Job Quality

Abstract

Wages are only one dimension of the quality of workers’ new jobs. Other aspects also matter for the workers’ life and career opportunities. It has been argued that work is intimately related to other social, economic and political issues (Kalleberg 2009: 8). For instance, job insecurity not only increases workers’ levels of stress (De Witte 1999), but may also make them risk averse. Parents with unstable jobs thus may invest less in their children’s education, which affects the children’s long-term opportunities and quality of life (Esping-Andersen 2008: 75).

These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.

Wages are only one dimension of the quality of workers’ new jobs. Other aspects also matter for the workers’ life and career opportunities. It has been argued that work is intimately related to other social, economic and political issues ( Kalleberg 2009: 8). For instance, job insecurity not only increases workers’ levels of stress (De Witte 1999), but may also make them risk averse. Parents with unstable jobs thus may invest less in their children’s education, which affects the children’s long-term opportunities and quality of life (Esping-Andersen 2008: 75).

Previous research suggests that workers who had experienced a spell of unemployment were at risk of being reemployed in jobs of lower quality ( Brand 2006: 290; Dieckhoff 2011: 242). One explanation comes from the human capital theory , suggesting that displaced workers’ skills may be less useful in their new job than in their former job and workers thus have to accept offers for jobs for which they are overqualified ( Becker 1962). Alternatively, reemployment in jobs of lower quality may also come about as a consequence of long spells of unemployment, which lead to skill depreciation, or loss of self-esteem and motivation (Pissarides 1992; Arulampalam 2001).

Our hypothesis with respect to job quality highlights the effects of long-term unemployment , expecting workers with long spells of unemployment to be most vulnerable to being reemployed in jobs of lower quality (hypothesis H7, see Sect. 1.4). Previous studies further indicate that workers with tertiary education and white-collar worker s are particularly vulnerable to a decrease in job quality upon reemployment. Our models therefore include education and collar. In addition, we control for age, sex, tenure , district unemployment rate and plant.

This chapter focuses on four dimensions of the quality of the new jobs. We examine the type of contract, job security , skill match, and job authority. Wherever possible, our focus lies on the comparison between the workers’ pre- and post-displacement situation. In each section we start with a descriptive analysis and then proceed with an analysis of potential determinants of job quality. Finally, we examine workers’ job satisfaction and discuss its strongest determinants.

8.1 Contract Type

Figure 8.1 presents the contract type in which workers were reemployed. Upon reemployment, 86 % of the workers had a permanent contract , 7 % a fixed-term contract, 4 % a temporary contract, 1 % a call-off contract and 2 % other types of contract. This result suggests that the great majority have found a comparatively stable job. This result is similar to other findings from Switzerland reported by (Wyss 2009: 27). She analyses whether workers experienced a change in their contact type, comparing their pre- and post-displacement job, and finds that, depending on the company, between 79 and 100 % of the workers were reemployed in the same type of contract. The study does not, however, indicate whether these workers were permanently employed before job loss.

Fig. 8.1

Contract type of the new job. N=481

In their literature review on non-standard employment, Hipp et al. (2015: 357) point out that collective bargaining systems that ensure wage flexibility tend to go along with a low prevalence of temporary employment. Accordingly, our finding may be explained by the high level of wage flexibility in Switzerland, which in turn may be a result of the high degree of coordination in collective bargaining. In addition, our finding has to be interpreted in the context of the Swiss labor market, which is characterized by weak employment protection ( OECD 2004: 72). Accordingly, a permanent contract guarantees less job security than in most other continental European countries. At the same time, non-permanent contracts may be a better alternative to unemployment. Although fixed-term and temporary jobs usually imply fewer fringe benefits and less opportunity for career development (Green 2008: 151), they can sometimes be a stepping stone towards permanent jobs and protect workers from long-term unemployment (Gerfin et al. 2005: 820).

If we examine the contract type after reemployment by age (not shown), we find that workers aged over 60 had a significantly lower chance of being reemployed in a permanent contract (64 % as compared to 82–94 % for the younger cohorts). 14 % of this oldest age group was reemployed in a fixed-term and 21 % in a call-off contract, much more than in the younger age groups. Among the age groups between 16 and 55 the proportion of workers reemployed in the different types of contracts does not vary greatly. This suggests that workers aged over 55 and especially those over 60, who are rapidly approaching the legal retirement age, have a particular – and weak – position in the labor market. Once they lose their job, they not only face high barriers to return to the labor force, but also have to put up with large wage losses and jobs on probation.

