Navigation – Plan du site

Voluntary early retirement: Between desires and necessities

Retraits anticipés du marché du travail : entre désir et nécessité
Retiros anticipados del mercado laboral : Entre deseo y necesidad
Nathalie Burnay
Cet article est une traduction de :
Retraits anticipés du marché du travail : entre désir et nécessité [fr]


L’objectif de cet article sera de déterminer quels sont les principaux facteurs personnels et professionnels qui conduisent à vouloir mettre un terme définitif à sa carrière professionnelle avant l’âge légal de la retraite.
Dans un premier temps, il s’agira de déterminer les profils des travailleurs tentés par un départ anticipé : quelles sont les logiques sociales qui permettent de différencier ce désir de retrait, inégalement réparti dans la population ?
Dans un second temps, l’analyse portera sur les raisons principales invoquées par les travailleurs pour quitter la vie active avant l’heure : raisons liées à un état de santé, raisons professionnelles (travail physique trop lourd, charges psychosociales, horaires de travail) et raisons privées.
Cet article s’appuie sur des données récoltées auprès de 812 travailleurs, répartis dans douze secteurs d’activité différents en Belgique en 2005.

Haut de page

Texte intégral

1. Demographic Aging and Societal Issues1

  • 1 This article is based on the results of the CAPA study, jointly funded by Le Fonds Social Européen (...)

1Since the mid-nineties, the aging of the population has worried many scientists working in different areas, who see the relationship between the number of people in the working population and the number of retirees as reversing on a long-term basis: the increase in life expectancy combined with the generational effects related to the pyramid of the elderly generate a generation of seniors whose retirement must be paid for. Thus, the reduction in the number of people active in the work force will lead to an increase in the dependency rate (Griffiths, 1997) and the public costs for retirees and healthcare (Kilbom, 1999).

2International bodies will gradually look at the issue of age and make it a priority. Starting in 1995, the OECD sounded the alarm, supporting its findings with demographic data: funding retirees and the entire social security system would be put back into question by the exponentially increase in expenditures. In this context, the OECD recommended « active aging » and the lifting of all barriers to extending activities, whether they are the early retirement packages or more radical retirement systems that discourage activity after a certain age (OECD, 1995 and 1998). In 1999, the European Union made the issue of aging workers a major item on the political agenda. In its 1999 report, the European « The European Job Market in Light of Demographic Development », the 50-64 age group is identified as being the main labour pool for the future. The Lisbon Summit in March 2000 proposed an active employment policy to the Member States in order to reach employment levels of 70% by 2010. The Stockholm Summit in March 2001 confirmed the directions of Lisbon, but also set a specific objective of 50% for the average employment rate of men and women aged 55 to 64 for 2010.

3If, from a macro-economic perspective, the collective management of the aging population becomes translates into the necessary raising of employment rates, this issue also has repercussions on the life of companies and employees: the risk of an increasing lack of skilled, experienced employees in companies (van der Velden, 1994), losses of social references for workers (Crespo & Beausoleil, 1999).

4However, employment rates of those 55 and older have drastically dropped in all western countries, especially in France and Belgium (Guillemard, 2003). In Belgium in 2004, employment rates of men aged 55 to 59 and 60 to 64 were only 53% and 20% respectively in 2004, even though the legal retirement age is set at 65.

2. Multiple Causes

5Many surveys have shown that employees wish to leave the working world well before the legal retirement age. The French survey Emploi, conducted in 1996, shows a staggering majority (71%) of employees who had taken early retirement did not want to continue with professional activities (cited by Guillemard, 2003). In a recent Belgian survey carried out by the VUB team (Elchardus & Cohen, 2003), the ideal age in Belgium to stop working was between 59 and 60. The initial results of the SHARE European survey showed how France was among the countries in which early retirement was most often taken (Blanchet & Debrand, 2005).

6There are many reasons and factors for low employment rates among seniors. It is difficult to truly pinpoint these reasons, given that the effects of interaction come into play. The scientific literature seems can be divided into three groups of explanation, which has complicated the attempts to come up with a single model.

2.1 Individual and Family Reasons

7Individual and family reasons are the first approach looked at for why people take early retirement. Stopping work in this case is seen as a positive transition for most workers who are responding to personal, family or social aspirations: the spouse’s professional situation, involvement with children and grandchildren, free time to carry out recreational activities (Hardy & Hazelrigg, 1999; Schultz, Morton & Weckerle, 1998; Szinovacz & De Viney, 2000). Leaving work is then explained by personal factors that go to the heart of broader social processes.

8Financial constraints and health also appear to be good predictors (Adam, 1999; Adam, Prescher, Beehr & Lespito, 2002; Barnes-Farrell, 2003; Feldman, 1994; Walker, 1985; Gratton & Haug, 1983, Friedman & Orbach, 1974; Parker, 1980; McGoldrick & Cooper, 1980, Saurel-Cubizolles et al., 1999, Barnay & Jeger, 2006). The more a family faces financial constraints, the less likely a person will permanently leave the job market: the loss in salary related to change in status cannot be taken on by the household. Changes in health contribute to leaving work, especially when the job involves difficult and strenuous working conditions (Lund & Borg, 1999). Health problems are thus a major reason for leaving the job market, either through sick leave processes, in particular disability, or through other available channels. Leaving the job market for health reasons occurs for 12% of men and 8% of women with rhumatological pathologies, and 7% of men and 5% of women for cardiovascular pathologies (Molinié, 2006). However, a significant link exists between these types of retirement and the difficulty of the professional activity (Molinié, 2006).

