Early Retirement and Exit from the Labor Force

Abstract

In the light of older workers’ difficulties in returning to employment after job displacement it is fundamental to analyze a third pathway older workers may take beside reemployment and unemployment: early retirement. The question here at stake is whether workers are involuntarily pushed out of the labor market and subsequently suffer from social exclusion or whether they experience this option as an alleviation of their critical situation. Some studies have shown that older job seekers choose this pathway primarily as a better alternative to long-term unemployment (Chan and Stevens 2001; Ichino et al. 2007). Other studies found that pull factors such as generous early retirement plans, being financially well off or having an economically inactive spouse may incite older workers to leave the labor force before the official retirement age (Knuth and Kalina 2002: 414).

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.

In the light of older worker s’ difficulties in returning to employment after job displacement it is fundamental to analyze a third pathway older workers may take beside reemployment and unemployment: early retirement. The question here at stake is whether workers are involuntarily pushed out of the labor market and subsequently suffer from social exclusion or whether they experience this option as an alleviation of their critical situation. Some studies have shown that older job seekers choose this pathway primarily as a better alternative to long-term unemployment (Chan and Stevens 2001; Ichino et al. 2007). Other studies found that pull factors such as generous early retirement plans, being financially well off or having an economically inactive spouse may incite older workers to leave the labor force before the official retirement age (Knuth and Kalina 2002: 414).

Starting from the hypothesis H2 (see Sect. 1.4) that push factors better explain the mechanisms underlying older workers’ transition into early retirement than pull factors, this chapter analyzes displaced workers’ transition into early retirement, focusing on workers over 55 but excluding those workers who reached the official retirement age (that is, 64 for women and 65 for men). We first present the overall probability of retiring early. We then describe the factors that may be linked to the workers’ transition into early retirement. In a next step we analyze the effect of explanatory factors by means of a regression analysis. Finally, we discuss whether the early retirement plans, which were provided by the plants, may have affected workers in their decision to retire early. We use well-being measures to help us in our interpretation.

4.1 Transition into Early Retirement

Our data reveals that a considerable proportion of workers went into early retirement. Taking workers of all age cohorts together, 8 % retired early, and among the workers aged over 55, the proportion of early retirees is 32 % – almost a third – while only 31 % of the workers in this age category were reemployed and 37 % were unemployed.

In a next step we examine whether this pathway was chosen by the workers rather because of push or pull factors. In Table 4.1 we present the descriptive analysis of the proportions of workers aged 56 to 64 in early retirement, reemployment and unemployment by the number of applications written, pre-displacement wage, presence of an economically active partner, plant and sex.

Table 4.1 Proportion of workers aged 56–64 in early retirement, reemployment and unemployment by different characteristics (in %)

We find that older worker s who retired early are more strongly represented among those who wrote less than 50 applications than among those who applied more than 50 times. For older workers in unemployment the opposite is the case. This suggests that the early retirees did not search for a job as intensively as the older workers who returned to employment or who still were unemployed.

Regarding the workers’ pre-displacement wage, workers who retired early have about the same probability of having an income over CHF 7000 (49 %) as reemployed and unemployed older workers (each 51 %). In addition, we find that 23 % of the workers who retired early had a pre-displacement income of under CHF 6000 while this is true for only 8 % of the reemployed. The pattern contradicts the possible expectation that workers with higher wages are more likely to retire early and that workers with lower wages are constraint to carry on working until their official retirement age. With respect to the availability of an economically active partner, the early-retired are much more likely to be without a working partner (51 %) than the reemployed (14 %) and slightly more likely than the unemployed (42 %). Thus, having a spouse who is not working may be an incentive for older workers to step back from the active work life.

With respect to the plant we find that among the early retirees most of them were employed in Plant 4 (76 %) before displacement . In contrast, workers from Plants 1, 2 and 5 make up together only 8 % while workers from Plant 3 represent 15 % of all early retirees. If we turn to reemployment it is striking that none of the older worker s in Plant 1 in Geneva was reemployed. This finding may be explained by workers being residents of France where they have access to unemployment benefits for an unlimited duration beginning at age 57.5. Evidence for this assumption is provided by the fact that workers from Plant 1 constitute 40 % of the unemployed workers. Workers from the other companies represent between 17 and 35 % of the reemployed. Finally, women constitute only 3 % of the early retirees while they make up 12 % of the reemployed. This may indicate that women less often have the opportunity to retire early because of financial reasons.

