
Table 3a: Descriptive results for sub-dataset 2: means, standard deviations and ranges for the
independent and control variables at baseline
Mean/proportion SD Range
Employability resources
Developmental proactivity
3.93
0.57
1-5
Professional ability
4.21
0.63
1-5
Demographic characteristics
Age (years)
53.05
4.85
45-64
Gender (ref=male)
0.37
0/1
Education (ref=high)¹
low
0.26
0/1
middle
0.39
0/1
Work-related characteristic
Job tenure (years)
10.85
9.60
0-45
Well-being related characteristic
Relaxation and recovery
2.91
0.72
1-5
Notes: Due to missing values (< 1,5% of total), the N ranges from 7,168 to 7,241. ¹low: (not) finished
primary school or completed lower vocational education; middle: finished secondary school or higher
vocational training; high: obtained a bachelor, master or PhD degree.
Conny Roobol, Hans Pruijt, Ferry Koster and Fenna Ruby Marie Leijten
78
Table 3b: Zero-order (Pearson r) correlation coefficients for the independent and control
variables at baseline (sub-dataset 2)
1. 2. 3. 4. 5. 6.
1. Developmental proactivity -
2. Professional ability
.173**
-
3. Age (years) -.027* -.164** -
4. Gender (ref=male) .037** -.035** .001 -
5. Education¹
.166**
.067**
-.019
-.051**
-
6. Job tenure (years) -.084** -.071** .222** -.062** -.060** -
7. Relaxation and recovery -.120** .070** .015 -.066** -.144** .010
Notes: *p<0.05; **p<0.01. ¹Spearman correlation used.
Due to missing values (< 1,1% of total), the N fluctuates between 7,168 and 7,241.
Table 3c displays the results from the binary logistic regression analysis for sub-dataset 2. Model 1,
which contains the employability resources, provides a statistically significant better fit than the
intercept-only model (χ² (2)=81.3, p<.001). The accompanying AIC is 3150.4 and R² is .031. Model 2,
which includes the control variables, shows a significant improvement in fit over Model 1, with χ²
(6)=69.3, p<.001 and AIC is 3070.2 (R²=.058). Model 3 shows the panel attrition analysis. Differences
between the findings of this and our full model (Model 2) were very small, indicating that respondents’
participation patterns did not distort our results.
Hypothesis 5 predicted that workers with a favourable self-assessment of their employability
(professional ability and developmental proactivity) are less likely to experience future unemployment.
Our logistic regression analysis showed a significant relationship for professional ability, that is, those
who self-assessed as having greater professional ability at baseline had a higher probability of having
retained work (‘sustainable employment’, Exp(B)=1.89, p<.001). However, the relationship between
developmental proactivity and sustainable employment was non-significant. Therefore, the data
partially support hypothesis 5. Figure 2 summarizes the outcomes.
Table 3c: Binary logistic regression analysis for sub-dataset 2: negative career events and
stability at follow-up as a function of employability resources and control variables assessed at
baseline, including panel attrition analysis
Model 1¹
Model 2^
Model 3^^
Exp(B)
Exp(B)
Exp(B)
Employability resources
Developmental proactivity
0.99
0.96
1.01
Professional ability
1.90***
1.89***
1.98***
Control variables
Age (years)
0.97***
0.97***
Gender (ref=male)
0.94
0.93
Education (ref=high)
low
0.71**
0.71**
middle
0.97
0.96
Job tenure (years)
1.04***
1.04***
Relaxation and recovery
0.77***
0.77***
Panel attrition
Participation in two waves (P1)
8.67²
Participation in three waves (P2)
2.64
Developmental proactivity*P1
0.79
Developmental proactivity*P2
0.90
Professional ability*P1
0.83
Int. Journal of Business Science and Applied Management/ Business-and-Management.org
79
Professional ability*P2
0.91
Constant
0.17
2.78***
2.30**
-2LL (χ²)
3144.4 (81.3***)
3052.2 (150.6***)
3041.0 (161.8***)
∆χ²(df)
-
69.3 (6)***
11.2 (6)
AIC
3150.4
3070.2
3071.0
Nagelkerke R²
.031
.058
.062
Notes: Reference category: transition to unemployment at follow-up.
