ABSTRACT
In view of the steady growth in life expectancy in recent decades the question is increasingly being raised whether and how older people should be encouraged to be more active, and particularly to engage in unpaid voluntary work. Taking adult life as a whole the conditions for such charitable involvement would appear to be especially favourable after retirement. However, these analyses, which are based on German longitudinal data, show that the effect of entering retirement is often exaggerated. Rather, the individual's previous volunteering experiences are of major importance in his decision to take up and continue voluntary work in later life. At the same time the analyses show how important the major resources of health and education are, particularly for participation in voluntary work during ageing.
1. Introduction
The structural demographic change in modern societies can be observed particularly in Europe. Low birth rates and growing life expectancy will clearly increase not only the absolute number of older people in these societies but also their share in the population as a whole (Eurostat 1999; Kohler et al. 2002). Even if it proves possible to change permanently the policy of ‘early retirement’ and to raise the retirement threshold for most of the workforce (OECD 2002), it must be expected that people will have many more years of retirement than in the past since life expectancy should remain constant or even rise further (Cutler and Meara 2001).
The shift in the relation between the number of employed people on the one hand and the number of pensioners who have left working life on the other is often described by the general public as a growing ‘age burden’, to which social policy needs to find answers. However, to call this an ‘age burden’ is to overlook the many productive contributions which older people make to the prosperity of society as a whole. While that view is only gradually gaining ground in Europe, in the United States, for example, attention was drawn to it more than 20 years ago with the use of the term ‘productive ageing’ (Butler 1985; Morris and Bass 1988) and in the general debate there has been more frequent reference to the voluntary work performed by seniors recently (Bass and Caro 2001; Fast et al. 2006).
Voluntary work differs from employment by a number of criteria. One of the most obvious is that volunteers receive no monetary compensation for their input (wages, fees, etc.). But one can go further and draw a distinction between formal and informal voluntary work. Formal voluntary work is performed for an organisation outside the volunteer's own household (like an association), while informal voluntary work is organised by the volunteers themselves. It is on principle performed for persons outside the volunteer's household, like neighbours, more distant relations and friends. So it must be distinguished from self-help, which is work for the more intimate circle of members of the household (Tilly and Tilly 1994; Wilson and Musick 1997a).
In the course of a person's life the performance of voluntary work increases over time, reaching a peak between the ages of about 35 to 55, and then declining again (Goss 1999; see also Wilson 2000: 226). However, particularly the shares of older people active in voluntary work have increased in recent years and decades (Goss 1999). But even if the share of older people performing voluntary work is already making a considerable contribution to the prosperity of society as a whole it certainly seems meaningful to encourage and stimulate this further. It is suggested that this involvement will counter the possible danger of social isolation for the volunteers and at the same time maintain their physical and mental well-being, if not, indeed, improve these (Van Willigen 2000; Thoits and Hewitt 2001; Morrow-Howell et al. 2003; Siegrist et al. 2004). For society as a whole the productive contribution from seniors would reduce the ‘age burden’, as they provide goods and services at favourable cost. So it is hoped that encouraging more seniors to engage in these activities will reduce social costs and at the same time increase general prosperity. It is hardly surprising, then, that politicians are becoming increasingly aware of the possibilities which the productive (informal) potential of older people seems to offer, out of working life as well (e.g., Chambre 1989; Bass et al. 1995; Baldock 1999).
Common to many of these political efforts is that they are aimed at people who are just before retirement or have just retired, and although this is little thematised in practice, such a strategy is often based primarily on implicit theoretical assumptions on the use of time. The belief is that pensioners who have left working life have lots of free time. At the same time today's pensioners have a high mental and physical performance capacity compared with earlier generations. So it seems appropriate, in the interests of both the seniors themselves and the welfare of society as a whole, to make better use of this seemingly unutilised resource by encouraging pensioners to engage more in voluntary work.
However, this theoretical relation between the resources of free time and participation in voluntary work that currently dominates political practice has so far hardly been subject to empirical analysis. This paper therefore aims to help close that gap in research, by not only examining the influence of retirement on engagement in formal and informal voluntary work, but also thematising the role of previous activities. We ask how experience of voluntary work further back in the past affects engagement in such work in retirement.
As a start, Section 2 surveys the current state of research on the influence of retirement and previous experience on engagement in voluntary work. Section 3 then gives an account of the data base and the method used in the study. Section 4 initially traces the development of the extent of engagement by older people in formal and informal voluntary work in Germany since the mid-1980s. This is followed by multivariant longitudinal analyses, which yield more exact information on the importance of retirement and previous experience of voluntary work. The paper closes with some conclusions (Section 5).