This age-related finding is confirmed by a binomial analysis presented in Fig. 8.2. However, the effects for the older age groups are not statistically significant because of large standard errors. Other covariates, in contrast, have provided us with significant results. In fact, workers with higher levels of education are less likely to be reemployed in non-permanent contracts than workers with less than upper secondary education. In comparison to the other covariates, education has a strong effect on the reemployment contract. It thus seems that education protects workers well from being reemployed in unstable jobs. The longer workers were unemployed, the more likely they were to be reemployed in a non-permanent contract. Workers who were unemployed between 3 and 12 months were 2 percentage points more likely to be reemployed in non-permanent contracts than workers who found their new job within 3 months. Workers who were unemployed for over 12 months had a 4 percentage points higher likelihood. Although the effects are statistically significant, they are small. Men and workers with longer tenure were less likely to be reemployed in a non-permanent contract than women and workers with tenure less than 2 years. It is possible that workers with tenure of less than 2 years were already hired on a temporary or fixed-term contract in their pre-displacement job. If this were the case, it is possible that certain workers have a higher risk of being repeatedly employed in temporary or fixed-term contracts. Finally, workers from Plants 2 to 5 were much less likely to be reemployed in non-permanent jobs than workers from Plant 1. The differences between Plant 1 – the reference category – and the other plants are large and highly statistically significant. This result shows that workers in Plant 1, located in Geneva , had by far the highest risk of being reemployed in temporary or fixed-term contacts. A reason for this result may be the more competitive labor market in Geneva.

Fig. 8.2

Average Marginal Effects (AME) for a binomial logistic regression for being reemployed in a non-permanent contract. Note: N=445. Significance levels: * p < 0.1, ** p < 0.05, *** p < 0.01. Standard errors are clustered at the plant level. The dependent variable is binomial and differentiates between two outcomes: (i) reemployed in a non-permanent (fixed-term, temporary, call-off or other) job and (ii) reemployed in a permanent job. Reading example: As compared to workers with less than upper secondary education, workers with upper secondary education are 14 percentage points and workers with tertiary education 12 percentage points less likely to be reemployed in a non-permanent contract

8.2 Subjective Job Security

An additional approach to examine the workers’ job security was made by asking reemployed workers how they estimate the risk of job loss in their new job. Figure 8.3 reveals that 5 % of the workers indicated a very high and 9 % a rather high subjective risk of job loss. 14 % of the reemployed workers thus experienced strong job insecurity in their new job. However, the great majority – altogether 75 % of the workers – reported a medium, rather low or very low risk of job loss. Three out of four reemployed workers thus felt rather safe in their new position. 11 % were not able to assess the risk of job loss.

Fig. 8.3

Subjective risk of job loss in the new job. N=482

These descriptive findings correspond approximately to those on contract type. While 4 % of the workers were reemployed in temporary contracts , 5 % feel the risk of job loss to be very high. The 7 % share of workers reemployed in fixed-termed contracts roughly corresponds to the 9 % of workers indicating a rather high risk of job loss. The assumption that probably the same workers indicated working in non-permanent contracts and experiencing low job security is supported by a study based on the European Community Household Panel that shows a clear link between subjective job security and type of contract (Clark and Postel-Vinay 2009). In the 12 European countries under study, they find that subjective job security is highest for workers employed in public service jobs and the lowest for those in temporary jobs. Interestingly, they observe that the more generous the unemployment benefits are in a country, the higher is the subjective job security in all types of jobs, including temporary and private sector jobs.

With respect to the risk of job loss, a descriptive analysis (not shown) reveals a similar pattern for all age groups from 30 to over 60. Here only the youngest cohort differs from the others. Among the under-30s none of the workers experienced a very high risk of job loss, while in all other age cohorts 5–7 % of the workers evaluate their risk of losing their job as very high .

A multinomial logistic regression analysis, presented in Fig. 8.4, shows that only our hypothesis regarding unemployment duration is supported. Workers with unemployment durations of over 1 year are 13 percentage points more likely to be reemployed in a job with very or rather high risk of job loss. Thus, even after controlling for socio-demographic factors, unemployment duration seems to be strongly associated with being reemployed in more insecure jobs. Moreover, unemployment durations over 1 year provide not merely the only statistically significant result of this regression analysis but also display the strongest effect in the entire model. Education provides us neither with statistically significant nor results in the expected direction. With respect to collar, there is a small effect but it is not statistically significant and goes in the opposite direction to what we would have expected .