*p<0.05; **p<0.01; ***p<0.001.
¹Listwise deletion reduced the number of records to n=7,213 (< 1% of total).
^Listwise deletion reduced the number of records to n=7,161 (=1,1% of total).
^^Listwise deletion reduced the number of records to n=7,161 (=1,1% of total).
²The non-significance of this parameter estimate may be due to an inflated Type II
error rate caused by the Wald statistic. However, since the difference in the -2LL
between model 2 and model 3 is non-significant, model 3 does not fit the data better
than model 2, leading us to conclude that it is not strictly necessary to control for panel
attrition when interpreting the findings for sub-dataset 2.
Figure 2 Corroborated theoretical framework
In addition, we performed two sensitivity analyses to assess whether the findings for sub-dataset 2
were robust against the unemployment patterns respondents experienced during follow-up. The first
analysis assessed whether the year in which respondents first became unemployed was important
(N=7,241). We performed a multinomial logistic regression analysis consisting of four categories:
transition to unemployment in year 2, transition to unemployment in year 3, transition to
unemployment in year 4, and sustained employment (reference category). The results aligned with our
binary regression analysis: respondents who positively self-assessed their professional ability at
baseline were less likely to have become unemployed at follow-up, regardless of the specific year in
which the transition occurred.
We conducted a second multinomial logistic regression analysis to test whether respondents who
had recovered from unemployment at follow-up showed any differences in their baseline employability
scores compared with those who did not (N=7,241). The odds ratios of both groups were comparable
Employability resource 1:
Developmental proactivity
Experiencing positive career
events
Promotion
Employability resource 2:
Professional ability
Circumventing negative career
events
Salary loss
Demotion
Salary loss + demotion
Unemployment
Conny Roobol, Hans Pruijt, Ferry Koster and Fenna Ruby Marie Leijten
80
for professional ability, implying that this employability facet, with a certain probability, protects
against unemployment, regardless of whether unemployment continues. (Results are available from the
first author upon request.)
Finally, it seems worth noting that the inclusion of age as a control variable provided exploratory
findings that pertain to potential differentiations within the spectrum between 45 and 65 years. We see
that professional ability declines with age, but the correlation was small (r= -.15, p<.01., Table 2b). The
results from the binary logistic regression analysis for sub-dataset 2, model 2, show a significant
positive relationship between age and sustainable employment (see Table 3c). Thus, in our sample of
workers aged 45 to 64, older participants were less likely to transition into unemployment.
6. DISCUSSION
Would it be possible to optimize organizations to boost developmental proactivity while
conserving professional ability? As far as developmental proactivity is concerned, a crucial factor
would be the extent to which the work organization is geared towards learning. Important features of
such a work organization are job rotation, job enlargement, and job enrichment, such as adding
managerial tasks to production work (Pruijt, 1997; Pruijt, 2003; Weichel et al., 2009). The latter
includes rotating workers to positions in which they can learn rather than keeping them in positions in
which they can maximize productivity (Svensson, 2004). The transparency that exists in a learning-
oriented organization enables workers to operate on the basis of a complex mental model of the work.
There is team learning, and workers are able to make changes to the work processes (Kira & van
Eijnatten, 2009). Teams engage in self-design (Stebbins & Shani, 2009). Finally, workers and
managers strive to neutralize prevalent stereotypes about older workers as being change-averse (Taylor
et al., 2010; Van Veldhoven & Dorenbosch, 2008). The ideal is to achieve polyvalence in flexible
teams in a learning organization.
As for the conservation of professional ability, the point of departure is the notion that the
variation in abilities tends to increase as workers get older (Hardy & Reyes, 2015). Because of this
variation, HRM practices that treat workers as standardized assets do not fit. Instead, job design gets
customized so that workers do what they can do well and avoid what they cannot (Zinsmeister &
Meerman, 2013). The organization has to be flexible in the sense that it can adapt the division of labour
to accommodate the growing variation in abilities that tends to characterize an ageing workforce
(Weichel et al., 2009). Employment protection for older workers can incentivize organizations to put in
the required effort to stabilize older workers in their jobs (Pruijt, 2013). In the Netherlands, such
employment protection exists in the form of tenure-dependent redundancy compensation (Pruijt &
Dérogée, 2010). This employment protection might explain our exploratory finding that the
unemployment risk did not increase with age. However, an ability-driven division of labour can make it
difficult for workers to cope with change, and the same holds for the organization (Mintzberg, 2009).