2. The current state of research
Beside the influence of individual resources, like health and education (Smith 1994; Wilson 2000), and of institutional context factors (Salomon and Anheier 1998; Erlinghagen and Hank 2006) the importance of previous experience for engagement in voluntary work has increasingly also been thematised in recent years. Distinctions must be drawn here between the short-term effect of individual events like marriage, divorce, death of the partner or the birth of a child (Wilson and Musick 1997a, 1998; Rotolo 2000; Okun and Michel 2006; and on voluntary work by couples Rotolo and Wilson 2006) and the long-term effect of past experience, like socialisation in the family (Janoski and Wilson 1995; Mustillo et al. 2004) or the cultural identity of whole cohorts (Goss 1999; Putnam 2000; Jennings and Stoker 2004; Rotolo and Wilson 2004).
The transition to retirement is a major, single event for the individual, comparable with marriage or the birth of a child. Reference is repeatedly made, particularly in the discussion on productive ageing, to the importance of the transition from working life to retirement in determining the participation of older people in voluntary work (e.g., Moen and Fields 2002; Smith 2004). For most people the transition to retirement marks a major change in their lives. Since the 1960s at the latest there has been debate over how far leaving working life necessitates a search for new interests and concerns, a new way to utilise time and new social contacts, and at the same time makes this possible (Burgess 1960; Cumming and Henry 1961).
As the individual approaches the end of working life voluntary work can be a promising way to organise the foreseeable retirement and give purpose to this new phase of his life (e.g., Caro and Bass 1997). Voluntary work can be a real option, particularly if the individual has already been involved in this in the past (e.g., Mutchler et al. 2003). So the question arises of if and how the single event of retiring (the ‘retirement effect’) on the one side and the more long-term effect of past experience (the ‘experience effect’) has an impact on the individual propensity to volunteer in later life.
As far as I know the study by Mutchler et al. (2003) is the only longitudinal empirical study to date that is explicitly concerned with the relevance of the ‘experience effect’ and the ‘retirement effect’ outlined above in explaining the voluntary activities of older people. The authors show firstly that the stability of participation in voluntary activities is not affected by the change in employment status. A major conclusion is that formal voluntary work and the provision of informal help are very strongly characterised by continuity over the years (see also Oesterle et al. 2004). On the other hand, their results also show that after leaving working life older people are more likely to take up new formal voluntary work compared with people in full employment. However, it is an open question whether this is in fact a retirement effect, as part-time workers also show more propensity to take up these activities between the two years surveyed.1 There is no clear retirement effect on informal help, by contrast. Altogether Mutchler et al. (2003) show that the involvement of older people in voluntary work is primarily a result of the ‘experience effect’; their past experience is of major importance. But there are also signs that there is a certain ‘retirement effect’ for persons who had not until then been engaged in formal voluntary work, as they are more likely to take this up on leaving working life.
This picture is substantiated by the cross-section analyses available on this subject. These analyses also suggest that the ‘retirement effect’ is rather of minor importance. Persons not in employment and pensioners are not more involved in voluntary work than persons in employment. Certainly older persons who are not in employment spend more time on voluntary work than persons in employment (Gallagher 1994; Choi 2003), but too much significance should not be attached to this, for ‘only a small percentage of the time available as a result of termination of work is reallocated to volunteering’ (Caro and Bass 1997: 432). Gauthier and Smeeding (2003) come to a similar conclusion, and they also show that with progressive ageing it is not involvement in voluntary work but the passive activities, like watching television, and consumption (e.g., hobbies) that become more important.
The lack of empirical longitudinal studies on the influence of past experience and entering retirement on older people's involvement in voluntary work is particularly astonishing. It is therefore the aim of this paper to help close that research gap. Using German panel data we can examine whether and to what extent the voluntary engagement of pensioners can be explained by experience effects and/or retirement effects. One advantage of this study over that by Mutchler et al. (2003) is that the German data clearly reflects final transitions from working life into retirement. That is chiefly because nearly all pensioners in Germany no longer engage in paid employment or do not take up any paid work (Brussig et al. 2006). So the status of ‘pensioner’ can be equated with finally leaving working life, at least in Germany. Secondly, the German data used here also contains information on past experience with informal voluntary work, and this data goes back further than that used in the US study.
3. Data and methods
3.1. The Socio-Economic Panel (SOEP)
These analyses are based on data of the Socio-Economic Panel (SOEP) (Wagner et al. 2007). SOEP is a representative repeat survey of households in Germany that has been carried out every year since 1984. As well as data on the households, which is obtained by questioning the head of the household, SOEP also provides information on the individual members of the household. Persons aged over 16 are questioned. This enables analyses to be made on both household and individual level. The panel nature of the survey enables both representative cross-section and longitudinal analyses to be made.
In the first year of its operation, 1984, SOEP contained information on around 12,000 persons questioned from just under 6,000 households. In 1990 the survey was extended to cover the territory of the former GDR, so that about 6,000 persons from 2,200 households were added. To counter problems of panel mortality and at the same time improve the scope for analysis further, a total of five supplementary random samples were taken in the next few years. The latest wave used here, 2005, covers a good 21,000 persons in 11,500 households (on the size of the random sample and panel mortality in SOEP see Kroh and Spiess 2006).
