Perceived Employability in Difficult Economic Times: The significance of education systems and labour market policies

Today's European labour markets are characterized by uncertainty and job insecurity (Erlinghagen 2008; Chung and Mau 2014; Marx 2014). Many employees perceive that their current employment is at risk, so they must be prepared for mobility and job change. In such a situation, perceived employability is essential; in other words, the employees’ assessment of their possibilities of finding another job in the labour market. A high level of perceived employability gives the employee confidence to handle restructurings and readjustments and thereby to cope with job insecurity (Vanhercke et al.2014). The aim of this article is to contribute to the understanding of factors influencing the perceived employability of employees. The research question addressed is whether perceived employability mainly depends on individual factors such as age and education, or if contextual factors, particularly education systems and labour market policies, are additional influential factors.

It is widely recognized that an individual's perceived labour market prospects are affected by the national economic situation in a procyclical direction (cf. Berntson et al.2006; Vanhercke et al. 2014). In good economic times, with many job openings, the employee is more confident in finding a new job than during weak economic periods. Besides economic cycles, different institutions and policy measures may also influence an individual's perceived labour market prospects. Among policy-makers, the concept of employability is frequently used as a catchphrase for the significance of developing ‘relevant skills for the twenty-first century’ (OECD 2011: 30). In the employability paradigm, the improvement of an individual's employability in order to remain attractive to employers is a priority (Jacobsson 2004). Education systems, providing different training opportunities, may from this perspective be of central relevance for the perception of employment prospects.

In another strand of policy discussions – the so-called flexicurity approach – the concept of employment security is used, referring to the possibility of an individual finding a new job in case of job loss (Wilthagen and Tros 2004). The flexicurity approach emphasizes measures such as generous unemployment benefits and active labour market measures to enhance the individual's chances in the labour market. Following this perspective, general labour market policies as well as the education system may also be relevant for perceived employability.

In this article, we take a comprehensive approach to the study of perceived employability by focusing on individual characteristics, country-level education systems and labour market policies, and how factors on these different levels may interact. Previous studies on perceived employability have mainly focused on its relationship to individual characteristics, national labour market policies, and economic conditions (cf. Chung and van Oorschot 2011). The present article contributes to a broader understanding of how macro-level characteristics impact on an individual's perceived employability, both by including dimensions of the education system and focusing on cross-level interactions.

We start by defining the concept of perceived employability. Thereafter, we compare the paradigms of employability and flexicurity and their emphasis on mobility and transitions in the labour market. We then discuss how the institutional dimensions of education and labour market systems may affect the individual's perceived employability. Following that, we present the data, methods of analysis and results of this study. Finally, a discussion of the findings stresses the importance of emphasizing the institutional context for employees’ perceived employability.

Perceived employability is usually described as an important factor that helps employees to handle turbulent organizational environments and is defined as an individual's perceived possibilities of either staying in employment or finding new employment (Berntson et al. 2006). The notion refers to a subjective assessment (Vanhercke et al.2014). Thus, individuals may assess comparable situations differently, depending on their characteristics related to skills, competences, attitudes and motivation (Van der Heijde and Van der Heijden 2006; Vanhercke et al.2014).

In this article, we define perceived employability even more narrowly by focusing on employees’ subjective assessments of their chances in the external labour market. This more circumscribed definition of perceived employability does not confuse the concept with job security as the previous wider definition risks doing. In the current article perceived employability refers only to the perceived chances of transiting between jobs (Berglund et al.2014).

In labour markets where employers prioritize flexibility and employees experience high risk of unemployment, perceived employability in its more narrow sense may constitute an important coping mechanism to handle job insecurity. As is well known, job insecurity has many detrimental effects, among others the impact on health and well-being (Sverke et al.2002; De Witte 2005). Greenhalgh and Rosenblatt (1984) stress that individuals need to find coping strategies to mitigate the negative effects of job insecurity. From this perspective, perceived job alternatives may be regarded as one central coping mechanism (Cuyper et al.2008). For example, Berglund et al. (2014) show that the experience of job alternatives reduces worries of losing a job. Perceived employability may therefore create a feeling of ‘security of the wings’, to quote the famous Swedish economist Rehn (1988: 2014), by reducing the employee's dependence on a specific employer.