Thus, there seems to be a contradiction between optimizing for the enhancement of developmental
proactivity and for the conservation of professional ability, but a radical choice between one or the
other seems risky with respect to career sustainability. This dilemma is not likely to disappear as
workers tend to retire later—in the Netherlands the retirement age rose from 63.8 years on average in
2014 to 65 years in 2018 (CBS, 2019)—while change due to restructuring and new technology seems
unrelenting.
7. CONCLUSION
This study tested a parsimonious model that relates two facets of employability—professional
ability and developmental proactivity—to older workers’ career events. The main findings are that
developmental proactivity led to a higher probability of internal promotion, and professional ability led
to a lower incidence of negative career events (salary loss, demotion, and unemployment). These
findings suggest an implication for theory: the role of proactivity should be put into perspective, at least
as far as older workers are concerned. In general, proactivity is central to ideas about employability,
and current thinking about sustainable organizational performance tends to reflect the same emphasis
(Dorenbosch, 2014). However, although proactivity is certainly relevant for older workers, the ability
to perform day-to-day work confidently seems critical.
The study shows that the conservation of resources (COR) model provides a valuable heuristic
framework for studying the employability of older workers. The model symmetrically covers potential
positive and negative developments, and therefore, directs attention not only to the ideal of proactivity,
but also to the less glamorous ability to cope with day-to-day work demands.
Our findings support the view that the value of resources depends on the specific goals that
individuals set for themselves (Hobfoll et al., 2018; Halbesleben et al., 2014). Our study shows how a
Int. Journal of Business Science and Applied Management/ Business-and-Management.org
81
particular resource—developmental proactivity—seems to be valuable mainly for older workers who
are pursuing promotion. Similarly, professional ability is especially important for older workers who
aim to maintain their positions.
The use of age as a control variable allows for a tentative conclusion about the spectrum between
45 and 65 years and potential differentiations. This concerns the exploratory finding that age is
associated with a declining risk of transition into unemployment. This could be due to the employment
protection that existed in the Netherlands during the data collection. Generally, employers were
required to pay compensation of one month’s salary per year of tenure when dismissing a worker.
Since then, this compensation has been halved.
7.1 Theoretical Implications
The main theoretical contribution of this paper is a conceptualization of employability geared to
older workers, and characterized by a sharp distinction between resource conservation and resource
acquisition. An additional theoretical implication of our analysis is that COR theory is a suitable
framework for explaining diverging outcomes.
In the light of recent renderings of COR theory, this is not self-evident. These recent renderings
posit that resources have equifinality, which means that different resources can contribute to the
achievement of the same goal, and that resources may be substituted for one another (Halbesleben et
al., 2014). Equifinality makes it hard to conceptualize how resource A can cause Outcome X and
resource B Outcome Y. Moreover, in recent COR theory, the notion of resource caravans—that
resources are highly correlated—has gained currency (Hobfoll et al., 2018). This notion leads away
from considering the possibility of diverging outcomes.
7.2 Practical Implications
To draw out practical implications, we need an interpretation of the questions and answers that
underlie the predictors. The items in the professional ability scale could be seen as measuring the extent
to which the employee is struggling. Not to be struggling protects against negative career events farther
down the road. In themselves, such events may or may not have negative consequences for the
organization; however, the trajectory that begins with struggling to cope at work and ends with a
negative career event tends to involve personal resource depletion, which is always costly for an
organization. One may think of impacts on productivity, error rate, quality, safety, and morale.
Whether an employee is struggling depends on characteristics of both the individual and the work.
Important areas for management interventions are issues that can cause experienced and suitably
qualified employees to struggle, and those issues are highly context-dependent. Examples are:
administrative duties on top of professional work; long working hours; lack of support to deal with
complex social problems that patients, students, or clients bring to the workplace; uncertainty-
generating reorganizations; or excessive physical demands. Efforts to address such issues may well do
more for older workers’ employability than interventions directed at attitudes or behaviours.
Organizations and managers are advised to monitor employees’ struggles and to address preventable
causes promptly.