3.2. The analysis strategy
The paper is interested in the individual dynamic of volunteering by older people,2 with particular focus on the influence of retirement and past experience in voluntary work on the assumption or ending of such activities between 2001 and 2005. In a first step, the paper presents some descriptive results on the dynamic of volunteering. These descriptive results are shown according to different groups of older people to give a first impression of possible socio-economic determinants of the individual dynamic of voluntary work. After that time-discreet transition models in the form of binary logistical regressions are estimated (Hosmer and Lemshow 2000; Cameron and Trivedi 2005: 602f). Separate estimates are made of the likelihood of persons starting or ending formal and informal voluntary work. As a person can only start volunteering if they were inactive in 2001, and correspondingly can only give up volunteering if they were active in 2001, not all persons in the data set are at risk of experiencing transition to (in)activity. So persons who were active in 2001 are excluded from the estimates of the number starting voluntary work and persons who were inactive in 2001 are not included in the estimates of the number giving up voluntary work.
3.3. Variables and data restrictions
3.3.1. Dependent variables
The participants in SOEP are also asked every one to two years about their involvement in formal voluntary work, and about their informal help to friends or relations. The question is:
Which of the following activities do you take part in during your free time? Please check off how often you do each activity: at least once a week, at least once a month, less often, never.
Volunteer work in clubs or social services;
Involvement in a citizens’ group, political party, local government;
Helping out friends, relatives or neighbors.
On principle a distinction must be drawn between starting an activity and increasing it, and between giving one up and reducing the time spent on it. As the SOEP data used here does not give sufficiently precise information on the time intensity of voluntary work, this study is limited exclusively to the question of how many started a new activity and how many gave these up entirely. Accordingly, the different dependent variables are modelled as follows:
- •
An active engagement has been started if a person inactive in 2001 reports engagement in voluntary work in 2005 (dependent variable=‘1’, otherwise ‘0’).
- •
An active engagement has ceased if a person active in 2001 is no longer engaged in voluntary work in 2005 (dependent variable=‘1’, otherwise ‘0’).5
3.3.2. Explanatory variables
In order to analyse the influence of past experience with voluntary work on the assumption or ending of such activities between 2001 and 2005 appropriate information from SOEP from 1996 to 1999 was used.6 If persons surveyed reported at least once in these years that they performed formal voluntary work or regularly gave informal help we assess this here as ‘Experience with formal or informal voluntary work’.
Leaving working life appears in different ways in the estimates. First the employment status in 2001 (employed, unemployed, pensioner, not earning) is included as a time-constant variable. Later the transition to retirement is included in the estimates as a variable that changes over time. A transition to retirement is made if the person questioned in 2001 had not then retired but gave ‘pensioner’ as their employment status in 2005.
The other control variables are all constant over time (as per 2001)7 and they form the influences that earlier studies have shown to greatly affect the involvement of older people in voluntary work. As well as gender they include equipment with human capital (measured by the highest professional or vocational qualification obtained), state of health (measured by subjective health status) and partnership status (measured by whether a person lives with a partner in the same household). Age is included in the estimates as a metric variable, with the square of age also included so that a possible, non-linear relation can be shown adequately. The region of residence (East or West Germany) is also checked, as engagement in voluntary work still differs between the old and the new federal states, even after reunification, owing to the differences in institutional framework conditions, traditions and socialisation experience (cf. e.g., Künemund 2006).
3.3.3. Data restrictions
The data set used for the analysis here consists exclusively of persons who were aged at least 50 in 2001 and who have taken part in every survey from 1996 to 1999 and in 2001 and 2005. Moreover, persons are excluded from the data set if no valid figures for the explanatory variables (formal or informal voluntary work) are available for them in either 2001 or 2005. Under these conditions 3,291 individuals remained in the data set. For the analyses that include entry into retirement as a feature varying over time only those persons are included who in addition to the above criteria had not entered retirement in 2001 nor reached age 65. Under these conditions 1,616 remained in the data set.8
4. Results
4.1. The individual dynamic of voluntary work
Table 1 shows, for formal and informal voluntary work separately, the shares of persons in the data set who (a) were inactive in both 2001 and 2005 (‘stable inactive, SI’), (b) were inactive in 2001 and active in 2005 (‘started voluntary work, SV’), (c) were active in 2001 and inactive in 2005 (‘ended voluntary work, EV’), and finally (d) were active in both 2001 and 2005 (‘stable active, SA’). Twenty percent of all the older respondents reported a constant formal engagement, and around 19 percent reported constant informal help between 2001 and 2005. There is, however, greater dynamic in engagement in informal help, as noticeably larger shares of the older generation reported starting or ending regular help of this kind.