Thus, the concept of perceived employability refers, as we have discussed, to the individual level, and the purpose of this article is to analyse both individual and contextual factors affecting employees’ assessment of the possibility to make transitions between jobs. However, employability is not only a concept used in psychologically oriented research trying to understand individuals’ labour market behaviour. Over recent decades, the concept of employability has also been central in policy discussions. The notion of employability originates from the beginning of the 1990s and fits into the project of modernizing Europe by promoting growth, competitiveness and employment (Jacobsson 2004). Today, employability is a central concept in the growth strategy of Europe 2020. It emphasizes education as an important tool for the achievement of high levels of employment, productivity and social cohesion (EU 2012). Individuals’ employability should be improved by activating citizens in lifelong learning activities.

In this discourse, the concept of employability emphasizes the relationship between the employee's responsibility for their own career development and the employer's need for labour (Forrier and Sels 2003). Some even regard employability as ‘the job security of modern labour market’, referring to the end of traditional job security through lifelong employment (Berntson et al. 2006: 224). Future employability rather than future job security are said to be the focus of the employment relationship in the twenty-first century (see Clarke and Patrickson 2008). This view attaches importance to the individual's ability to deal with turbulent organizational environments, toning down governments’ responsibilities to directly create jobs. Instead, the profile of governmental policies and programmes is supposed to activate citizens to develop their own employability. This means that social expenditures should be rechanneled from passive to active social policies, including motivational aspects (an entrepreneurial self) as well as skills and qualifications. From this perspective, a well-functioning education system promoting lifelong learning is a key priority (cf. Morel et al. 2012).

In another policy discussion, proponents of flexicurity argue in a similar way by stressing a concept of ‘employment security’ (Wilthagen and Tros 2004). Within this discourse, employment security differs from job security where the latter concept refers to the protection of a specific job. Employment security, on the other hand, is defined as ‘security of staying employed, though not necessarily in the same job’ (Bredgaard et al. 2005: 23; see also Leschke et al. 2007: 340). However, compared to the employability discourse, the concept of employment security focuses on contextual factors per se rather than on the characteristics of the individual (e.g. the human capital). The flexicurity country par excellence, Denmark, has had a specific institutional setup in this regard which is believed to promote employment security (OECD 2004: 97–98). This institutional combination (called ‘The Golden Triangle’) comprises, firstly, liberal employment protection legislation (EPL). Low formal protection of jobs may imply high levels of job insecurity. Secondly, to mitigate the risks of unemployment Denmark has instituted very generous unemployment benefits and, thirdly, massive investment in active labour market programmes to help job searches to transit into new employment. The Danish labour market is characterized by high mobility between employment and unemployment, and especially between jobs (Berglund et al.2010). This constitutes evidence of employment security in the Danish labour market, although less is known about whether the institutions identified by the flexicurity proponents also influence the perceived employability of employees.

The employability and flexicurity discourses display similarities. One of the more obvious is the devaluation of job security, replaced by an emphasis on transitions and flexibility during the life course. However, the main difference between the two paradigms is the focus on the individual in the employability discourse, while the flexicurity discourse remains at the system level in discussing how complementary institutional configurations should be organized (Berglund 2015). In the first case, the human capital of the individual is stressed as the essential factor of labour market success. To sustain and develop the human capital, the individual is mainly responsible, while governments are required to provide necessary educational resources. In the flexicurity paradigm, on the other hand, the government should provide a security system that facilitates transitions between jobs. The starting point is not the characteristics of the individuals – their employability – but how the characteristics of the labour market system – the Golden Triangle – promote employment security.

This brief overview of the two paradigms makes clear that employees’ own perceptions of their chances in the labour market – their perceived employability – may be affected by institutional factors in the labour market. The employability paradigm emphasizes the cultivation of human capital by the educational system, while the flexicurity approach stresses the significance of unemployment insurance and active labour market policies (ALMPs). In the following sections, we elaborate on how institutional factors at the country level can affect perceived employability at the individual level. However, we will first present factors at the individual level which previous research has shown as important determinants of perceived employability.