9Other individual data can also help explain early retirement, but their effect in the retirement models varies based on the analyses. This primarily means sex and age (Walker, 1985; Gratton & Haug, 1983; Szinovacz & De Viney, 2000, Van Solinge & Henkens, 2005). These two factors combine with others, resulting in powerful interactions and making it relatively difficult to evaluate the actual contributing factors. For example, aging can sometimes have a hand in leaving work early, but sometimes it has the opposite effect, and can be a powerful incentive for staying at work. To better understand the impact of age, it must be combined with status, working conditions, family composition… As well, the effects of selecting the population can also have an effect and influence the results of surveys: only active workers answer certain surveys, which makes it impossible to make generalizations about age-related data.

10The family structure, in particular the professional situation of the spouse, is also a determining factor in decision-making in a correspondence of social time (Szinovacz & De Viney, 2000, Van Solinge & Henkens, 2005, Gustman & Steinmeier, 1994 cited by Hansson R., Dekoekkoek P.D., Neece W. & Patterson D., 1997). The spouse affects the decision to leave when the spouses decide to unite their destinies. The future thus develops based on a common destiny instead of professional requirements.

2.2 Organizational Reasons

11Strictly professional factors can also favour early retirement from the job market. For Hansez (2005), these factors are three-fold.

12First of all, organizational context, or structural changes of the professional context (Bartel, Sicherman, 1993) can lead the worker to not want to go through the heavy, costly process of changing his working environment.

13Next, changing working conditions also influence early retirement, especially if there are increased psychosocial risks at work (Herzog, House & Morgan, 1991), by acting on the ability to resist work pressures. The difficulty and intensification of work also influence the choice of retiring early: tired bodies long for retirement (Feldman, 1994). In a context where working conditions are becoming tougher, many surveys have pointed to the risk of extreme weakening of seniors through the effects of age and generations (Molinié, 2006).

14Lastly, the lack of career opportunities, social recognition and personal development also contribute to favouring early retirement (Feldman, 1994; Schmitt, White, Coyle & Rauschenberger, 1979). Moreover, these different items can be combined or mutually strengthen each other. As well, professional dissatisfaction can reinforce the choice (Schmitt, White, Coyle & Rauschenberger, 1979), but this factor is closely correlated to social recognition as well as motivation at work.

2.3 Societal Reasons

15The different employment policies carried out since the late 70s have most likely contributed to the development of early retirements from the job market. Indeed, during the seventies and eighties, all European countries were facing the economic crisis head on. Unemployment rates soared in a context of drastic industrial restructuring. Public policies, especially in Belgium, immediately showed a marked preference for young unemployed workers over older workers (Simoen, 1980). Workers 50 and older were encouraged by the government to leave the job market through early retirement, unemployment, inability to work and career sabbaticals. The OECD (2003) confirmed this diagnosis of setting aside seniors by moving up the retirement age and implementing early retirement.

16These policies are based on a large consensus between political bodies, social partners and workers: the purpose of these practices was to give priority to young people to the job market by cutting back the senior workforce. However, the impact of these measurements is still difficult to determine, and especially to measure effectively.

17As for employers, the early retirement model also has legitimized the practices of cutting back the workforce » based on age, even more so when a compensation system is used that is based on seniority, which considerably increases salaries as people get older. This becomes for the employer a way of reducing work costs. Directly or indirectly, workers over 40 are faced with discriminatory practices upon hiring: employability drastically reduces with age. Two types of discrimination based on age can thus be identified: direct and indirect discrimination (Leleu, 2001). The first refers to measure that target older workers based strictly on age. An example of direct discrimination, for example, is when age limits are specified in job offers or training. This form of discrimination is subject to a specific regulation, which introduces measures for protecting workers. Indirect discrimination is a measure that is not directly related to age but negatively and disproportionately affects older workers versus other age groups (Drury, 1993). This more subtle form of discrimination is more common than the first direct form of discrimination (McEvan, 1990; Lackzo & Phillipson, 1991; Itzin & Phillipson, 1993). Stereotyping and prejudices that colour the selection of candidates for jobs are forms of indirect discrimination since they are likely to influence attitudes and behaviours, both for the person being subjected to them, as well as those who are the instigators. Many surveys on these social stereotypes regarding older workers have been conducted since the initial American studies of the fifties (Tuckman & Lorge, 1952; Kirchner, Lindbom, & Paterson, 1952; Kirchner & Dunnette, 1954). The stereotypes studied in the different surveys concern both the development of physical and mental health of an older worker and his professional integration. Certain stereotypes are well ingrained in people’s minds: physical difficulties (Rosen & Jerdee, 1976; Taylor & Walker, 1994; Hasell & Perrewe; Palonimi & Tikkanen, 2000; Mc Gregor, 2002), cognitive or mental difficulties (Aaronson, 1966; Rosen & Jerdee, 1976; Taylor & Walker, 1994), resistance to change (Rosen & Jerdee, 1976; Taylor & Walker, 1994; Lyon & Pollard, 1997; Remery, Henkens, Schippers, & Ekamper, 2003), difficulties dealing with younger people (Taylor & Walker, 1994; Hassell & Perrewe, 1995) as well as learning difficulties (Rosen & Jerdee, 1977; Taylor & Walker, 1994) are largely confirmed by the studies.

18This macro level will not be discussed in this study.