. | Formal voluntary work . | Informal voluntary work . | . | . | ||||||
---|---|---|---|---|---|---|---|---|---|---|
. | SI . | SV . | EV . | SA . | SI . | SV . | EV . | SA . | % . | n . |
Total | 63.5 | 7.4 | 9.0 | 20.0 | 55.3 | 12.1 | 13.5 | 19.1 | 100.0 | 3,291 |
Male | 57.9 | 7.2 | 9.1 | 25.8 | 55.1 | 12.4 | 14.1 | 18.5 | 47.2 | 1,553 |
Female | 67.8 | 7.6 | 9.0 | 15.6 | 55.4 | 12.0 | 13.1 | 19.5 | 52.8 | 1738 |
Aged 50–64 | 59.3 | 8.4 | 9.2 | 23.0 | 52.7 | 14.5 | 12.7 | 20.2 | 63.2 | 2,080 |
65–74 | 63.3 | 7.8 | 7.5 | 21.5 | 49.9 | 11.5 | 15.6 | 23.0 | 26.0 | 856 |
≥75 years | 80.2 | 2.8 | 11.3 | 5.7 | 75.8 | 4.4 | 12.9 | 6.9 | 10.8 | 355 |
Poor health | 65.6 | 7.2 | 9.1 | 18.2 | 58.1 | 11.7 | 13.1 | 17.2 | 71.9 | 2,367 |
(Very) good health | 57.6 | 8.2 | 8.9 | 25.2 | 47.3 | 13.3 | 14.9 | 24.5 | 28.1 | 924 |
No vocational training | 71.5 | 6.3 | 6.9 | 15.3 | 56.3 | 13.9 | 12.8 | 17.0 | 27.4 | 901 |
With vocational training | 63.3 | 7.6 | 9.2 | 19.9 | 54.4 | 11.2 | 15.5 | 18.9 | 53.4 | 1,756 |
College/University | 45.6 | 10.6 | 13.7 | 30.1 | 55.6 | 11.5 | 9.5 | 23.5 | 18.0 | 591 |
In employment | 55.7 | 8.9 | 10.4 | 25.0 | 52.4 | 12.7 | 12.7 | 22.2 | 38.5 | 1,267 |
Unemployed | 37.8 | 21.2 | 7.9 | 33.0 | 55.8 | 22.1 | 4.3 | 17.7 | 3.7 | 122 |
Pensioner | 70.6 | 5.8 | 8.1 | 15.5 | 58.2 | 11.0 | 14.4 | 16.4 | 50.9 | 1,674 |
Otherwise not earning | 57.6 | 6.3 | 10.5 | 25.6 | 45.1 | 13.5 | 15.5 | 26.0 | 6.9 | 228 |
Lives alone | 69.4 | 7.8 | 8.2 | 14.7 | 58.1 | 10.3 | 14.9 | 16.7 | 20.1 | 660 |
With partner | 60.7 | 7.3 | 9.8 | 22.3 | 53.5 | 12.8 | 13.1 | 20.5 | 76.1 | 2,504 |
Other households | 60.3 | 6.8 | 2.5 | 30.5 | 61.5 | 17.3 | 7.0 | 14.2 | 3.9 | 127 |
West Germany | 62.0 | 7.7 | 9.4 | 20.9 | 55.7 | 11.9 | 12.8 | 19.7 | 68.9 | 2,268 |
East Germany | 69.7 | 6.3 | 7.4 | 16.6 | 53.6 | 13.2 | 16.6 | 16.5 | 31.1 | 1,023 |
No formal volunteering experience | 88.1 | 5.6 | 4.2 | 2.1 | – | – | – | – | 54.3 | 1,786 |
With formal volunteering experience | 35.8 | 9.5 | 14.4 | 40.3 | – | – | – | – | 45.7 | 1,505 |
No informal volunteering experience | – | – | – | – | 77.2 | 9.4 | 8.2 | 5.3 | 39.8 | 1,309 |
With informal volunteering experience | – | – | – | – | 39.2 | 14.1 | 17.5 | 29.2 | 60.2 | 1,982 |
No transition to retirement | 55.6 | 9.5 | 10.6 | 24.4 | 53.3 | 12.3 | 11.8 | 22.6 | 34.2 | 1,125 |
Transition to retirement | 52.3 | 9.4 | 9.5 | 28.8 | 47.6 | 16.3 | 13.9 | 22.2 | 14.9 | 491 |
. | Formal voluntary work . | Informal voluntary work . | . | . | ||||||
---|---|---|---|---|---|---|---|---|---|---|
. | SI . | SV . | EV . | SA . | SI . | SV . | EV . | SA . | % . | n . |
Total | 63.5 | 7.4 | 9.0 | 20.0 | 55.3 | 12.1 | 13.5 | 19.1 | 100.0 | 3,291 |
Male | 57.9 | 7.2 | 9.1 | 25.8 | 55.1 | 12.4 | 14.1 | 18.5 | 47.2 | 1,553 |
Female | 67.8 | 7.6 | 9.0 | 15.6 | 55.4 | 12.0 | 13.1 | 19.5 | 52.8 | 1738 |
Aged 50–64 | 59.3 | 8.4 | 9.2 | 23.0 | 52.7 | 14.5 | 12.7 | 20.2 | 63.2 | 2,080 |
65–74 | 63.3 | 7.8 | 7.5 | 21.5 | 49.9 | 11.5 | 15.6 | 23.0 | 26.0 | 856 |
≥75 years | 80.2 | 2.8 | 11.3 | 5.7 | 75.8 | 4.4 | 12.9 | 6.9 | 10.8 | 355 |
Poor health | 65.6 | 7.2 | 9.1 | 18.2 | 58.1 | 11.7 | 13.1 | 17.2 | 71.9 | 2,367 |