In the discussion of policies of employability, human capital is assumed to be the most important quality for the individual's labour market success. One main indicator of human capital is usually educational level, and is thus one of the individual-level variables examined in the present analysis. A related important individual-level variable is age. Young people in general run a higher risk of unemployment compared to other age categories. For young employees, employment prospects in the external labour market must therefore be of central concern.

Previous research show rather convincingly that level of education affect perceived employability: High educated employees perceive more job opportunities than low educated employees (Berntson et al.2006; Andeson and Pontusson 2007; Berglund et al.2014). This may be explained by theories of human capital (Becker 1997) where individuals with limited education and skills are more vulnerable and have fewer opportunities in the labour market (Åberg 2002; Lindberg 2009). Besides mirroring ‘objective’ opportunities, education also have a subjective side by affecting the self-confidence of finding a new job (Andeson and Pontusson 2007).

The other individual level factor of concern is age, which also shows a strong relationship to perceived employability; younger individuals typically assess better chances in the labour market than older ones (Dixon et al.2013; Berglund et al. 2014). This pattern is probably explained by a lifecycle effect related to labour market attachment. Older workers may assess it as more difficult than younger workers to find a job equally as good as the one they currently hold because of their more firm-specific human capital related to longer tenure. Young people, on the other hand, are at the beginning of their labour market career and thus looking for a job that matches their preferences and qualifications.

Both education and age are individual-level characteristics that are usually addressed by institutions and labour market measures as factors affecting individuals’ behaviours. For example, adult educational facilities are developed to add new competencies at a later stage of a career, and many young people are especially exposed to active labour market programmes because of their higher risk of unemployment. In the coming sections, we will elaborate on the possible effects of country-level institutional variables, and how they may interact with educational level and age on the individual level.

Besides these focal individual-level factors, the analyses in the current study include some controls. We include gender as research shows that women perceive lower employability due to their generally weaker labour market position, irrespective of their level of human capital (Fogde 2011). Another factor included is country of birth: To a lesser extent, non-native workers are believed to perceive good chances in the labour market (cf. Acker 1990; Likic-Brboric et al. 2013). Two other controls are domicile and type of household. People living in rural areas are expected to have a lower level of perceived employability compared to people in urban areas (Dixon et al.2013). Single households are expected to see better labour market opportunities than dual households, because of less restriction in geographical mobility (Berglund et al. 2010). Furthermore, three factors related to the employment conditions of the employees in the study are included in the analysis. These are occupational category, employment contract and sector of employment (cf. Clark and Postel-Vinay 2009; Chung and van Oorschot 2011).

We now turn our focus to the structural and institutional factors that can affect an employee's perceived employability. We start by discussing the significance of national education systems which are generally believed to be a vital promoter of employability. Table 1 shows the education profiles of the 21 European countries in our study, defined by 3 dimensions: the proportion of tertiary educated, the type of skills obtained (general or specific) and degree of participation in adult education and training. In the empirical analyses, we use these three educational characteristics as indicators of differences in national education systems. The choice of these three educational variables is explained below.

The figures in Table 1 show differences between countries with regard to the proportion of inhabitants with completed tertiary education, defining our first country-level determinant. Especially evident are differences between Eastern Europe and countries of other regions. Despite these differences a general trend both within countries in the OECD and globally is that the proportion of people with tertiary education has increased during recent decades (cf. Schofer and Meyer 2005).

In general, level of education is believed to become more and more vital for the chance of finding a job (Eichhorst et al.2013). This would suggest that education systems that produce a high share of tertiary educated workers improve the general level of perceived employability in a country when compared to countries with a lower output of the tertiary educated. However, this rather straightforward hypothesis relies on a compositional effect of the labour force, which means that in the type of multi-level regression analysis that is conducted in the present study, no direct effect of the proportion of tertiary educated should be visible if the individual-level composition of the sample is accounted for. A complementary hypothesis of the significance of the proportion of tertiary educated on perceived employability would suggest an indirect relationship. Investments in higher education are generally believed to cause prosperity and a dynamic economy (cf. Lundvall and Johnsson 1994; Lundvall and Lorenz 2012). If this is the case, an effect of the proportion of tertiary educated on perceived employability would be present despite compositional effects.