19The article discussed the individual and organizational dimensions in a given analysis model. It will involve showing that, beyond the causal relationships already highlighted in the literature, early retirement from the job market responds to contemporary logic, which is currently affecting the whole of western society. In this sense, early retirement is only one particular way of understanding contemporary societal changes that connect the individual to the world of work. Thus, we will see to what extent retiring from professional life is both the result of deteriorating working conditions, especially in terms of increased professional stress (Gollac & Volkoff, 2000; Neboit & Vézina, 2007), and of a specific relationship that individuals maintain with work. The latest surveys on values (European Value Surveys, World Value Survey, International Social Survey Programme) only confirm that, although work remains a fundamental value, it is not the only form of symbolic investment of the individual. We will then talk about identify polycentring. These two events will then allow us to understand to extent to which early retirement oscillates between desire and necessity.

20This article is organized in two parts. The first part attempts to determine the profiles of workers who want early retirement: What are the social factors that determine the desire of whether or not to retire, which is unequally distributed among the population?

21In the second part, the analysis focuses on the main reasons given by workers for quitting their working life early: health-related reasons, professional reasons (heavy physical work, psychosocial stress, work schedules) and personal reasons.

3. Methodology

22This article is based on data collected from 812 workers, in 12 different sectors of activity: chemical industry, tool or machine manufacturing, car industry, construction, large-scale distribution, restaurants, finance and insurance, research & development, public administration, teaching, healthcare and social sector. The objective was to cover the job market as broadly as possible. The subjects were selected by occupational health physicians as part of an annual consultation that took place between October 2004 and June 2005. Two occupational protection and prevention departments were thus directly involved in the research project. For researchers, there were two reasons for involving these field doctors: to provide more detailed information and to create the questionnaire based on their professional practice, to ensure the questionnaire is filled in. For the doctors, their interest lied in a better understanding of the end of career period. It thus involved understanding a structural change that had been a recent concern in their professional practice: the aging of the work force causes new pathologies and, in particular means new management of age in a company. Thus, although job restructuring was possible in specific cases in the past, given the proliferation of situations, it leaves them slightly impotent. Therefore, stopping work permanently for health reasons tend to be quite common.

  • 2 The models for early retirement from the job market that attempted to identify individual component (...)

23The questionnaire was given in two stages: in the first stage, the questionnaire was self-administered by respondents in a way that would ensure maximum autonomy in the answers collected; in the second stage, occupational health physicians checked the content of the responses with the respondents to check the reliability of data collected. The questionnaire had two parts. The first part was made up of questions to identify the respondent worker: sociodemographic variables (age, sex, education level, professional status, etc.), sociability variable and level of satisfaction, career path variables and subjective health and stress variables. The second part included questions to do with early retirement: « Ideally, at what age would you like to retire? »; « What is the main reason for which you would take early retirement? »2

24The average age that a worker wants to leave work is around 58 years of age, with two main peaks: the most answered age is 60 years old for 35% of the people questioned, followed closely by 55 years for 30%. The age of 65, which is the legal age of retirement in Belgium, was only chosen by 8% of the sample. If we dichotomize the variable (whether the desired age is less than or equal to 59 years, or the desired age is greater than or equal to 60), 48% of people want to draw a pension at 60 years or later, and 52% before the age of 60 (n= 748, i.e., 92% of 816 workers).

4. The Desire to Retire Early

25The desire to retire early from the job market is very widespread: 92% of workers queried planned to leave before the legal retirement age of 65. This desire is also unevenly distributed among workers. The objective will therefore be to determine the objective and subjective factors that come into play for wanting to retire early based on individual and organizational data.

26Two hypotheses were then formulated:

  • H1: The desire for early retirement would be related to the symbolic investment of work. Identity at work is then only one of many forms of personal fulfillment and self-realization. This is a hypothesis of identity polycentring.

  • H2: The desire for early retirement from the job market would be due to objective and subjective working conditions: pressure at work and difficulty could lead a worker to retire early from the job market.

  • 3 The following variables were introduced in the analysis: age, sex, marital status, presence of chil (...)

27The following table presents the logistic regression based on different variables.3 This regression must allow to identify variables that come into play in the choice of early retirement before the age of 60, i.e., at least five years before the legal retirement age.

Table 1. Logistic Regression on the Willingness to Retire Early. Before 60 Years of Age


Odds ratios


0.15 (***)


Male (ref.)



3.53 (***)

Number of children(1)

1.37 (*)


Blue-collar worker (ref.)


White-collar worker

0.75 (*)

Lower management

0.66 (***)

Upper management

0.60 (***)

Professional satisfaction(1)

0.55 (***)

Stress level

No stress (ref)


Slight stress

2.16 (***)

Average to a lot of stress

3.46 (***)

Willingness to reduce working time at the end of career




0.29 (***)

Member of a social or cultural association (volunteer)




2.12 (**)

(1) variables considered continuous or ordinal, according to Appendix 1.
(*)= 0.10, (**)= 0.05, (***)= 0.001

28The higher the odds ratio, the more the worker will want to leave work before the age of 60, i.e., well before the legal retirement age. Different variables are significant for understanding this phenomenon: age, sex, having raised children, professional status, professional satisfaction, stress levels, willingness to reduce time at work at the end of career, and belonging to a social or cultural association.

29The older the worker is, the less he wants to leave work before 60. This relationship must be qualified, because it could be the result of population selection: only workers remaining on the job market answered this survey, which may bias the results obtained. Indeed, the relationship observed would be less due to age than to a certain category of the older population.

  • 4 This relationship could also be explained by the recent changes in public policy on retirement in B (...)

30Regarding sex, women were far more likely to want to retire from professional activities before the age of 60. There are several explanations for this. The first result falls under our first hypothesis: it could involve investing less in professional activities, but it is also a form of identity polycentring where professional activities are not the only source of self-actualization: in this case, women would prefer leaving the working world early to devote their time to other forms of symbolic investment within the family structure, in a volunteer activity or for themselves. This hypothesis would also be reinforced by two explanatory elements: the importance of wanting to volunteer time to an association (this investment would also contribute to choosing early retirement) and the family structure. The fact of having raised children also seems to have an impact on the desire to leave the professional world early.4

  • 5 We notice that subjective health does not directly come into play in the choice of early retirement (...)