4.2 Determinants of Early Retirement

In order to measure the net effect of the factors discussed above, we continue with a multinomial logistic regression analysis, presented in Fig. 4.1. Our dependent variable is “going into early retirement” as compared to being “reemployed” or “unemployed or out of the labor force”. We run a model for going into early retirement, relative to being reemployed. We enter the variables age, sex, education, plant (as a proxy for the redundancy plan ), wage and economic activity of the partner. It was not possible to include the number of workers’ job applications in the model because this information is available for too small a number of early-retired workers to produce robust results. We restrict our analysis to workers aged 56–64 and exclude workers who retired regularly. In order to test the robustness of the results, we also estimated two types of binomial models where we compare first workers who retired early with workers who were reemployed and second workers who retired early with workers who were reemployed, unemployed or out of the labor force. The results – that are similar to the findings from the multinomial analysis .

Fig. 4.1

Average marginal effects (AME) for a multinomial logistic regression for going into early retirement for workers aged 56–64. N=141. Note: The dependent variable is multinomial and differentiates between three outcomes: (i) early-retired, (ii) reemployed and (iii) unemployed or out of the labor force. Only workers aged 56 to 64 are included and the regular retirees are excluded. Standard errors are clustered at the plant level. Significance levels: * p < 0.1, ** p < 0.05, *** p < 0.01. Significance levels: * p < 0.1, ** p < 0.05, *** p < 0.01. Reading example: With each year in age workers’ probability of retiring early increases by 6 percentage points

Our analysis shows that the closer the workers are to the regular retirement age, the higher is the likelihood of retiring early, a result that is entirely plausible as with each year the cuts in pension benefits that workers have to accept when retiring early becomes less burdensome. We find a large and significant effect for sex, men being 32 percentage points more likely than women to retire early relative to being reemployed, unemployed, or out of the labor force. Workers with tertiary education were 14 percentage points more likely to retire early than workers without upper secondary education. For the variable plant we use Plant 2 as reference category since this plant was the only one without a redundancy plan . Workers in Plants 1 and 5 seem to be slightly (2–3 percentage points) less likely to retire early as compared to workers in Plant 2 (reference category). Workers in Plant 3 are 12 and – most noteworthy – workers in Plant 4 60 percentage points more likely to retire early than workers in Plant 2. A possible explanation for these results may be the redundancy plans. Whether they are relevant will be discussed later. With respect to wage we find that workers with an income of CHF 5000–6999 are 15 percentage points less likely to retire early as compared to those with an income below CHF 5000. Interestingly, workers in the highest wage category (CHF > 7999) are the least likely to retire early relative to becoming reemployed. This result suggests that workers with higher wages have more enjoyable jobs and continue to work until the regular retirement age. Finally, workers with an economically active partner are 7 percentage points less likely to retire early than workers with a non-active partner.

When we tested different models introducing all or only a selection of the discussed covariates, we observed that the introduction of the variable plant into the model strongly increases the models’ goodness of fit (pseudo R²). This suggests that the differences between plants decisively contribute to the explanation of workers’ entry into early retirement. Perhaps workers were incited to retire early by the redundancy plans negotiated between the employees’ and employer s’ representatives, some redundancy plans indeed containing extensive early retirement provisions.

In Table 4.2, we present an overview of the redundancy plan s and calculate the benefits for an exemplary worker aged 60 with 15 years of tenure and a pre-displacement wage of CHF 6000. Plant 1 located in Geneva provides men over the age of 62 and women over the age of 61 with early retirement benefits that depend on their tenure. Workers with at least 25 years’ tenure received a one-off payment of CHF 10,000 while workers with at least 36 years received CHF 20,000. Moreover, workers received a top-up of their unemployment benefits up to 60 % of their former wage. This measure is of relevance only for workers who live in France and who are subject to the French unemployment benefit system. Workers living in Switzerland do not need this provision since they have a replacement rate of at least 70 % of their former wage if they apply for unemployment benefits. The worker in the example in Table 4.2 does not receive early retirement provisions in Plant 1 since s/he is below the age limit.