In the study, the construct of developmental proactivity represents behavioural and attitudinal
aspects. The fact that developmental proactivity did not seem to protect against negative career events
is a reason to doubt the benefits of attitude- or behaviour-directed interventions for organizations.
Developmental proactivity does predict promotions; therefore, it seems especially relevant for
individual employees. Therefore, as Cheung et al. (2019) noted, it is a suitable theme for career advice.
Topics could include cultivating a habit of looking for new things to learn, routinely reflecting on ways
to improve the work, networking with an eye toward learning opportunities, and assessing future skill
and knowledge requirements.
7.3 Limitations
We acknowledge three limitations to this study. First, we did not study COR’s resource gain and
loss spirals or the cumulative gain in or loss of resources (Ungerath, 2012). Second, we did not include
mediators in the link between employability and career events. Third, the findings in this study are
strictly about the employability of older workers, not employability in general.
7.4 Future Research Suggestions
A longitudinal design with a longer view of individuals’ working lives than the STREAM study
could shed light on resource gain and loss spirals or the cumulative gain in, or loss of, resources. When
probing gain and loss spirals, researchers could use latent transition modelling or sequence analysis in
combination with logistic regression analysis.
Conny Roobol, Hans Pruijt, Ferry Koster and Fenna Ruby Marie Leijten
82
An additional interesting avenue for research would be to include mediating factors, such as job
performance and well-being. Recent research (De Cuyper, Mäkikangas et al., 2012; Van Harten, 2016;
Vanhercke et al., 2015) has shown that employability predicts the possession of these and related
resources. Some competencies such as corporate sense, which figures as an independent variable in
existing research (Van der Heijde & Van der Heijden, 2006; Van der Heijden et al., 2009), could
arguably also be considered a mediating factor.
It would also be interesting to know if, and to what extent, the pattern uncovered in this paper
applies to younger workers, especially because these workers have to survive in an unpredictable world
of work where a stable intra-firm career is no longer the prerogative of many. Including both younger
and older workers also opens up the possibility of treating age as a moderator in the link between
employability and career events. According to life course theorists (Baltes et al., 1999), workers in their
early and mid-careers (<45 years) attach importance to knowledge acquisition. In contrast, workers in
their late careers (≥45 years) value protection of losses. These motivational differences between age
groups suggest that the effects of investments in personal resources such as employability on career
outcomes might vary between workers of different ages.
REFERENCES
Arthur, M. B., & Rousseau, D. M. (1995). Introduction: The boundaryless career as a new employment
principle. In M. B. Arthur, & D. M. Rousseau (Eds.), The boundaryless career: a new employment
principle for a new organizational era.(pp. 3–20). Oxford University Press.
Bakker, A., Emmerik, H., & Euwema, M. (2006). Crossover of burnout and engagement in work
teams. Work and Occupations, 33(4), 464–489.
Baltes, P. B., Staudinger, U. M., & Lindenberger, U. (1999). Lifespan psychology: Theory and
application to intellectual functioning. Annual Review of Psychology, 50, 471–507.
Baruch, Y., & Holtom, B. C. (2008). Survey response rate levels and trends in organizational research.
Human Relations, 61(8), 1139–1160.
Bosch, N., & Ter Weel, B. (2013). Labour-market outcomes of older workers in the Netherlands:
Measuring job prospects using the occupational age structure. CPB Netherlands Bureau for
Economic Policy Analysis.
CBS. (2019). Werknemers in 2018 gemiddeld 65 jaar bij pensionering. [In 2018, workers average 65
when retiring.] Statistics Netherlands, https://www.cbs.nl/nl-nl/nieuws/2019/32/werknemers-in-
2018-gemiddeld-65-jaar-bij-pensionering,
Cheung, F., Yeung, D. Y., & Wu, A. M. S. (2019). Occupational future time perspective and successful
aging at work. Journal of Career Development, 46(1), 3–16.
Cohen, J. (1977). Statistical power analysis for behavioral sciences (Rev. ed.). Academic Press.
De Cuyper, N., Mäkikangas, A., Kinnunen, U., Mauno, S., & De Witte, H. (2012). Cross-lagged
associations between perceived external employability, job insecurity, and exhaustion: Testing
gain and loss spirals according to the Conservation of Resources Theory. Journal of
Organizational Behavior, 33, 770–788.