(Very) good health | 57.6 | 8.2 | 8.9 | 25.2 | 47.3 | 13.3 | 14.9 | 24.5 | 28.1 | 924 |
No vocational training | 71.5 | 6.3 | 6.9 | 15.3 | 56.3 | 13.9 | 12.8 | 17.0 | 27.4 | 901 |
With vocational training | 63.3 | 7.6 | 9.2 | 19.9 | 54.4 | 11.2 | 15.5 | 18.9 | 53.4 | 1,756 |
College/University | 45.6 | 10.6 | 13.7 | 30.1 | 55.6 | 11.5 | 9.5 | 23.5 | 18.0 | 591 |
In employment | 55.7 | 8.9 | 10.4 | 25.0 | 52.4 | 12.7 | 12.7 | 22.2 | 38.5 | 1,267 |
Unemployed | 37.8 | 21.2 | 7.9 | 33.0 | 55.8 | 22.1 | 4.3 | 17.7 | 3.7 | 122 |
Pensioner | 70.6 | 5.8 | 8.1 | 15.5 | 58.2 | 11.0 | 14.4 | 16.4 | 50.9 | 1,674 |
Otherwise not earning | 57.6 | 6.3 | 10.5 | 25.6 | 45.1 | 13.5 | 15.5 | 26.0 | 6.9 | 228 |
Lives alone | 69.4 | 7.8 | 8.2 | 14.7 | 58.1 | 10.3 | 14.9 | 16.7 | 20.1 | 660 |
With partner | 60.7 | 7.3 | 9.8 | 22.3 | 53.5 | 12.8 | 13.1 | 20.5 | 76.1 | 2,504 |
Other households | 60.3 | 6.8 | 2.5 | 30.5 | 61.5 | 17.3 | 7.0 | 14.2 | 3.9 | 127 |
West Germany | 62.0 | 7.7 | 9.4 | 20.9 | 55.7 | 11.9 | 12.8 | 19.7 | 68.9 | 2,268 |
East Germany | 69.7 | 6.3 | 7.4 | 16.6 | 53.6 | 13.2 | 16.6 | 16.5 | 31.1 | 1,023 |
No formal volunteering experience | 88.1 | 5.6 | 4.2 | 2.1 | – | – | – | – | 54.3 | 1,786 |
With formal volunteering experience | 35.8 | 9.5 | 14.4 | 40.3 | – | – | – | – | 45.7 | 1,505 |
No informal volunteering experience | – | – | – | – | 77.2 | 9.4 | 8.2 | 5.3 | 39.8 | 1,309 |
With informal volunteering experience | – | – | – | – | 39.2 | 14.1 | 17.5 | 29.2 | 60.2 | 1,982 |
No transition to retirement | 55.6 | 9.5 | 10.6 | 24.4 | 53.3 | 12.3 | 11.8 | 22.6 | 34.2 | 1,125 |
Transition to retirement | 52.3 | 9.4 | 9.5 | 28.8 | 47.6 | 16.3 | 13.9 | 22.2 | 14.9 | 491 |
SI, stable inactive; SV, started voluntary work; EV, ended voluntary work; SA, stable active.
Source: SOEP (author's calculations), shares weighted longitudinally.
In accordance with the explanatory variables used in the later multivariate analysis Table 1 presents the descriptive results separately for the different features. As this paper is particularly concerned with the effect of transition to retirement and of past experience, only those results are discussed below.
As Table 1 shows, pensioners are not only much less frequently active in formal voluntary work than persons in employment, for example, or the unemployed, they also show a very low engagement dynamic compared with other groups. Pensioners are also less frequently active in informal help, although here the dynamic of starting or ending voluntary work is comparable with that of persons in employment. If entry into retirement is modelled as a variable changing over time we see that 9.4 percent of the older generation who retired between 2001 and 2005 took up formal voluntary work during the same period. Persons who had not yet retired in 2005 showed an almost identical figure for this at 9.5 percent. The picture is rather different for informal voluntary work, where 16.3 percent of the new pensioners also started giving regular help parallel to this, while the corresponding figure for persons without transition to retirement is lower, at 12.3 percent.