Table 4.2 Retirement-related features of the redundancy plan s

Plant 2 close to Biel did not provide a redundancy plan. In Plant 3 in North-Western Switzerland male workers received early retirement benefits from the age of 60 and female workers from the age of 59. The provisions amounted to 70 % of the former wage inclusive of the 13th monthly wage. This leads for our exemplary worker to a monthly income of CHF 4500. Workers in Plant 4 in the Canton of Bern could request early retirement provisions from the age of 56.5 years. Dependent on their age, they received 90–100 % of the pension benefits that they would have received if they had gone into regular retirement. Since the amount of the regular retirement benefits depends on the requirements of the decentralized pension funds, we can only approximately estimate the amount of the benefits for early-retired workers in Plant 4, which is about CHF 2500.¹ In Plant 5 in North-Western Switzerland received early retirement benefits from the age of 58. The benefits depended on the tenure . For our exemplary worker this scheme results in a monthly income of CHF 1500.

To come back to the results of our regression analysis presented in Fig. 4.1, we contend that the early retirement provisions in the redundancy plan do not offer a clear-cut explanation of the much higher likelihood of retiring early of workers in Plant 4 (Bern) and the slightly higher likelihood of workers in Plant 1 (Geneva) as compared to workers in Plant 2 (Biel). Regarding Plant 4, the early retirement benefits are not as generous as in Plant 3 (NWS 1) but the age limit for receiving early retirement provisions is very low, i.e. 56.5 years. This condition probably led to the early retirement of a large proportion of workers. With respect to Plant 1 the higher propensity to retire early may stem from the fact that almost half of the workforce live in France where the transition into this gateway is facilitated by social security institution s. While in France workers over the age of 50 are entitled to 36 months of unemployment benefits, in Switzerland the maximum duration is 24 months (Schwab and Weber 2010: 40).

Finally, given the generous early retirement provisions, we would have expected workers from Plant 3 to be particularly likely to go into early retirement. This is, however, not the case and we therefore conclude that in addition to the redundancy plan other mechanisms seem to be at work. A possible factor behind the differences may have been how the provisions offered by the redundancy plan were communicated to the workers. For instance, it is conceivable that many older worker s in Plant 3 (NWS 1) were not as well informed about their opportunities as workers in Plant 4 (Bern). Another possibility is that there was an effect of conformity motivating workers in Plant 4 to retire early together.²

Do our results rather point to the presence of pull or push mechanisms? On the one hand, the impact of age and wage suggest that push mechanisms are at work: as we have seen in the section on reemployment , older worker s have much more difficulty in finding a job. Exiting the labor force into early retirement thus may be a smooth pathway after job displacement . The finding that workers with lower wages are more likely to retire early may suggest that high wages do not have a pull effect. Alternatively, it is possible that workers with lower wages had bleaker reemployment prospects than workers with higher wages – probably related to their occupational status – and therefore had to retire early.

On the other hand, the result from the descriptive analysis regarding the number of job applications and the result from the regression analysis that belonging to Plants 1 or 4 has a positive effect on retiring early suggest that pull mechanisms are at work. The fact that the early retirees were less likely to apply more than 50 times for a job than workers who were reemployed or unemployed indicates that many of them did not search highly intensively. Moreover, the higher likelihood of retiring for individuals working in a plant with early retirement provisions suggests that, not surprisingly, these benefits had an inciting effect.

Overall, it is therefore difficult to tell whether pull factors are more important than push factors. As a last resort to find out whether one mechanism predominates over the other, we analyze the early retirees’ change in subjective well-being between their pre- and post-displacement situation. We compare the outcome for the early retirees with the one for the reemployed and unemployed workers aged 56 to 64. The results are presented in Fig. 4.2.

Fig. 4.2

Change in life satisfaction before and after displacement (in points on an 11-point scale) for workers aged 56–64. N: Early-retired=63, Reemployed= 54, Unemployed=66

We assessed workers’ life satisfaction on a scale from 0 to 10 where 0 stands for not satisfied at all and 10 for very satisfied. We measure change in life satisfaction between the situation before displacement when workers were still employed and the moment of the survey. We calculate the difference between the two values by simply subtracting one value from the other. A negative value means that workers were more satisfied before displacement than after, and a positive value that they were more satisfied after displacement than before. A value of zero indicates that there was no change.

We find that the early retirees experience an average decrease in life satisfaction by 0.1 points, the reemployed by 0.3 points and the unemployed by 2.6 points. With respect to the distribution of the change in well-being, Fig. 4.2 suggests that for the early retirees the 25- and the 50-percentile levels are at zero change in life satisfaction and the 75-percentile level at 1 point increase. Accordingly, half of all early retirees lie within a very narrow range of almost no change in life satisfaction. At the same time, the dots outside of the whiskers indicate that the early retirees experience strong decreases or intermediate increases in life satisfaction. The pattern of the reemployed older worker s is similar: half of the workers experienced either no change in life satisfaction or a decrease in 1 point. In contrast, the unemployed clearly experienced the strongest drop in life satisfaction, the median being a 2.5 points decrease and the 25-percentile level a 4 points decrease. Taken together, these descriptive results show that among the workers over 55, the early retirees were the least negatively affected by plant closure. This suggests that they were rather pulled than pushed out of the active labor force.