However, a high level of tertiary educated in the labour market can have less favourable side effects. Over recent decades, the question of over-education has been salient in research and policy discussions (Åberg 2002; Verhaest and Van der Velden 2013). A high proportion of tertiary educated in a country could lead to severe competition for the places that match one's qualifications. At the individual level, this may impair the advantages of those with higher education. At the country level, the proportion of tertiary educated may therefore interact with individual level of education, reducing the effect of education on perceived employability. A similar interaction effect may appear also in relation to age, reducing the positive employability prospects expected among the young. For older age categories, formal qualification can be compensated by the human capital gained by work experience. Inflated educational qualifications may therefore hit young people the hardest.

Beside the share enrolled in tertiary education, European education systems also vary with regard to systems oriented towards general or specific training at the upper secondary level. This difference defines our second country-level determinant.

Table 1 indicates differences among countries for the proportion of the total population with general or specific skills. It is widely debated whether an education should lead to general or specialized skills for graduates (cf. Schofer and Meyer 2005; Green 2011; Eurodyce 2012). Investment in more specific skills makes individuals more dependent on specific employers, as their skills may be less transferable in case of a job loss. According to the varieties of capitalism literature, individual incentives to invest in more specific and less portable skills are therefore connected to strict employment protection and generous social security systems (Estevez-Abe et al.2001).

In this paper, we focus on the impact of specific and general education systems on perceived employability. We expect individuals to judge their skills as more transferable in a general education system than in a specific system. Consequently, employees may see more opportunities outside their current employment in a country promoting general education. The argument is based on the assumption that the education system mirrors the kind of educational profile the labour market demands.

The last indicator presented in Table 1 is the percentage of employed participating in adult education and training. Current policies promote lifelong learning as the key to improved employability (Rubenson and Desjardins 2009; Fogde 2011; OECD 2011). Looking at lifelong learning in a European context, one can discern great variation in the participation rate. Green (2011) argues that the availability of lifelong learning activities varies depending on regime type, that is, whether a liberal, social market or social democratic regime prevails. The Nordic countries especially share a long history of supporting and fostering a rich adult learning culture (Rubenson and Desjardins 2009). Studies show that people constantly engaging in career-related education are more employable, given their updated knowledge and skills (Hanushek et al. 2011). Our general hypothesis is therefore a direct positive effect of the extent of adult education in a country (indicated by the participation rate) on perceived employability. Furthermore, we believe that the extent of adult education decreases the individual-level effect of education. Adult education may compensate for low educational achievement in the standard education system, thereby equalizing differences in perceived employability between lower and higher educated. Also, with regard to age, adult education may reduce the difference between age categories on perceived employability. In this case, older age categories may benefit more from adult education than the young by adding complementary skills to their work experience.

We now turn to the significance of the country-level characteristics of labour markets for the perceived employability of employees. We will focus on the generosity of passive and active labour market programmes, the strictness of EPL and unemployment. Table 2 presents the labour market indicators for the 21 European countries in the study.

The first dimension is the generosity of passive labour market programmes (PLMPs). PLMPs include spending on unemployment insurance and early retirement schemes. The major part goes to unemployment insurance, while spending on early retirement schemes has decreased in most European countries (Eichhorst et al.2013). Our discussion will mainly focus on the possible effects of the generosity of the unemployment insurance system.

The generosity of unemployment insurance is believed to affect the job search efforts of the unemployed: a more generous benefit can prolong the unemployment spell (OECD 2010). However, the degree of generosity may affect those still employed as well. Sjöberg (2010) argues that employees may be more prone to risky transitions if there is generous unemployment insurance in place (see also Estevez-Abe et al. 2001). Thus, employees in countries with more generous unemployment benefits may be more willing to search for another job by examining their opportunities in the labour market. With less generous unemployment benefits, the employed may be more risk-averse, avoiding external mobility. Especially, this may be detrimental for people with higher education, who may be locked into jobs that do not match their competences. Consequently, the generosity of passive measures may affect the higher educated, rendering them more prone to mobility compared to employees with lower education, and ultimately, more aware of their labour market opportunities. A similar effect may also apply for older workers because of their more firm-specific skills. The generosity of passive measures may compensate for this potential lock-in effect.