Past experience with voluntary work seems to play a much greater part than the employment status, both in the engagement and the dynamic of activity. Nearly 90 percent of the older generation without experience in formal voluntary work, and a good 75 percent of this generation without experience in informal help, remain permanently inactive. But if the older persons had experience with voluntary work in the past five years the comparable inactivity shares are clearly lower, at just under 36 and 39 percent, respectively. The shares of the older generation that took up voluntary work between 2001 and 2005 are also clearly bigger in the group with experience. The same does apply to the share of persons who had ended their engagement by 2005, but this higher rate needs to be relativised in view of the greater participation on principle of persons with experience.
The descriptive results shown in Table 1 thus primarily indicate the existence of experience effects, and they suggest that the transition to retirement plays only a minor role in participation in voluntary work. Insofar these descriptive results largely accord with the results of earlier studies. Finally, however, only the multivariate analysis, which now follows, will tell us how important the transition to retirement on the one side and experience gained on the other are in taking up and ending voluntary work.
4.2. The retirement effect versus the experience effect
Table 2 shows the coefficients of the binary logistical regression estimate of starting and ending formal voluntary work. For each, two models are estimated, with experience in formal voluntary work in the past five years included as an additional explanatory variable in both the second estimates.
. | Starting . | Ending . | ||
---|---|---|---|---|
Age | 0.175 | 0.180 | −0.428*** | −0.475*** |
Age2 | −0.001 | −0.001 | 0.003*** | 0.004*** |
(Very) good health | 0.259* | 0.311** | −0.517*** | −0.558*** |
No vocational training | −0.603*** | −0.559*** | 0.183 | 0.156 |
College/University | 0.485*** | 0.374** | −0.176 | −0.133 |
Unemployed | 0.186 | 0.053 | −0.555 | −0.487 |
Pensioner | −0.323 | −0.320 | −0.019 | 0.096 |
Otherwise not earning | −0.031 | −0.164 | 0.018 | 0.050 |
With partner | 0.198 | 0.116 | 0.156 | 0.206 |
Other households | −0.116 | −0.268 | 0.005 | 0.199 |
Female | 0.001 | 0.057 | 0.408*** | 0.287* |
East Germany | −0.239 | −0.191 | 0.140 | 0.086 |
Formal volunteering experience | 1.444*** | −1.627*** | ||
Constant | −7.30** | −0.07** | 12.05*** | 15.13*** |
Pseudo-R2 | 0.037 | 0.104 | 0.035 | 0.082 |
N | 2,372 | 2,372 | 919 | 919 |
. | Starting . | Ending . | ||
---|---|---|---|---|
Age | 0.175 | 0.180 | −0.428*** | −0.475*** |
Age2 | −0.001 | −0.001 | 0.003*** | 0.004*** |
(Very) good health | 0.259* | 0.311** | −0.517*** | −0.558*** |
No vocational training | −0.603*** | −0.559*** | 0.183 | 0.156 |
College/University | 0.485*** | 0.374** | −0.176 | −0.133 |
Unemployed | 0.186 | 0.053 | −0.555 | −0.487 |
Pensioner | −0.323 | −0.320 | −0.019 | 0.096 |
Otherwise not earning | −0.031 | −0.164 | 0.018 | 0.050 |
With partner | 0.198 | 0.116 | 0.156 | 0.206 |
Other households | −0.116 | −0.268 | 0.005 | 0.199 |
Female | 0.001 | 0.057 | 0.408*** | 0.287* |
East Germany | −0.239 | −0.191 | 0.140 | 0.086 |
Formal volunteering experience | 1.444*** | −1.627*** | ||
Constant | −7.30** | −0.07** | 12.05*** | 15.13*** |
Pseudo-R2 | 0.037 | 0.104 | 0.035 | 0.082 |
N | 2,372 | 2,372 | 919 | 919 |
Significance: ***P<0.01; **P<0.05; *P≤ 0.1.
Reference categories: satisfactory/poor health, with vocational training, in employment, lives alone, male, West Germany (no formal volunteering experience).
Source: SOEP (author's calculations).
Before we turn to the question of the effect of retirement and past experience let us first briefly look at the other control variables. Evidently qualifications and the state of health play a decisive role in whether a person starts formal voluntary work when aged 50 or over. Well qualified and healthy seniors are significantly more often inclined to become active in this field. Good health also increases the continuity of the engagement by this age group. At the same time a humped age progression is evident. The continuity first increases after 50. However, the significant positive relation to the square of the age in this estimate indicates that, as was only to be expected, formal voluntary work is ended more frequently as ageing progresses (turning point approximately 59 years).
Regarding the significance of the retirement and experience effects the results evidently point to the latter. Pensioners do not differ from persons in employment in their propensity to take up or end formal voluntary work between 2001 and 2005. But past experience with voluntary work does appear to be a very strong explanatory factor in the models estimated. Older persons with the appropriate experience not only show a highly significant greater propensity to start voluntary work, they are also distinguished by a highly significant stability in this activity over time.