De Cuyper, N., Raeder, S., Van der Heijden, B. I. J. M., & Wittekind, A. (2012). The association
between workers’ employability and burnout in a reorganization context: Longitudinal evidence
building upon the conservation of resources theory. Journal of Occupational Health Psychology,
17(2), 162–174.
De Cuyper, N., Sulea, C., Philippaers, K., Fischmann, G., Iliescu, D., & De Witte, H. (2014). Perceived
employability and performance: moderation by felt job insecurity. Personnel Review, 43(4), 536–
552.
De Rooij, J. (2014, October 30). Je werkgever stelt demotie voor: wat nu? [Your employer proposes
demotion. What to do?] Intermediair.
https://www.intermediair.nl/carriere/carriereontwikkeling/je-werkgever-stelt-demotie-voor-wat-
nu?
De Vos, A., De Hauw, S., & Van der Heijden, B. I. J. M. (2011). Competency development and career
success: The mediating role of employability. Journal of Vocational Behavior, 79(2), 438–447.
Int. Journal of Business Science and Applied Management/ Business-and-Management.org
83
Dorenbosch, L. (2014). Striking a balance between work effort and resource regeneration. Vitality as a
sustainable performance concept. In I. Ehnert, W. Harry, & K. Zink (Eds.), Sustainability and
human resource management: Developing sustainable business organizations (pp. 156–180).
Springer.
Eby, L. T., Butts, M., & Lockwood, A. (2003). Predictors of success in the era of the boundaryless
career. Journal of Organizational Behavior, 24, 689–708.
Field, A. (2009). Discovering statistics using SPSS (3
rd
ed.). Sage Publications Ltd.
Forrier, A., & Sels, L. (2003). Temporary employment and employability: Training opportunities and
efforts of temporary and permanent employees in Belgium. Work, Employment and Society, 17(4),
641–666.
Fugate, M., Kinicki, A. J., & Ashforth, B. E. (2004). Employability: A psycho-social construct, its
dimensions, and applications. Journal of Vocational Behavior, 65, 14–38.
Gazier, B. (2001). Employability—The complexity of a policy notion. In P. Weinert, M. Baukens, P.
Bollerot, M. Pineschi-Gapenne, & U. Walwei (Eds.), Employability: From theory to practice (pp.
3–23). Transaction Publishers.
Ghoshal, S., Bartlett, C., & Moran, P. (1999). A new manifesto for management. Sloan Management
Review, 40(3), 9–20.
Hair, J. F., Jr., Black, W. C., Babin, B. J., & Anderson, R. E. (2010). Multivariate data analysis (7
th
ed.). Pearson Prentice Hall.
Halbesleben, J. R. B., Neveu, J. P., Paustian-Underdahl, S. C., & Westman, M. (2014). Getting to the
“COR”: Understanding the role of resources in conservation of resources theory. Journal of
Management, 40(5), 1–31.
Hardy, M. & Reyes, A. (2015). Aging and work performance. In G. Ritzer (Ed.), The Blackwell
Encyclopedia of Sociology. John Wiley & Sons.
https://onlinelibrary.wiley.com/doi/book/10.1002/9781405165518
Hillage, J., & Pollard, E. (1998). Employability: Developing a framework for policy analysis (Research
Report No. 85). The Institute for Employment Studies. Department for Education and
Employment.
https://www.researchgate.net/publication/225083565_Employability_Developing_a_framework_f
or_policy_analysis_London_DfEE
Hirschi, A. (2012). The career resources model: An integrative framework for career counsellors.
British Journal of Guidance & Counselling, 40(4), 369–383.
Hobfoll, S. E. (1989). Conservation of resources: A new attempt at conceptualizing stress. American
Psychologist, 44(3), 513–524.
Hobfoll, S. E. (2002). Social and psychological resources and adaptation. Review of General
Psychology, 6(4), 307–324.
Hobfoll, S. E., Halbesleben, J., Neveu, J. P., & Westman, M. (2018). Conservation of resources in the
organizational context: The reality of resources and their consequences. Annual Review of
Organizational Psychology and Organizational Behavior, 5, 103–128.