The second determinant that can affect an employee's general perceived employability is ALMPs. Similar to the unemployment benefits, ALMPs are primarily aimed at helping the unemployed back to work, either by activating them in their job search efforts, preserving their human capital by training, or directly subsidizing employment (Bonoli 2010). In the present analysis, we will use an overall indicator of ALMPs. Instead of focusing on the effects on the unemployed, the present analysis will test whether there is an additional effect on employees’ perceived employability. Governments’ investments in ALMPs may affect employees’ general trust that the potential transition in the labour market will turn out well. Similar to the effect of PLMPs, this may increase attention to labour market opportunities. However, the effect may vary between categories. For instance, investments in ALMPs may have a stronger impact among young categories or people with low education compared to the older and highly educated by implying a potentially larger impact on preserving or developing their human capital (cf. Andeson and Pontusson 2007).

Table 2 also shows an OECD index of EPL, which will be used as the third country-level determinant in this study. The index refers to rules concerning the dismissal of employees: rules for individual layoffs, collective layoffs and rules for temporary contracts (see Venn 2009). Research indicates that the strictness of EPL predominantly influences the behaviour of employers by slowing down the dismissal process and making them more fastidious in their hiring decisions (Skedinger 2011). On an aggregate level, this may cause fewer job openings in countries with strict EPL, thereby affecting perceived employability. The expectation of a general effect of EPL on perceived employability received weak support from a study by Reeskens and Oorschot (2012), who found a negative impact for the young cohort. Andeson and Pontusson (2007) found no effect. Following the tendency found in Reeskens and Oorschot (2012), we expect EPL strictness to affect younger employees more strongly than older employees by reducing their perceived employability. Still, it is hard to make any conclusive predictions. Strict EPL may make workers with longer tenure (often older workers) less prone to mobility because of the protection they invested in, which, in the end, may affect their perceived employability negatively.

The final country-level indicator is unemployment. An expected result is that the unemployment rate has a negative effect on employees’ level of perceived employability (cf. Andeson and Pontusson 2007; Reeskens and Oorschot 2012). The period (2010) in focus, immediately after the financial crises of autumn 2008 and at the beginning of the Euro crises, was characterized by a rather large dispersion in unemployment (see Table 2). Unemployment rate and change in unemployment rate are related to labour market mobility (see Berglund et al. 2010) and influence intentions to change jobs, as well as perceived employability (Berntson et al. 2006; Berglund 2012). Furthermore, we expect that the unemployment rate interacts with education and age: the relative significance of education and the work experience that older employees often possess becomes more important for job prospects the higher the unemployment rate.

The data used in the current study originate from the European Social Survey (ESS), a biennial cross-sectional survey performed since 2002.¹ The empirical findings are based on Round 5 from 2010. The total survey included 27 European countries, whereas the analyses below include 21 countries due to missing country-level data. The target group were employees aged 16–65 who were living in the specific countries at the time of data collection. With regard to sample design, all countries followed the same basic principles of strict probability and representativeness. The sample size varied between 1403 and 3031, and the response rate was between 47.1% and 74.7%.

The data are nested and exist at different levels, so multi-level regression modelling is a useful approach (Hox 2012). The first level comprises individuals and individual characteristics. At the second level, country-level variables are included in the analysis. The reason for combining these two levels in an interactive analysis is that national variations are assumed to impact on the individual inhabitants.

The operationalization of perceived employability in this study focuses on employees’ assessment of the possibility to move into a new job. The data include a sample of 35,002 individuals, out of which 23,433 were in paid work. A total of 15,375 (65.6% of the employees) responded to the question: How difficult or easy would it be for you to get a similar or better job with another employer if you had to leave your current job? The respondent could answer on an 11-point scale ranging from 0 = Extremely difficult to 10 = Extremely easy.

The independent variables on the individual level comprise individual characteristics (gender, age, civil status, if born within country where employed, domicile and level of education) and work-oriented characteristics (occupational category, type of employment contract and employment sector). Educational level is defined by the International Standard Classification of Education (ISCED).