31The socio-professional status is also relevant for understanding the decision to retire early: the higher status a worker has, the more likely he will leave work after 60. This relationship could be correlated more to professional investment, professional satisfaction and subjective health levels. We can also see a significant relationship between the professional level and professional satisfaction: managers have greater professional satisfaction than workers. This same relationship is also seen with the subjective health level.5

  • 6 In our sample, stress is attributed in over 90% of cases to professional factors, and not solely fa (...)

32Stress levels6 are also a major factor in understanding the desire to retire before the age of 65: the more stressed an individual feels, the more likely he will want to leave work before the age of 60. Stress levels are the highest among young managers nowadays. This result may initially appear paradoxical, but it shows in particular the importance of stress as part of the choice for leaving: even manager-level workers, who would have tended to reject early retirement, considered this solution because of excessive stress levels. This factor thus divides the group of managers based on two logics: unstressed managers who carry out their professional activities in a climate of relative well-being (good subjective health, involvement and high professional satisfaction, etc), and those worn out by stress who are thinking more of ending their career early (poor subjective health, reduced professional satisfaction, etc.). This indicates that there are interactions effects between status and stress.

33Stress also is felt differently by men and women: 19.6% of men say they are stressed or very stressed versus 27% of women. Moreover, men attribute stress only to professional conditions, whereas women qualify it more: it is more a stress coming out of professional factors related to a difficult family organization. In this case, we are dealing with issues of reconciling private and professional life at a so-called age when people are not only taking care of their own children, but may also be looking after their parents.

Table 2. Main Reasons for Stress per Sex (results in%)

 % of Stressed Workers



Absolute %

Relative %

Absolute %

Relative %

For professional reasons





For private reasons





For professional and personal reasons





Not stressed at all










5. Reasons for Early Retirement

34The desire to leave professional life is also unevenly distributed in the work population with certain categories of workers planning more than others to retire early. There are many complex reasons stated for early retirement: they can be due to factors related to health, professional factors related to working conditions in the broadest sense of the word, or private reasons that elude purely professional logic. The table below allows to better understand the main reasons for early retirement.

Table 3. Types of Primary Reasons Stated for Potential Early Retirement (%)

Types of Reasons




Health reasons




Professional reasons

Heavy physical work




Psychosocial load




Work schedule




TOTAL professional reasons




Private reasons








35The two main reasons why people take early retirement from the job market are not directly related to working conditions, but rather to deteriorated health and personal reasons.

36In fact, nearly 40% of workers surveyed (42% women, 36% men) want to leave the working world before legal retirement age for reasons that have little to do with the world of work. This observation reinforces the hypothesis of the polycentring of the individual: these workers invest in other areas, and the subjective importance of this investment pushes them to want to retire early. The gradual dissemination of a cultural model in which professional activity no longer is the main reason for personal investment can probably be illustrated here: other collectively adhered to central standards and values ingrain themselves in people’s minds, inciting them to take early retirement in what some are calling « a culture of early retirement » (Guillemard, 2003: 47).

  • 7 This light increase is probably a manifestation of a healthy worker effect: workers in bad health l (...)

37But, in addition to this aspiration for other horizons, there is a second group of reasons, almost nearly as important as the first: health reasons would lead almost 37% of workers to think about early retirement. Perceived health also tends to significantly deteriorate starting at the age of 45. Although nearly 90% of workers surveyed say they are in good health before 45, this drops to 70% between 45 and 54, and to 75% after age 55.7 In the same way, fewer than 25% of workers under the age of 45 feel their health is not as good as it was five years before, but 40% over 45 do think this. However, the current ability to work does not seem to be influenced by age: it is estimated at 90% of its best level in all age categories. Although subjective health does not seem to be a significant factor for understanding the choice to retire early (see previous analyses), deterioration of health could not only justify voluntarily leaving work, but also make retirement necessary. In this way, it would then no longer be a choice.

38These statistics can help understand the issue involved in ending work. Indeed, subjective health condition tends to reduce with age. Indeed, subjective health condition tends to deteriorate with age, except for those over 55. This phenomenon, at first sight paradoxical, comes through in the effect it has on population selection: only those over 55 years of age who work answered our questionnaire; the others who are probably in worse health are no longer part of our sample. However, this deterioration of subjective health condition does not seem to affect abilities at work. It is as though the health condition, even deteriorated manages to offset the investment needed to carry out a job. Early retirement would not then affect a worker with poor health, i.e., the worker with deteriorated health who can no longer face the pressures, strains and stresses of professional life.

39Only 23% of the reasons given for early retirement come directly from professional reasons, more than 50% of which are related to a major psychosocial load. This weighing of psychosocial factors, especially in men, reiterates the importance of stress, which was identified in previous analyses. However, only 23% of workers appear to have directly mentioned the importance of professional factors in their reasons for taking early retirement; we can therefore only relate these strictly professional reasons with those specifying health conditions. In fact, although deteriorating health is explicitly mentioned, nothing allows us to assert that it is not partially related to working conditions, and therefore would be indirectly related to the working world. We can also formulate the hypothesis that retiring for health reasons would be more socially legitimate than for strictly professional reasons.

  • 8 The same variables as before were introduced into the model. Note that age is no longer an explicat (...)