Hox, J. J. (2010). Multilevel analysis. Techniques and applications (2
nd
ed.). Routledge.
Jongsma, R. (2014). Demotie: het laatste taboe doorbroken. [Demotion: the last taboo broken.]
Vakmedianet.
Kanter, R. M. (1993). The view from the 1990s: How the global economy is reshaping corporate power
and careers. afterwoord to the 1993 edition. In R. M. Kanter, Men and women of the corporation
(pp. 289-328). New York: Basic Books.
Kira, M., & Van Eijnatten, F. M. (2009). Sustained by work: Individual and collective sustainability in
work organizations. In P. Docherty, M. Kira, & A. B. Shani (Eds.), Creating sustainable work
systems: Developing social sustainability (pp. 233–246). Routledge.
Leijten, F. R. M., De Wind, A., Van den Heuvel, S. G., Ybema, J. F., Van der Beek, A. J., Robroek, S.
J. W., & Burdorf, A. (2015). The influence of chronic health problems and work-related factors on
loss of paid employment among older workers. J Epidemiol Community Health, 69(11), 1058–
1065.
Littleton, S. M., Arthur, M. B., & Rousseau, D. M. (2000). The future of boundaryless careers. In A.
Collin, & R. A. Young (Eds.), The Future of Career (pp. 101–114). Cambridge University Press.
Conny Roobol, Hans Pruijt, Ferry Koster and Fenna Ruby Marie Leijten
84
Ma, C., Wu, C.-H., Chen, Z. X., Jiang, X., & Wei, W. (2019). Why and when leader humility promotes
constructive voice: A crossover of energy perspective. Personnel Review, 49(5), 1157–1175.
Mintzberg, H. (2009). Structure in 5’s: Designing effective organizations. Pearson Custom Publishing.
Ng, T. W. H., Eby, L. T., Sorensen, K. L., & Feldman, D. C. (2005). Predictors of objective and
subjective career success: A meta-analysis. Personnel Psychology, 58(2): 367–408.
Ng, T. W. H., & Feldman, D. C. (2014). A conservation of resources perspective on career hurdles and
salary attainment. Journal of Vocational Behavior, 85(1), 156–168.
Pruijt, H. (1997). Job design and technology. Taylorism vs. anti-Taylorism. Routledge.
Pruijt, H. (2003). Teams between neo-Taylorism and anti-Taylorism. Economic and Industrial
Democracy, 24(1), 77–101.
Pruijt, H. (2013). Employability, empowerment and employers, between debunking and appreciating
action. Nine cases from the ICT sector. International Journal of Human Resource Management,
24(8), 1613–1628.
Pruijt, H., & Dérogée, P. (2010). Employability and job security, friends or foes? The paradoxical
reception of employacurity in the Netherlands. Socio-Economic Review, 8(3), 437–460.
Raemdonck, I., Tillema, H., De Grip, A., Valcke, M., & Segers, M. (2012). Does self-directedness in
learning and careers predict the employability of low-qualified employees? Vocations and
Learning, 5, 137–151.
Rosseel, Y. (2012). Lavaan: An R package for structural equation modeling. Journal of Statistical
Software, 48(2), 1–36.
Savickas, M. L. (2000). Renovating the psychology of careers for the twenty-first century. In A. Collin,
& R. A. Young (Eds.), The Future of Career (pp. 53–68). Cambridge University Press.
Seibert, S. E., Crant, M. J., & Kraimer, M. L. (1999). Proactive personality and career success. Journal
of Applied Psychology, 84(3), 416-427.
Son, H., Friedmann, E., & Thomas, S. A. (2012). Application of pattern mixture models to address
missing data in longitudinal data analysis using SPSS. Nursing Research, 61(3), 195–203.
Spurk, D., Hirschi, A., & Dries, N. (2019). Antecedents and outcomes of objective versus subjective
career success: Competing perspectives and future directions. Journal of Management, 45(1), 35-
69.
Stebbins, M., & Shani, A. B. R. (2009). Toward a sustainable work systems design and change
methodology. In P. Docherty, M. Kira, & A. B. Shani (Eds.), Creating sustainable work systems:
Developing social sustainability (pp. 247–267). Routledge.
Svensson, L. (2004). Lifelong learning: A clash between a production and a learning logic. In C.