If participation in regular informal help is now considered (Table 3) we see here, too, that good health not only increases the take up of voluntary work, it also encourages continuity. At the same time older people with an academic degree show greater stability in giving regular informal help than less highly qualified respondents. As age increases the effects are similar to those for formal engagement, which means on the one hand an initially increasing propensity of starting informal volunteer activities with increasing age that decreases in later life (turning point approximately 69 years). On the other hand we find an initially decreasing propensity of ending informal volunteering that increases year by year (turning point approximately 51 years).
. | Starting . | Ending . | ||
---|---|---|---|---|
Age | 0.442*** | 0.412*** | −0.334*** | −0.305** |
Age2 | −0.004*** | −0.003*** | 0.003*** | 0.003*** |
(Very) good health | 0.303** | 0.303** | −0.353*** | −0.338** |
No vocational training | 0.126 | 0.106 | −0.067 | −0.020 |
College/University | 0.055 | 0.096 | −0.358** | −0.383** |
Unemployed | −0.223 | −0.148 | –0.026 | −0.015 |
Pensioner | −0.200 | −0.227 | 0.383** | 0.356* |
Otherwise not earning | −0.128 | −0.155 | 0.057 | 0.110 |
With partner | 0.133 | 0.120 | 0.099 | 0.070 |
Other households | 0.556* | 0.648** | −0.242 | −0.280 |
Female | −0.017 | −0.055 | −0.108 | –0.138 |
East Germany | −0.103 | −0.140 | 0.228* | 0.263* |
Informal volunteering experience | 0.936*** | −0.932*** | ||
Constant | −14.59*** | −14.05*** | 9.47** | 9,30** |
Pseudo-R2 | 0.031 | 0.064 | 0.036 | 0.055 |
N | 2123 | 2123 | 1168 | 1168 |
. | Starting . | Ending . | ||
---|---|---|---|---|
Age | 0.442*** | 0.412*** | −0.334*** | −0.305** |
Age2 | −0.004*** | −0.003*** | 0.003*** | 0.003*** |
(Very) good health | 0.303** | 0.303** | −0.353*** | −0.338** |
No vocational training | 0.126 | 0.106 | −0.067 | −0.020 |
College/University | 0.055 | 0.096 | −0.358** | −0.383** |
Unemployed | −0.223 | −0.148 | –0.026 | −0.015 |
Pensioner | −0.200 | −0.227 | 0.383** | 0.356* |
Otherwise not earning | −0.128 | −0.155 | 0.057 | 0.110 |
With partner | 0.133 | 0.120 | 0.099 | 0.070 |
Other households | 0.556* | 0.648** | −0.242 | −0.280 |
Female | −0.017 | −0.055 | −0.108 | –0.138 |
East Germany | −0.103 | −0.140 | 0.228* | 0.263* |
Informal volunteering experience | 0.936*** | −0.932*** | ||
Constant | −14.59*** | −14.05*** | 9.47** | 9,30** |
Pseudo-R2 | 0.031 | 0.064 | 0.036 | 0.055 |
N | 2123 | 2123 | 1168 | 1168 |
Significance: ***P<0.01; **P<0.05; *P≤ 0.1.
Reference categories: satisfactory/poor health, with vocational training, in employment, lives alone, male, West Germany (no informal volunteering experience).
Source: SOEP (author's calculations).
In regard to the question of the retirement and experience effects the results, first for the determinants of starting activity, are similar to those for formal engagement. Pensioners are not more likely to start this work, but help given in the past made it much more likely that they would (again) become active in informal voluntary work between 2001 and 2005. This past experience also significantly reduced the likelihood of ending their participation in informal help during the period examined. But the most surprising result of these estimates is that pensioners show a greater likelihood – of at least some significance – to end regular help over time.
Table 4 finally shows the estimated results of the regressions, in which entry into retirement was included as a variable changing over time. As already mentioned, only those respondents that had not yet retired in 2001 and were also not yet over 65 were included in these estimates. Unlike Tables 2 and 3 these results are presented in the form of Odds Ratios, in order to provide information not only on the direction of the influence but also on the intensity of the effect of the various factors.