40The following table lets us even better identify the logic behind early retirement. It has been prepared on the basis of logistic regression using three independently constructed models: each reason given thus corresponds to a model. High odds ratios mean that the worker who wants to retire early will tend to use this as the main reason.8

Table 4. Logistic Regression on Reasons for Retirement

Model Variables

Model 1 : Health Reasons

Model 2 : Prof. Reasons

Model 3 : Personal Reasons

Socio-demographic variables







0,52 (***)

1,31 (**)

Education level

Higher education or university (ref)




Secondary school

2,03 (**)


Primary school

1,58 (***)

0,52 (*)

Professional variables

Professional status

Blue-collar worker (ref)




White-collar worker

0,41 (***)

1,67 (*)

1,51 (**)

Lower management

0,48 (***)

2,65 (***)

0,91 (n.s.)

Upper management

0,59 (**)

1,85 (**)

1,21 (n.s.)

Sector of activity

Secondary sector (ref)




Tertiary sector

0,64 (**)

Financial satisfaction (income)

Very poor to average




Good to very good

1,4 (**)

Work time

Part-time work




Full-time work

1,58 (**)

Wish to reduce working time at the end of career






0,53 (***)

1,60 (**)

Health variables

Subjective health

Poor or acceptable (ref)





0,61 (***)

Very good or excellent

0,71 (**)

Subjective stress

No stress



Slight stress


3,49 (***)


Average to a lot of stress

5,05 (***)

0,62 (*)

Sociability variables

Primarily sociability (family)

Never or rarely (ref)





0,34 (***)

3,60 (**)



0,41 (**)

1,84 (*)

2,21 (*)

Secondary sociability (friends)

Never or rarely (ref)





1,54 (*)


1,84 (**)

Tertiary sociability (social or cultural activities)

Never (ref)





0,43 (***)

2,82 (***)

1,68 (*)

- than 2 hours/week

0,54 (**)

2,17 (**)

1,80 (**)

+ than 2 hours/week

0,61 (*)

1,51 (*)

1,89 (**)

(*)= 0,10, (**)= 0,05, (***)= 0,001

41Different variables can be used to explain the reasons for early retirement. These different variables are not always significant for all reasons. They have been grouped into four homogenous categories in order to provide a better description of the results: socio-demographic variables, professional variables, health variables and sociability variables.

42There are only two significant sociodemographic variables (sex and education level), although other variables were tested: number of children, presence of a spouse, type of housing, place of residence, etc. Although sex is not significant in Model 1 (health reasons), it affects models 2 and 3 differently. Women, more so than men, specify personal reasons for retiring, whereas men are more likely to talk about professional reasons.

43Education level does not seem to come into play in Model 2, which is dedicated to professional reasons, whereas it is relevant in models 1 and 3. Workers with a graduate degree are less likely to specify health as the reason for retiring. The higher the education level, the less likely personal reasons are behind the choice to retire. We should say that this is the main reason, but not necessarily the only reason.

44Professional variables also come into the choice of early retirement. Professional status is thus an important variable. More than for other statuses, blue-collar workers mentioned health reasons, which can partly be explained by lower education levels. Conversely, someone with a worker status was more likely to state reasons for retirement as being mainly professional. Model 3 seems not to be very significantly influenced by this status.

45Sector of activity only influenced Model 1: workers in the secondary sector are more likely than workers in tertiary sectors to leave their professional activities for health reasons.

  • 9 For ethical reasons, the “income” variable could not be included in the questionnaire. We therefore (...)

46Financial satisfaction9 only influences Model 2 since workers with good financial conditions are more likely to use it as a reason to leave than workers without good financial resources.

  • 10 This variable is interesting, because it allows to measure the relationship between the two tempora (...)

47Then there are two other variables related to work time: one is measured in the effective reduced work time, and the other as desired reduced work time. Although part time work does not seem to significantly influence models 1 and 3, it has an effect on retiring for professional reasons: the risk increases for full-time workers. It also seemed important to test the impact of reduced work time that is not effective, but desired. This variable is significant for models 1 and 2. In the first model, the willingness to reduce work time at the end of one’s career allows to reduce the risk of retiring for health reasons: the more the worker wants to end his part-time career, the less likely he will leave for health reasons. In Model 2, this relationship is the opposite: workers wanting to reduce their professional activities are more likely to leave for professional reasons than those who continue working full time.10 In such cases, plans to reduce working time at the end of a career does not does not mean that one will not retire for professional reasons. However, workers who do work part time are less likely to want to leave early than those working full time (previous result). There is thus a gap between what is projected and professional reality.

48Subjective health appeared as a significant variable in Model 1. This result seems logical enough: workers with health problems want to retire from work for health reasons! The stress variable appears more interesting, since, although it is not significant in Model 1, it is strong in Model 2: stress seems to be the most important predictor for retiring for professional reasons. We should also keep in mind that subjective stress is above all professional stress, strongly related to working conditions.

49The last group of variables hinges on the concept of sociability. It is used to indirectly measure the social relationships of the worker outside his professional sphere, i.e., a form of symbolic investment outside the work collective. Thus, primary sociability refers to family relationships, and, more specifically, to the frequency of relationships formed with the close family. While a person with strong primary sociability is less likely to take early retirement for health reasons, the reverse is true in models 2 and 3. The stronger primary sociability is, the greater the chances that an employee will retire for professional or personal reasons.

50As well, the greater secondary sociability is, the greater the risk of retiring for personal reasons. This relationship is also true for tertiary sociability: the more the worker is involved in social and cultural associations, the more likely he is to leave work before legal retirement for professional or personal reasons.

6. Discussion

51The previous analyses identified three profiles, all relatively different, but consistent with the hypotheses that had been formulated.