Garsten & K. Jacobsson (Eds.), Learning to be employable: New agendas on work, responsibility,
and learning in a globalizing world (pp. 83–106). Palgrave Macmillan.
Taylor, P., Brooke, L., McLoughlin, C., & Biase, T. D. (2010). Older workers and organizational
change: Corporate memory versus potentiality. International Journal of Manpower, 31(3), 374–
386.
Thijssen, J. G. L., Van der Heijden, B. I. J. M., & Rocco, T. S. (2008). Toward the employability-link
model: Current employment transition to future employment perspectives. Human Resource
Development Review, 7(2), 165–183.
Trougakos, J. P., Beal, D. J., Cheng, B. H., Hideg, I., & Zweig, D. (2015). Too drained to help: A
resource depletion perspective on daily interpersonal citizenship behaviors. Journal of Applied
Psychology, 100(1), 227–236.
Ungerath, S. (2012). Towards a refined understanding of resource gain spirals—A theoretical
discussion of existing research, and novel empirical evidence from two longitudinal intervention
studies. [Unpublished doctoral dissertation]. Bergische Universität Wuppertal.
Van Dalen, H. P., & Henkens, K. (2015). Why demotion of older workers is a no-go area for managers.
Center Discussion Paper, vol. 2015–025. Tilburg University Department of Economics.
https://research.tilburguniversity.edu/en/publications/why-demotion-of-older-workers-is-a-no-go-
area-for-managers
Van den Broeck, A., De Cuyper, N., De Witte, H., & Vansteenkiste, M. (2010). Not all job demands
are equal: Differentiating job hindrances and job challenges in the Job Demands-Resources model.
European Journal of Work and Organizational Psychology, 19(6), 735–759.
Int. Journal of Business Science and Applied Management/ Business-and-Management.org
85
Van den Heuvel, S., Geuskens, G., Van der Meer, L., Wind, A., & Leijten, F. (2014). Study on
transitions in employment, ability, and motivation (STREAM). Report, TNO, Leiden, The
Netherlands.
Van der Heijde, C. M., & Van der Heijden, B. I. J. M. (2006). A competence-based and
multidimensional operationalization and measurement of employability. Human Resource
Management, 45(3), 449–476.
Van der Heijden, B. I. J. M., De Lange, A. H., Demerouti, E., & Van der Heijde, C. M. (2009). Age
effects on the employability-career success relationship. Journal of Vocational Behavior, 74, 156–
164.
Van Harten, E. J. (2016). Employable ever after: Examining the antecedents and outcomes of
sustainable employability in a hospital context. [Unpublished doctoral dissertation]. Utrecht
University.
Vanhercke, D., Kirves, K., De Cuyper, N., Verbruggen, M., Forrier, A., & De Witte, H. (2015).
Perceived employability and psychological functioning framed by gain and loss cycles. Career
Development International, 20(2), 179–198.
Van Veldhoven, M., & Dorenbosch, L. (2008). Age, proactivity and career development. Career
Development International, 13(2), 112–131.
Volmer, J., & Spurk, D. (2011). Protean and boundaryless career attitudes: Relationships with
subjective and objective career success. Zeitschrift für Arbeitsmarktforschung [Journal for Labour
Market Research], 43(3), 207–218.
Waterman, R. H., Waterman, J. A., & Collard, B. A. (1994). Toward a career-resilient workforce.
Harvard Business Review, 72(4) 87–95.
Weichel, J., Buch, M., Urban, D., & Frieling, E. (2009). Sustainability and the ageing workforce—
Considerations with regard to the German car manufacturing industry. In P. Docherty, M. Kira, &
A. B Shani (Eds.), Creating sustainable work systems: Developing social sustainability (pp. 70–
83). Routledge.
Wolf, F. M. (1986). Meta-analysis: Quantitative methods for research synthesis. Sage.
Xanthopoulou, D., Bakker, A. B., Demerouti, E., & Schaufeli, W. B. (2009). Reciprocal relationships
between job resources, personal resources, and work engagement. Journal of Vocational
Behavior,74, 235–244.
Xantopoulou, D., Bakker, A. B., Kantas, A., & Demerouti, E. (2012). Measuring burnout and work
engagement: Factor structure, invariance, and latent mean differences across Greece and the
Netherlands. International Journal of Business Science and Applied Management, 7(2), 40–52.