Country-level variables for the education system include the share of those with tertiary education completed as a percentage of the total 30–34 age group in 2010, the percentage of total population with an upper secondary level of general education (as opposed to vocational education), and the percentage of employed 25–64 age group who participated in adult education activities in 2009. Indicators for the labour market variables are unemployment rate, spending on active and PLMPs, and EPL. All country-level variables were collected from either the OECD Statistics or Eurostat, and are presented in Tables 1 and 2. Concerning ALMPs and PLMPs, spending was standardized against unemployment level; the percentage of gross domestic product (GDP) spending was divided by the unemployment rate. These variables give a better indication of governmental ambition and generosity (cf. Chung and van Oorschot 2011). We tested how strongly the country-level variables correlate. The strongest correlation is found for ALMP and PLMP (0.78), while all other correlation coefficients for variables are lower than 0.5.

Dependency between the individual and country level was studied in a first null model, including the dependent variable and country-level variance. The null model explained the country-specific intercepts and was used as a reference point to estimate if additional variables could improve the model. Within country and between country variances were measured using the intraclass correlation (ICC) (see Hox 2012: 15).

A means of improving our model was the inclusion of individual-level variables. These controlled for distributional differences between countries. The last steps included country-level variables for explaining variation between countries. This was followed by an attempt to introduce cross-level interactions between contextual and the focal individual variables of education and age. For exploring the hierarchical patterns, Restricted Maximum Likelihood estimation was used due to the small number of cases at the higher level. For the model fit statistics, the Bayesian Information Criterion (BIC) functioned as an adjustment to deviance for assessing model parsimony (cf. Hox 2012).

For each independent variable on the individual level, we tested if the coefficient could be found to vary significantly between countries. This was the case regarding age, and we therefore decided to use a mixed-model with age as a random slope. As the model included both fixed and random slope estimates, a mode of centring the dependent and the independent non-dummy variables was used. Hox (2012) describes the reason for centring related to problems with multi-collinearity, reducing the problem with variations of slopes for independent variables. By centring the variables, the effects were instead estimated at mean value of all variables included, which is particularly important when cross-level interactions are included in the model.

As can be seen from Figure 1, the residuals are plotted in ascending order of magnitude with their confidence intervals. The confidence intervals for Portugal, Slovenia, Poland and France cross zero. These four countries are not significantly different from the overall European outcome, while several other countries vary considerably from the mean. Moreover, the graph clearly shows that the Nordic countries have a relatively positive predicted perceived employability. To perform a robustness test of the model fit, we therefore subsequently compared all of the analyses, excluding the Nordic countries. In addition, several of the countries with negative predicted perceived employability were negatively affected by the economic downturn in late 2008. This motivated our decision to include unemployment as an important country-level variable for understanding employees’ perceived labour market opportunities (Table 3).

The analysis continues with a null model specification. This contains only the intercepts and error terms on individual and country-level, making it possible to calculate the ICC (Hox 2012). The country-level variance in our case is 8.8%. In model 1, individual-level variables control for compositional effects of each country, therefore no country-level variables were included. In the analysis, a random effect for age is included. This was the only individual-level variable with a significant between-country variance of the slope coefficient.

Furthermore, model 1 shows the effect of each variable at the individual level. Firstly, our two focal independent variables, education and age, are both significant in the expected direction: higher educated and younger age categories both have a higher level of perceived employability than lower educated and older categories. Concerning our controls, men perceive better opportunities than women, and living with a partner, being born in the country of employment or living in a big city or suburban area is also beneficial for perceived employability. Finally, we found a significant impact of occupational group (manual workers perceive lower employability than the other occupational groups) and employment sector (private sector employees perceive higher employability).