52The first profile refers to workers who want to take early retirement for health reasons. More specifically, we are talking about workers with lower education levels who are blue-collar workers. Their health is not good and may have even deteriorated over these past five years. They have little contact with their family and rarely are involved in social and cultural activities. This profile involves retirement and social isolation with the weight of work and poor health weighing on individual destinies. These are worn-out, tired, but only slightly stressed workers.

53However, it is interesting to note that looking at reducing work time at the end of career would contribute to significantly delaying their permanent retirement from professional activities.

54The second profile concerns workers who want to quit the professional world for professional reasons. These are hyperactive men with high primary sociability who are involved in many cultural pastimes. They are highly qualified and often hold managerial positions. Early retirement is deeply linked with working conditions, more specifically referring to psychosocial loads: too much stress is the prime element.

55In this profile, we can also establish a link between the idea of reducing work time, and permanent early retirement from the job market. The more the worker looks at reducing working time, the more likely he is to want to retire early from work. It is as though a professional withdrawal process has been activated, and nothing could truly go against it. This is genuine professional saturation: leaving faster, partially or completely. However, it is interesting to note that part-time workers are more likely to continue working longer than those working full time. Perhaps then moving from full-time to part-time work could indeed encourage a worker to remain in his job even if he is not actually considering it.

56The third profile fits more the relatively educated female. These workers develop many social contacts both through the family and with a large circle of friends. They tend to invest a lot in association and cultural pastimes. Their symbolic investment is not centred on professional activities. This is where the hypothesis of polycentring comes clear. Stress, although not highly present, affects both the professional and personal life. The fundamental issue for these workers is to be able to reconcile private and professional life, but the balance could at any time shift to professional disinvestment. Moreover, there is a significant relationship between high sociability and part-time professional integration, which has probably been chosen and not forced.

57While this third profile confirms the hypothesis of polycentring, the first two profiles lend understanding to early retirement from the perspective of necessity and not desire. Whether directly or indirectly, they show how much being worn out at work can lead to various ways of breaking away from professional activities.

7. Conclusion

58The issue of early retirement is complex and multidimensional, because it involves an individual in his relationship to the world, at the heart of the normative provisions that underlie the social basis. This article has attempted to unravel the many aspects involved in choosing to leave work before the legal retirement age and the reasons for this choice. This choice can ultimately take on meaning from on a two-pronged interpretative key: that of desire and that of necessity.

59The issue of desire refers so to speak to the creation of normative models that take on meaning at a given moment in society, but also in an individual and identity path. The public policies developed since the eighties in France, like in Belgium, have deeply contributed to changing the social representation of career ends and thus the associated normative models, thereby creating alongside the established model other equally legitimate forms of retiring. However, this spreading by itself does not allow to understand this incredible eagerness to retire before the legal age. Other factors intervene, especially the deployment of another normative imperative: self-realization and personal growth. This trend also helps in understanding the phenomenon: the individual no longer or does not express himself through focusing on a single area of identity, but builds an identity based on many areas of symbolic investment. This polycentring allows in particular to qualify his professional investment regarding other forms of self-development. In this sense, managing the end of a career becomes a good time for taking stock of one’s identity, used by the social actor in other forms of investment, either within the family, or in other areas, such as volunteer work.

60The issue of necessity refers to the absence of choice, to having an obligation and therefore to renouncing, and has to be put into perspective with professional imperatives and their evolutions. Therefore, although the aging worker to date could benefit from adapted working conditions (Pailhé, 2005), the intensification of work and recent changes in the form of management condemns him in much the same way as younger workers: when age no longer provides protection… It is not so much the heavy physical work that is the problem, but rather the increased psychosocial load. What our survey brought to light was that stress ultimately leads to major health problems, which no longer spares the senior. In both the choice to retire early and the reasons for making that choice, the importance of health factors and especially stress is undeniable. The problem could also grow as we see how many young workers today are experiencing increasingly unsatisfying working conditions, which are jeopardizing their health. And the young people of today are the seniors of tomorrow…

61In either case of whether the choice is desired or borne out of necessity, the collective management of end of work life cannot be resolved only by a transformation, even a radical one, of public policies. According to a Belgium policy manager, it would be enough to « get older workers where it counts—by their pocketbooks » and change the provisions for early retirement to raise employment rates. The worker is clearly not taken into account in his reasoning, and all the more so since companies no longer see it in their interest to keep seniors and do not integrate the issue of aging at work in a bona fide provisional human resources management policy. It is also difficult to measure the real impact of the restrictive policies on the end of working life. Indeed, often the issue arises about bias in selecting populations for surveys on experiences regarding the end of working life: how to understand the impact of a population that has already left the job market on working conditions? Does it not constitute a population with its own characteristics? The Netherlands, which has adopted very restrictive policies in this area over the past few years, has seen a major increase in other aspects of social security, for example, disability insurance.

62Does the future of the end of working life lie more in implementing public policies that try to combine the financial constraints of the aging population and respect for the aspirations of workers? Should we not be innovative by implementing gradual retirement processes that would perhaps take into account professional (dis)engagement, individual aspiration, forms of transferring knowledge and collective constraints? The challenge would then be to move from public policies based on early retirement to progressive retirement policies, and not to policies based on maintaining activities. This article points to some avenues of thought on this topic, in particular regarding the reduction of work time as working life winds down.

Haut de page


Aaronson B.S. (1966). Personality Stereotypes of Aging. Journal of Gerontology, vol. 21, p. 458-462.

Barnay T., Jeger F. (2006). Quels dispositifs de cessation d’activité pour les personnes en mauvaise santé ? Drees, Études et Résultats, 492, 8 p.