Yang, F., & Chau, R. (2016). Proactive personality and career success. Journal of Managerial
Psychology, 31(2), 467–482.
Ybema, J. F., Geuskens, G. A., Van den Heuvel, S. G., De Wind, A., Leijten, F. R. M., Joling, C., &
Bongers, P. M. (2014). Study on transitions in employment, ability and motivation (STREAM):
The design of a four-year longitudinal cohort study among 15,118 persons aged 45 to 64 years.
British Journal of Medicine & Medical Research, 4(6), 1383–1399.
Zacher, H., & Kooij, D. T. A. M. (2017). Aging and proactivity. In S. K. Parker & U. K. Bindl (Eds.),
Proactivity at work: Making things happen in organizations (pp. 258–294). Routledge.
Zinsmeister, J., & Meerman, M. G. M. (2013). The deployment of enriched capacities: The rights of
older workers. Journal of Workplace Rights, 17(3), 303–
Conny Roobol, Hans Pruijt, Ferry Koster and Fenna Ruby Marie Leijten
86
APPENDICES
Appendix I: Older workers’ intra- and inter-firm mobility patterns in the cohort study STREAM
(2010-2013)¹
Measurement wave
Time 1 (2010)
Time 2 (2011)
Time 3 (2012)
Time 4 (2013)
Sub-dataset →
1
2
1
2
1
2
1
2
Mobility pattern ↓
Same employer, same job
7,039
6,213
6,180
5,579
5,940
5,294
5,468
4,732
Same employer, different job
422
651
567
575
538
532
475
446
Different employer, same job
196
233
162
150
149
148
143
140
Different employer, different job
92
140
67
78
64
56
59
59
Total n responses³
7,749
7,237
6,976
6,382
6,691
6,030
6,145
5,377
Total n unemployed²
NA
-
NA
109
NA
202
NA
315
Total n participation in wave
7,751
7,241
6,978
6,493
6,692
6,233
6,145
5,692
¹Figures showcase the number of respondents who participated in a specific wave and indicated having switched
employers and/or jobs or otherwise reported having been at the same employer and job in the 12 months
preceding the measurement wave.
²Respondents who indicated becoming unemployed during a specific follow-up wave did not receive the question
regarding employer/job switch.
³The question regarding employer/job switch was not mandatory, meaning that the total n responses is not
necessarily equal to the total n participation in that wave (minus the number of unemployed persons in case of
sub-dataset 2). Differences are, however, very small.
Int. Journal of Business Science and Applied Management/ Business-and-Management.org
87
Appendix II: Flowchart of the responses to all measurements of the STREAM questionnaire
Persons invited at T1
N=26,601
Source: Van den Heuvel et al., 2014: 8 (adapted version)
Conny Roobol, Hans Pruijt, Ferry Koster and Fenna Ruby Marie Leijten
88
Appendix III: Participation pattern of respondents comprising sub-dataset 1 and 2 compared to
the participation pattern of respondents included in the entire STREAM study
Sub-dataset
STREAM
1
2
Total
Participation pattern¹ ↓
n
%
n
%
n
%
1-1-1-1
5,257 67,8 4,827 66,7 9,639 63,8
1-1-1-0
804 10,4 792 10,9 1,313 8,7
1-1-0-1
355 4,6 339 4,7 628 4,2
1-1-0-0
562 7,3 535 7,4 850 5,6
1-0-1-1
391 5,0 392 5,4 724 4,8
1-0-1-0
240 3,1 222 3,1 381 2,5
1-0-0-1
142 1,8 134 1,9 246 1,6
1-0-0-0
NA NA 1,337 8,8
Total
7,751 100 7,241 100 15,118 100
¹A 1 signifies that a respondent has participated in a specific wave; a 0 signifies that a respondent has skipped that
wave.
Appendix IV: Survey items employability
Scale professional ability
In the coming 12 months, will you be able to…
1.…perform the work?
2.…keep up with the speed of the changes in your work?
In the coming 12 months, will you be able to continue working if…
3.…your work becomes physically more demanding?
4.…your work becomes emotionally more demanding?
5.…your work becomes more difficult?