Fig. 8.4

Average Marginal Effects (AME) for a multinomial logistic regression for being reemployed in a job with higher risk of job loss . N=397.Note: The dependent variable is multinomial and differentiates between three outcomes: reemployed in a job where the worker is at (i) high risk, (ii) intermediate risk and (iii) low risk. Significance levels: * p < 0.1, ** p < 0.05, *** p < 0.01. Standard errors are clustered at the plant level. Reading example: As compared to workers with less than upper secondary education, workers with upper secondary education are 4 percentage points less and workers with tertiary education 0.4 percentage points more likely to be reemployed in a job with high risk of job loss

8.3 Skill Match

Skill match refers to the correspondence between workers’ skills and the requirements of their jobs. In our survey we asked the workers in a straightforward way whether they thought that the skill requirements of their new job corresponded to their own skills. Possible answers were that the skill match is adequate, that the worker is underqualified, i.e. the skills available are below the requirements, or that the worker is overqualified, i.e. the skills available are over the requirements.

Overall, two out of three reemployed workers (64 %) declared they had found a job with an adequate skill match, 12 % indicated that they were underqualified and 24 % that they were overqualified in their new job (Fig. 8.5). Thus, a majority of the workers found a job that corresponded to their skill level. However, twice as many workers were overqualified as are underqualified and thus more workers encountered a situation where the skill requirements of their new job were below the skill profile they possess. It is possible that workers’ self-reported skill match is biased because they rather overestimate than underestimate their skills. Moreover, workers may also have been overqualified in their former job. Still, this result suggests that workers who found a job in which they cannot tap the full potential of their skills were more numerous than those in the opposite situation.

Fig. 8.5

Skill match in the new job. N=474

Overqualification may hinder workers’ career advancement and usually goes hand in hand, other things being equal, with lower returns on their skills than for workers in a job with an adequate skill match. Moreover, based on data from the US and Spain it has been shown that overqualifaction is strongly associated with low job satisfaction (Johnson and Johnson 2000: 547; Badillo Amador et al. 2012: 322). In the Spanish study it has been shown, however, that underqualification is even more detrimental to workers’ job satisfaction.

Were older worker s more often confronted with overqualification in the new job than younger workers? A descriptive analysis (not shown) suggests that this is the case for the oldest category of workers: 36 % of the over 59 year olds indicated that they were overqualified as compared to 27 % for the 50–59 year olds, 22 % for the 40–49 year olds, 29 % for the 30–39 year olds and 20 % for the under 30s. Compared to the extent of the disadvantage that older workers experienced in terms of reemployment chances and wages, their disadvantage in terms of skill match is small.

Nevertheless, using inferential statistics and thus measuring the net effect of the different factors reveals that workers over 59 have a 17 percentage point higher risk of being overqualified in their new job (see Fig. 8.6). For the other age groups however, there is no statistically significant effect. Our multinomial regression analysis on being overqualified in the new job as compared to being adequately qualified or underqualified furthermore shows that workers with tertiary education were 14 percentage points more likely to be overqualified in their new job than workers with the lowest level of education. Workers with upper secondary education were about 6 percentage points more likely to be overqualified than workers with less education, although the effect does not reach statistical significance. This result is not surprising, since low-educated workers are generally unlikely to be overqualified.

Fig. 8.6

Average Marginal Effects (AME) for a multinomial logistic regression for being overqualified in the new job. Note: N=440. Significance levels: * p < 0.1, ** p < 0.05, *** p < 0.01. Standard errors are clustered at the plant level. The dependent variable is multinomial and differentiates between three outcomes: reemployed in a job where the worker is (i) overqualified, (ii) adequately qualified and (iii) underqualified. Reading example: As compared to workers with less than upper secondary education, workers with upper secondary education are 6 percentage points and workers with tertiary education 14 percentage points more likely to be overqualified in their new job

Being unemployed for 3–12 months is linked to a 7 percentage points higher risk and being unemployed for more than a year to a 14 percentage points higher risk of being overqualified in the new job as compared to workers who return to a job within 3 months. We find this effect although we control for other socio-demographic characteristics. Accordingly, the duration per se seems to have a negative effect on skill match in the new job. This finding may be due to a selection effect, unobservables driving both long-term unemployment and a decrease in job quality, or due to the fact that long unemployment durations are a negative signal to employer s. Overall, we can maintain that workers with long spells of unemployment seem to receive the low-end jobs in terms of job quality.