In the rest of this section, we explain the variance left across countries while controlling for various compositional effects of the individual-level variables (not shown). Table 4 shows the effect of country-level variables. In model 2, eight different analyses with each country-level variable analysed separately are summarized. The result indicates that the proportion of people with higher education and the share of participants in adult education have a positive effect on perceived employability among employees. This is in line with the theoretical expectations. However, the proportion of people with general education is negatively related to perceived employability, contrary to what we expected. Concerning the variables related to labour market institutions and unemployment, country-level spending on ALMPs has a positive effect on perceived employability, whereas the unemployment rate and change in unemployment have negative effects. EPL shows no significant relationship to perceived employability, while PLMP has a p-value below 0.10 with a positive relationship to the outcome variable. The lowest BIC value can be found in the single country-level model including ALMP, indicating that ALMP is of high importance for improving our model. When excluding the Nordic countries in the single country-level analysis (model 2), most coefficients remain rather stable even if the effect from tertiary level of education and EPL is reduced. However, only the country-level variables ALMP, percentage participating in adult education, and inhabitants with general education remain significant.

The full model (model 3) includes all macro-level variables. In this inclusive model, spending on PLMPs becomes significant on the 5% level. However, in the multivariate country-level model, the effect is reversed compared to the single level-2 model. Furthermore, the variable indicating general or specific education systems still remains significant. The coefficients for the share of tertiary educated, spending on ALMP, and change in unemployment are reduced (significant on 0.10 level), while the variables indicating adult education and unemployment level become insignificant. Concerning adult education, the reduction of the effect takes place when ALMP or PLMP is inserted in the model. This pattern remains when excluding the Nordic countries, although the general effect of adult education is overall less significant.

The final part includes five cross-level interaction effects, shown in Models 4–8 in Table 5. Each cross-level interaction is analysed separately, although both the individual-level and country-level variables are controlled for in each interaction model. None of the macro-level variables show any significant interaction with individual-level education, which the non-significant variation of this coefficient also indicate. Therefore, these results are not presented. However, significant cross-level interactions to age category were found, which is plausible due to the significant variation between countries for this slope.

Model 4 shows an interaction between age and the share participating in adult education, indicating a stronger effect for the youngest and middle-aged categories compared to the oldest age category, the larger the share participating in adult education. This result contradicts the expectation of more beneficial effects of adult education the older one gets. Model 5 shows that the difference between the middle-aged and oldest age category increases, the larger the spending on ALMPs. This result partly supports our expectation.

A similar relationship is revealed with regard to spending on PLMPs. As can be seen in Model 6, spending on PLMPs adds to age differences, in particular between the middle-aged category and the oldest. However, one must keep in mind the general negative relationships between spending and perceived employability, which is strongest for the oldest age category. In our two last models, a reversed pattern appears, and the interactions show a decreased difference between the oldest and the two younger age categories. In the first case, this is a consequence of the unemployment rate; in the second case, it is a consequence of the number of inhabitants possessing a general education.

Finally, a robustness test indicates that when the Nordic countries are excluded from the models (17 countries included), all cross-level interactions with individual age remain significant except for the percentage of inhabitants possessing a general education. Furthermore, when excluding the Nordic countries, analysis shows that the difference between the oldest age category and younger employees increases even more, the larger the share participating in adult education.

Contemporary political discourse has increasingly come to emphasize the workforce as responsible for their own employability (Jacobsson 2004; Fogde 2011). This is a shift in policy perspective from an earlier view saying that the government had the responsibility to uphold full employment by affecting general demand. Within this new paradigm one can argue that the government still has a responsibility to create the best institutional framework possible to enable individuals to stay employed. This article has focused on central institutional factors in order to study their impact on perceived employability among employees. We assessed by means of multi-level analysis how education systems and labour market policies, besides unemployment, affect employees’ perceived chances in the labour market.

Our results show the importance of the education system. In countries with a high proportion of tertiary educated, employees show more confidence in their ability to find an equal or better job. The positive effect of investment in tertiary education is in line with other research (Reeskens and Oorschot 2012). Still, the relationship between macro-level education and perceived employability is not clearly identified, especially when the compositional effect of education is accounted for in the analysis. Three scenarios may explain the result: firstly, that a high proportion of highly educated creates a more dynamic economy and labour market, that is, investment in education causes prosperity. Secondly, that prosperity possibly causes investment in education or, thirdly, that an interplay between these two appears. Clearly, the observed national variations between education systems are vital for understanding individual perceived labour market opportunities, yet we recommend further research to explore why this relationship is of importance. In addition, we find no effect of over-qualification related to a high share of tertiary educated. Neither the positive effect of educational level nor the negative effect of age on perceived employability is reduced by the share of tertiary educated in the country. This result, however, must be interpreted in the light of the categories we studied, namely, individuals in employment. These have managed to find employment, and no effect of over-education is visible in this category. However, this finding does not exclude that over-education can affect the chances to enter into employment.