Bartel A.P., Sicherman N. (1993). Technological Change and Retirement Decisions of Older Workers. Journal of Labour Economics, n° 11, vol. 1, p. 162-183.

Bird C.P., Fisher T.D. (1986). Thirty Years later. Toward the Employment of Older Workers. Journal of Applied Psychology, vol. 71, n° 3, p. 515-517.

Blanchet D., Debrand T. (2005). Aspiration à la retraite, santé et satisfaction au travail : une comparaison européenne. Insee Première, n° 1052, 4 p.

Crespo S., Beausoleil J. (1999). La sortie anticipée d’activité des travailleurs de 45 à 64 ans. Fiche technique, Montréal, INRS Culture et Société.

Drury E. (1993). Discrimination fondée sur l’âge exercée contre les travailleurs âgés de la communauté européenne. Une analyse comparative. Bruxelles, Eurolink Age.

Elchardus M., Cohen J. (2003). Attitude et attentes en rapport avec la fin de carrière professionnelle. Rapport partiel n° 3 : les déterminants de la fin de carrière, Louvain.

Feldman D.C. (1994). The Decision to Retire Early: a review and Conceptualization. vol. 19, n° 2, p. 285-311.

Friedman E.A., Orbach H.L. (1974). Adjustment to retirement. In American Handbook of Psychiatry, eds S. Arieti, vol. 1, New York: Basic Books.

Gollac M., Volkoff S. (2000). Les conditions de travail. Paris, La Découverte, Repères.

Guillemard A.M. (2003). L’âge de l’emploi : les sociétés à l’épreuve du vieillissement. Paris, Armand Colin.

Gratton, B., Haug M.R. (1983). Decision and Adaptation. Research on Aging, vol. 5, p. 59-76.

Griffiths A. (1997). Ageing, Health and Productivity: a challenge for the new millennium. Work and Stress, vol. 11(3), p. 197-214.

Hansez I., Bertrand F., De Keyser V., Pérée F. (2005). Fin de carrière des enseignants : vers une explication du stress et des retraites prématurées. Le Travail Humain, vol. 68, p. 193-223.

Hansson R., Dekoekkoek P.D., Neece W., Patterson D. (1997). Successful aging at work. Annual Review, 1992-1996: the older workers and transitions to retirement, Journal of Vocational Behavior, vol. 51, n° 2, p. 202-233.

Hardy M.A., Hazelrigg L. (1999). A multilevel Model of early retirement decisions among autoworkers in plants with different futures. Research on aging, vol. 21, p. 275-303.

Hassell B., Perrewe P. (1995). An examination of Beliefs about Older Workers: do stereotypes still exist? Journal of organizational Behaviour, vol. 16(5), p. 457-468.

Kilbom A. (1999). Evidence-based programs for the prevention of early exit from work. Experimental Aging Research, vol. 25 (4), p. 291-299.

Kirchner W., Lindbom T., Paterson D. (1952). Attitudes toward the Employment of Older People. Journal of Applied Psychology, vol. 36(3), p. 154-156.

Kirchner W., Dunnette M. (1954). Attitudes toward Older Workers. Personnel Psychology, n° 7, p. 257-265.

Leleu M. (2001). Stéréotypes et travailleurs âgés. Le cas de la Belgique et des pays limitrophes : Luxembourg, Pays-Bas. Genève, Rapport pour le bureau international du travail, ILO.

Lund T., Borg V. (1999). Work environment and Self-rated health as predictors of remaining in work five years among Danish employees 35-39 years of age. Experimental Aging Research, vol. 25 (4), p. 429-434.

Lyon P., Pollard D. (1997), Perceptions of the Older Employee: is anything Changing? Personnel Review, vol. 26(4), p. 245-252.

Mc Goldrick A., Cooper C. (1980). Voluntary early Retirement - taking the decision. Employment Gazette, p. 859-864.

Mc Gregor, J. (2002). Stereotypes and older Workers: the New-Zealand Experience. Social Policy Journal of New Zealand, vol. 18, p. 163-177.

Molinié A.-F. (2006). Les salariés quinquagénaires, entre fragilisation et protection. Retraite et Société, n° 49, p. 12-37.

Neboit M., Vézina M. (2007). Stress au travail et santé psychique. Toulouse, Octares Éditions.

OCDE, 2003, Vieillissement et politiques de l’emploi : Belgique. Paris : Author.

Pailhé A. (2005). Les conditions de travail : quelle protection pour les salariés âgés en France ? Population, vol. 1-2, p. 99-126.

Paloniemi S., Tikkanen T. (2000). Victims of Ageist Attitudes - But how do the Older Workers themselves view their competences. Paper presented in ESREA research Network meeting Adult Education and the labour Market VI, October 2000, Seville.

Parker S. (1980). Older Workers and Retirement. London, HMSO.

Remery C., Henkens K., Schippers J., Ekamper P. (2003), Managing an aging Workforce and a tight labor market: views held by Dutch Employers. Populations Research and Policy Review, vol. 22, p. 21-40.

Rosen B., Jerdee T.H. (1976). The Nature of Job-related age Stereotypes. Journal of Applied Psychology, vol. 61(4), p. 428-432.

Rosen B., Jerdee T.H. (1977), Too Old or not too Old ? Harvard Business Review, vol. 11-12.

Saurel-Cubizolles M.-J., Bardot F., Berneron B. et collab., (1999), État de santé perçu et perte d’emploi, Travail, santé et Vieillissement. Relations et évolutions. Toulouse, Éditions Octarès, p. 53-66.

Schmitt N., White J.K., Coyle B.W., Rauschenberger J. (2004). Retirement and Life Satisfaction. The Academy of Management Journal, vol. 22, n° 2, p. 282-291.