8.4 Job Authority

Pre- and post-displacement job authority was measured by asking workers which hierarchical function they had in their former and their new job. We define change in job authority as a situation where workers were in a position of supervision or management before displacement but without such a position after displacement or the other way around. 77 % of the reemployed respondents indicated having the same level of job authority in the new job as in the old job. 16 % experienced a downgrading and 7 % an upgrading (see Fig. 8.7). Although the majority of the workers indicated that they are reemployed in a job with the same level of job authority and some workers experienced an increase in job authority, still a considerable proportion of workers experienced a downgrading (16 % on average).

Fig. 8.7

Level of job authority in the new job as compared to the former job. N=456

These findings confirm earlier research showing that workers who experienced a spell of unemployment were more likely to experience a loss of job authority than workers who had not been unemployed. A study based on European data measuring the impact of job loss on job authority about 2 years after job loss indicates a negative effect of about 8 percentage points for Denmark and the UK ( Dieckhoff 2011: 242). Based on data from the US, the negative effect of job displacement on job authority has been estimated to be about 7 percentage points ( Brand 2006: 286).

A descriptive analysis (not shown) of the change in job authority by age reveals that the youngest cohort of the under 30 year old workers had by far the largest proportion (16 %) of workers who experienced an increase in job authority.

This result is confirmed by a multinomial logistic regression analysis, presented in Fig. 8.8, with the usual controls. This finding is not surprising since 94 % of the workers in the youngest cohort did not have a position of supervision before displacement. With respect to education, we find that workers with upper secondary and tertiary education were 8–13 percentage points more likely to experience a decrease in job authority than workers with less than upper secondary education. Again this finding is due to the fact that 90 % of the workers with the lowest level of education did not have a supervision function before displacement, as compared to 70 % of the workers with upper secondary education and 46 % for workers with tertiary education.

Fig. 8.8

Average Marginal Effects (AME) for a multinomial logistic regression for being reemployed in a job with lower job authority . Note: N=424. Significance levels: * p < 0.1, ** p < 0.05, *** p < 0.01. Standard errors are clustered at the plant level. The dependent variable is multinomial and differentiates between three outcomes: reemployed in a job where the worker has a (i) lower level of job authority, (ii) the same level of job authority and (iii) a higher level of job authority. Reading example: As compared to workers with less than upper secondary education, workers with upper secondary education are 13 percentage points and workers with tertiary education 8 percentage points more likely to be reemployed in a job with lower job authority

8.5 Job Satisfaction

The experience of the occupational transition after plant closure also depends on whether workers are satisfied with their new job, or more precisely, how job satisfaction changes as a consequence of displacement. Information about workers’ job satisfaction provides us with a subjective evaluation of more objective dimensions of job quality. In fact two recent studies suggest that job satisfaction is closely linked to workers’ job quality . Dieckhoff (2011: 243) shows for Austria and Spain that decreases in the availability of permanent contracts lead to decreases in satisfaction with job security. Based on the British Workplace Employee Relation Survey, Gazioglu and Tansel (2006: 1167) report a strong association between job security and job satisfaction as well as between income and job satisfaction. Moreover, they find that women are more satisfied with their jobs than men, low-educated workers more than highly educated ones, workers who are single more than married ones and that age has a U-shaped relation to job satisfaction.

We first describe workers’ pre- and post-displacement job satisfaction, before we try to identify the factors associated with change in workers’ job satisfaction. Our analysis takes only reemployed workers into account. Before displacement, workers’ job satisfaction was on average 7.7 points while after displacement it was significantly lower at 6.9 points. Figures 8.9a and 8.9b present the distribution of job satisfaction before and after displacement on a scale from 0 to 10 points where 0 represents “Not at all satisfied” and 10 represents “Completely satisfied with the job”. Figure 8.9a shows that 31 % of the workers indicated a job satisfaction of 8 points and 84 % between 7 and 10. After displacement, 24 % of the workers reported a job satisfaction of 8 points and 64 % indicated a level of satisfaction between 7 and 10. At the same time, 20 % of the workers reported a job satisfaction of 5 or 6 – a substantially larger proportion than before displacement (8 %). The displacement has thus flattened the distribution of job satisfaction and shifted it to the left .