4.3 Exit from the Labor Force

A small proportion of workers (n = 23) in our sample quit the labor force for training (n = 9), childcare (n = 2), disability (n = 7) or other unspecified reasons (n = 5). The theory suggests that there is an interaction effect between sex and civil status explaining labor force dropout, married women being most likely to quit the labor force. However, because of the small number of observation, we can only carry out a very rough analysis and the results are not in line with the theoretical predictions. In fact, among those in training, men and – not surprisingly – young workers are overrepresented. If we consider those who do childcare, both genders are represented. Among the disabled, the majority worked in typically manual occupations (craft worker s, machine operator s or elementary occupations ) and were over 55 at the moment of displacement . In the category of other situations no particular socio-demographic pattern is identifiable.

With respect to these workers, it may be interesting to examine if we can find evidence for whether these workers were pushed out of the workforce (e.g. because they could not find a job) or whether they voluntarily quit the labor market. Because of small n, we cannot thoroughly analyze this question, but descriptive statistics can still give an idea of their potential motives. Among the labor force dropouts two-thirds searched for a job. This suggests that some of the labor force exits were a consequence of low reemployment prospects. Among the workers in “other” situations one searched for 7–9 months, one for 13–18 months and one for 19–24 months. One worker who does childcare searched for 5–6 months. In contrast, workers who have gone into training did not search longer than 2 months and among workers with disability no one indicated having searched for a job. If we look at the number of job applications that workers who quit the labor force have written, we find that among those who went into training, most of them wrote fewer than six application letters. The same is true for those with disability benefits. In contrast, the two workers in childcare and the workers in “other” situations wrote either 50–100 or more than 100 applications. This shows that the workers in childcare or other situations tried intensively to find a job and exited the labor market only when they were not successful.

Finally, an interesting finding is that workers both with an economically active partner and with an inactive partner tended to leave the labor market. This indicates that not only workers who have a second income – and thus are financially secure – dropped out. Overall, it seems difficult to generalize on the situation of the labor force dropouts. From the small number of observations available we can only maintain that the workers in this group followed individualized strategies and probably had a variety of reasons to quit the labor force.

4.4 Conclusion

In this chapter we have shown that although older worker s faced considerable barriers to returning to the labor force, the evidence suggests that there were not only push but also pull mechanisms that led workers to choose to retire early. In fact, workers with a non-working partner and with early retirement provisions were more likely to retire early than those with an economically active partner and without provisions. Accordingly, our findings seem to contradict our hypothesis H2.

How do our findings compare with the literature about early retirement? Our outcomes do not completely confirm the results from most continental European countries that workers are involuntary pushed into early retirement. By contrast, they seem to conform to the observation by Dorn and Sousa-Poza (2010: 432) that the pathway into early retirement in Switzerland is often voluntary. Understanding the link between job displacement and retirement is crucial since it may lead to informed policy decisions about how to support older displaced workers and how to promote longer working lives (Tatsiramos 2010: 517). Teyssier and Vicens (2001: 25) contend that facilitating early retirement is a success in addressing old-age poverty, but does not enhance the social integration of workers towards the end of their career s. It thus seems that there is a trade-off between these two concerns. A potential solution would be that in the case of a plant closure early retirement provisions could not exclusively be accessed immediately after the closure but for several years after displacement. This would motivate older workers to first search for a job and only transit into early retirement if they were not able to find one. If – as is the case with most redundancy plan s – workers have to decide directly after the plant closure whether to retire early, they are likely to choose the most secure option and accept the early retirement provisions instead of trying to return to the labor force.

This chapter shows that for older workers’ who belonged to the age group required for early retirement usually chose this option; there is little evidence that the early-retired workers were forced into this gateway. For policy decisions this finding implies that offering older displaced workers early retirement plans, ideally including financial contributions to the social security system, allows them to overcome job loss without major calamity. If, in contrast, workers over 55 years have to search for a job, they are likely to suffer a difficult occupational transition often encompassing long-term unemployment and – as we will see in the next chapters – reemployment in precarious jobs.