Schultz K.S., Morton K.R., Weckerle J.R. (1998). The influence of push and pull factors on voluntary and involuntary early retirees’ retirement decision and adjustment. Journal of Vocational Behavior, n° 53, p. 45-57.

Simoens D. (1980). Hoe pijnloos is de crisis ? Kritische analyse van de ontwikkeling van de wetgeving over de sociale zekerheidsprestaties (1976-1979). Crisiswetgeving en social Zekerheid, Anvers, Kluwer, p. 70-82.

Szinovacz M.E., De Viney S. (2000). Marital Characteristics and retirement decisions. Research on aging, n° 22, p. 470-498.

Taylor P., Walker A. (1993). Employers and Older Workers. Employment Gazette, vol. 101(8), p. 371-378.

Taylor P., Walker A. (1994). The Ageing Workforce : Employers’ Attitudes towards Older People. Work, Employment and Society, vol. 8(4), p. 569-591.

Taylor P., Walker A. (1998). Employers and Older Workers : attitudes and Employment practices. Ageing and Society, vol. 18, p. 641-658.

Tuckman, J., Lorge I. (1952). Attitudes toward older Workers. Journal of Applied Psychology, vol. 36(3), p. 149-153.

Van der Velden J. (1994). Age-conscious social policy. In Work and Aging : a European prospective, eds J. Snel and R. Cremer, London, Taylor and Francis, p. 353-364.

Van Solinge H., Henkens K. (2005). Couples’ Adjustment to Retirement : a multi-actor Panel Study. Journal of Gerontology : social sciences, 60B, S11-S20.

Walker A. (1985). Early Retirement : Release or Refuge from the labour Market. The Quarterly Journal of Social Affairs, vol. 1 (3), p. 211-229.

Haut de page


Characteristics of the Population Studied (n =826)




Age Group

< 34 years



35-44 years



45-54 years



> 55 years











Live together



Do not live together



Number of children

No children



1 child



2 children



3 or more children



Social status

Blue-collar workers



White-collar workers



Lower management



Upper management



Education level




Trade school









Financial situation










Very good



Subjective health







Very good



Subjective stress




Not stressed



Professional satisfaction

Not satisfied






Very satisfied



Sectors of activity




Public administration



Chemical industry



Financial sector & insurance



Research & development






Social sector



Metallurgical industry (cars)



Machine/tool manufacturing









Wholesale business



Haut de page


1 This article is based on the results of the CAPA study, jointly funded by Le Fonds Social Européen and Service Public Fédéral, Direction Humanisation du travail.

2 The models for early retirement from the job market that attempted to identify individual components distinguish three successive steps in the process that leads to the end of working life (Feldman, 1994, cited by Gaillard & Desmette, 2007)
The first step refers to the preference of the individual for early retirement (I would like to stop my professional activities); the second step involves the intention of leaving (I plan to leave); the third step represents effectively making the decision (Gaillard & Desmette, 2007: 325). In this case, the model for understanding the factors influencing retirement was designed based on data collected from moving from the action, and not “post hoc” (Barnes-Farrell, 2003). This methodological position can be justified through the strong concordance between the expression of the intention and the actual act of retiring (Henkens & Tazelaar, 1994; Prothero & Beach, 1984, cited by Gaillard & Desmette, 2007).
This article is based on this methodological position, which fosters the collection of data well before the actual act takes place.

3 The following variables were introduced in the analysis: age, sex, marital status, presence of children or grandchildren, education level, professional status, work time, sector of activity, financial satisfaction, professional satisfaction, stress level, subjective health, primary sociability (contact with family), secondary sociability (contact with friends), tertiary sociability (member of a social or cultural association). They have been integrated into a previously developed theoretical framework at both the individual and organizational levels.

4 This relationship could also be explained by the recent changes in public policy on retirement in Belgium: the legal age for retirement of women was increased to 65 from the previous 60 years of age. Certain women would have had difficulty assimilating this change fully.

5 We notice that subjective health does not directly come into play in the choice of early retirement. It only intervenes indirectly via professional satisfaction or via stress.

6 In our sample, stress is attributed in over 90% of cases to professional factors, and not solely family factors.

7 This light increase is probably a manifestation of a healthy worker effect: workers in bad health left the labor market prematurely in response to an array of social policies implemented by the Belgian government over the last twenty five years.

8 The same variables as before were introduced into the model. Note that age is no longer an explicative variable in the reasons for early retirement.

9 For ethical reasons, the “income” variable could not be included in the questionnaire. We therefore do not have information on satisfaction related to income to understand financial effects.

10 This variable is interesting, because it allows to measure the relationship between the two temporal projection variables: one is based on a desire to reduce professional activities at the end of one’s working life and the other to prematurely end professional activities. In both cases, it involves projecting into a relatively distant future.

Haut de page

Pour citer cet article

Référence électronique

Nathalie Burnay, « Voluntary early retirement: Between desires and necessities »Perspectives interdisciplinaires sur le travail et la santé [En ligne], 10-2 | 2008, mis en ligne le 01 novembre 2008, consulté le 13 juillet 2020. URL :; DOI:

Haut de page


Nathalie Burnay

Maître de conférence FUNDP, Département de sciences politiques, sociales et de la communication, Rempart de la Vierge 8, 5000 Namur, Belgique

Haut de page
  • Logo FQRSC – Fonds de recherche Société et culture Québec
  • Logo IRSST – Institut de recherche Robert-Sauvé en santé et en sécurité du travail
  • Logo SELF – Société d’ergonomie de langue française
  • Logo DOAJ - Directory of Open Access Journals
  • OpenEdition Journals