Fig. 8.9a

Distribution of job satisfaction before displacement. N=474. Reading example: Before displacement, 15% of the workers indicated being completely satisfied with their lives

Fig. 8.9b

Distribution of job satisfaction after displacement. N=474. Reading example: After displacement, 12% of the workers indicated being completely satisfied with their lives

To determine which factors are linked to workers’ changes in job satisfaction, we run an OLS regression (see Fig. 8.10). We use within-individual change rather than level of job satisfaction as a dependent variable since satisfaction measures are typically subject to selection bias . It has been argued that most individuals have a baseline level of satisfaction that remains relatively stable across the life course , although major life events may lead to oscillations around this baseline level ( Clark et al. 2008). Our model contains the covariates sex, education, age, civil status, plant, and measures which assess the change between before and after displacement in terms of wages, weekly working hours and job authority .

Fig. 8.10

Coefficients for an OLS-regression analysis for change in job satisfaction for the reemployed. Note: N=332. Significance levels: * p < 0.1, ** p < 0.05, *** p < 0.01. Standard errors are clustered at the plant level. Reading example: As compared to women, men experienced a decrease in job satisfaction by 0.11 points

We find a strong positive effect of tertiary education: reemployed workers with tertiary education experienced on average an increase of over 1 point in life satisfaction as compared to workers with less than upper secondary education. In addition, workers who were single experienced a more positive change in job satisfaction than workers who were married. Workers in Plant 2 experienced a much more positive development of job satisfaction than the workers in other plants. This may be due to the fact that they had known for some time that the plant was in a difficult economic situation – so difficult that they even accepted wage cuts the year before the plant closed – and thus suffered from job insecurity. A 10 % increase in wages was associated with an increase in job satisfaction of 0.3 points. Moreover, longer working hours have a negative effect on job satisfaction: an increase of 10 % in working hours was associated with a decrease in job satisfaction of about 0.7 points. In contrast, changes in job authority do not provide significant effects. In sum, this analysis shows that tertiary education and, curiously, being single cushioned workers against a decrease in job satisfaction. Changes in wages also mattered for workers’ job satisfaction, but less than changes in working hours .

8.6 Conclusion

The analysis of the quality of the workers’ new jobs has provided us with a more in-depth picture about the adjustment process that workers undergo after plant closure. However, we do not find as bleak a picture as might have been expected. Interestingly, 86 % of the workers were reemployed in a permanent contract . However, because of the weak employment protection in Switzerland a permanent contract does not automatically imply high job security – but does give higher security than temporary or fixed-term jobs. An analysis of the subjective risk of job loss suggests that while half of the workers think they have a low risk of losing their job, about 40 % of the workers indicate a high or a medium risk.

We are not able to identify one single factor that best explains the decreases in all dimensions of job quality . However, what we find is that long unemployment duration is associated with reemployment in non-permanent contracts, a high risk of job loss and overqualification – but is not correlated with changes in job authority. More precisely, our analysis has shown that an unemployment duration of over 1 year goes along with a substantially higher risk of experiencing a decrease in job quality. Higher levels of education are linked to two dimensions of low job quality, overqualification and lower job authority. Finally, an older age seems to be the driver of overqualification, lower job authority and a lower social status.

In a nutshell, unemployment duration seems to be particularly relevant with respect to the quality of displaced workers’ new jobs. This finding corroborates our hypothesis H7. However, it needs to be recalled that unemployment duration is not exogenous and is thus likely to be determined by unobservable variables.

This chapter shows that losing one’s job goes along with a decline in job quality. From our descriptive analysis we can conclude that in terms of all job quality dimensions some workers are disadvantaged in their new job. Some were reemployed in better jobs – for instance with higher job authority – but those who experienced an improvement are clearly fewer than those who experienced a deterioration. We therefore can maintain that plant closure has a negative impact on workers’ career s in terms of job quality. Nevertheless, the outcome is not the same for everyone and a more precise look at the data shows that some workers were reemployed in jobs of about the same and some in jobs of better quality.