Furthermore, our study shows that specific education systems are more beneficial for employees’ perceived opportunities in the labour market than general education systems. Initially, we expected a relationship in the other direction; general education would generate perceptions of larger opportunities in the labour market because of competencies that are transferable between jobs. Moreover, compared to the oldest age category, the negative effect is stronger for the younger age categories. Our result adds to studies indicating that vocational training is related to higher propensity of a successful school-to-work transition and lower youth unemployment (Eichhorst et al.2013). European countries, to which the present study is limited, mainly consist of coordinated market economies (Hall and Soskice 2001). Perhaps the result would have been different if more liberal market economies were included. Furthermore, the present study focuses on employees’ perceptions and expectations, and our results indicate that employees in education systems promoting specific skills are more confident of their value in the labour market. Still, in this study we cannot determine if the individuals make misjudgements in this regard or if these are realistic expectations.

The last indicator of the education system, the proportion taking part in adult education, is, as well, clearly positively related to employees’ perceived employability. It may be reasonable to believe that the opportunities to re-learn and reskill created by adult education convinces people that they can find a suitable job. Therefore, opportunities of lifelong learning seem important to reinforce employees’ confidence to cope with labour market uncertainties. However, we expected that adult education would be especially beneficial for older age categories. Instead, our results show this educational opportunity is more advantageous for the youngest and the middle age categories.

Concerning labour market policies, investment in ALMPs are important to help increase employees’ assessed chances in the labour market. A supportive system related to transition in the labour market seems to reduce the anxiety of failure. However, the results on PLMPs point in a contrary direction. We believed that a generous system would affect an employee's perceptions in a positive manner by making transitions less risky and facilitating the possibility of finding suitable jobs (Schmid 2008). The results indicate that the generosity is connected to less confidence among employees in finding another job. However, this effect only appears when controlling for the other factors, especially ALMPs to which PLMPs are strongly correlated.

Furthermore, the results indicate that investments in ALMP and PLMP are more beneficial for younger age categories, increasing their confidence in finding another job. Since our data also show that younger employees are more vulnerable in periods of economic downturns, investment in labour market policies and adult education can be an important factor to reduce youth employment insecurity.

Our general conclusion is that employability cannot solely be regarded as an individual phenomenon relying on individual characteristics. The increase of employability is also related to the institutional context which needs to enable the individual to remain in employment in the existing mobile and risky labour market. If institutions provide resources to cope with job insecurity, the negative effects of insecurity (stress, ill-health, less risk taking) may be reduced. However, it needs to be noted that in times of economic downturns, the unemployment rate becomes a crucial factor behind perceived labour market opportunities, predominantly among younger employees. This emphasizes general demand of labour as a necessary condition for perceived employability. Employees are required to constantly reskill as contemporary education policy emphasizes increasingly supply-oriented solutions. Europe 2020 mainly promotes a general increase in individual human capital as a vital factor in the creation of a dynamic European labour market, without mass unemployment. Our results indicate the need for a focus on the combination of individual competences/skills, national education systems and labour market policies, except for measures that increase general demand in the economy.

No potential conflict of interest was reported by the authors.

Tomas Berglund is Associate Professor of Sociology at the Department of Sociology and Work Science, University of Gothenburg. Research areas are labour markets and working life, for example, labour market mobility, job security and attitudes towards work. Another research field is political sociology with a focus on class and political attitudes.

Ylva Wallinder is a Ph.D. Candidate in sociology at the Department of Sociology and Work Science, University of Gothenburg. Her research focuses on individual's access to employability, how structural factors such as social policy, educational policy and the economic cycle intervenes with individual's perceived employment opportunities and their labour market mobility.