. | Formal voluntary work . | Informal voluntary work . | ||||||
---|---|---|---|---|---|---|---|---|
. | Starting . | Ending . | Starting . | Ending . | ||||
Age | 0.429 | 0.415 | 0.344 | 0.350 | 0.757 | 0.753 | 0.907 | 0.776 |
Age2 | 1.007 | 1.008 | 1.009 | 1.009 | 1.003 | 1.003 | 1.001 | 1.002 |
(Very) good health | 1.242 | 1.368 | 0.692* | 0.679* | 1.066 | 1.135 | 0.836 | 0.855 |
No vocational training | 0.515*** | 0.556** | 1.064 | 1.053 | 1.067 | 1.033 | 0.958 | 1.23 |
College/University | 1.338 | 1.188 | 0.723 | 0.734 | 0.989 | 1.019 | 0.592** | 0.575** |
With partner | 1.158 | 0.912 | 1.012 | 1.012 | 1.055 | 1.061 | 0.807 | 0.793 |
Other households | 0.583 | 0.403 | 0.825 | 0.939 | 1.784 | 1.835 | 0.696 | 0.630 |
Female | 1.077 | 1.093 | 1.483* | 1.388 | 0.980 | 0.918 | 0.809 | 0.766 |
East Germany | 0.676* | 0.760 | 1.227 | 1.165 | 1.215 | 1.211 | 1.144 | 1.188 |
Retired 01/05 | 1.453 | 1.005 | 1.347 | 0.677 | ||||
Interaction effects | ||||||||
Not retired 01/05 & with experience | 6.536*** | 0.227*** | 2.841*** | 0.457*** | ||||
Retired 01/05 with no experience | 2.100** | 1.130 | 1.350 | 0.826 | ||||
Retired 01/05 with experience | 6.547*** | 0.205*** | 3.557*** | 0.276*** | ||||
Pseudo-R2 | 0.026 | 0.120 | 0.026 | 0.072 | 0.010 | 0.050 | 0.014 | 0.032 |
N | 1,028 | 1,028 | 525 | 525 | 995 | 995 | 568 | 568 |
. | Formal voluntary work . | Informal voluntary work . | ||||||
---|---|---|---|---|---|---|---|---|
. | Starting . | Ending . | Starting . | Ending . | ||||
Age | 0.429 | 0.415 | 0.344 | 0.350 | 0.757 | 0.753 | 0.907 | 0.776 |
Age2 | 1.007 | 1.008 | 1.009 | 1.009 | 1.003 | 1.003 | 1.001 | 1.002 |
(Very) good health | 1.242 | 1.368 | 0.692* | 0.679* | 1.066 | 1.135 | 0.836 | 0.855 |
No vocational training | 0.515*** | 0.556** | 1.064 | 1.053 | 1.067 | 1.033 | 0.958 | 1.23 |
College/University | 1.338 | 1.188 | 0.723 | 0.734 | 0.989 | 1.019 | 0.592** | 0.575** |
With partner | 1.158 | 0.912 | 1.012 | 1.012 | 1.055 | 1.061 | 0.807 | 0.793 |
Other households | 0.583 | 0.403 | 0.825 | 0.939 | 1.784 | 1.835 | 0.696 | 0.630 |
Female | 1.077 | 1.093 | 1.483* | 1.388 | 0.980 | 0.918 | 0.809 | 0.766 |
East Germany | 0.676* | 0.760 | 1.227 | 1.165 | 1.215 | 1.211 | 1.144 | 1.188 |
Retired 01/05 | 1.453 | 1.005 | 1.347 | 0.677 | ||||
Interaction effects | ||||||||
Not retired 01/05 & with experience | 6.536*** | 0.227*** | 2.841*** | 0.457*** | ||||
Retired 01/05 with no experience | 2.100** | 1.130 | 1.350 | 0.826 | ||||
Retired 01/05 with experience | 6.547*** | 0.205*** | 3.557*** | 0.276*** | ||||
Pseudo-R2 | 0.026 | 0.120 | 0.026 | 0.072 | 0.010 | 0.050 | 0.014 | 0.032 |
N | 1,028 | 1,028 | 525 | 525 | 995 | 995 | 568 | 568 |
Significance: ***P<0.01; **P<0.05; *P≤0.1.
Reference categories: satisfactory/poor health, with vocational training, in employment, lives alone, male, West Germany, did not retire between 2001 and 2005 (and no volunteering experience).
Source: SOEP (author's calculations).
In the first step models are estimated that include entry into retirement beside the control variables already known, but without including information on past voluntary experience. It is clear that the transition to retirement does not affect the likelihood of either starting or ending formal or informal voluntary work. In the second step interaction effects between the transition to retirement and active experience are taken into account. Seniors who did not report entry into retirement between 2001 and 2005, and who had no experience with voluntary work, here function as the reference category. Persons are distinguished from them who (a) did not retire between 2001 and 2005 but had past experience, (b) who did retire between 2001 and 2005 but did not have past experience, and (c) who retired between 2001 and 2005 and had past experience.
The results show firstly that there certainly are retirement effects. The likelihood of becoming active in formal voluntary work between 2001 and 2005 is twice as high for seniors without past experience (Odds Ratios 2.100) if they retired during that period. However, experience effects play a much greater part, for independent of entry into retirement seniors who were active in the preceding five years are more than six times more likely to start a formal voluntary engagement (Odds Ratios 6.536 and 6.547) and three times more likely to start regular informal help (Odds Ratios 2.841 and 3.557). The continuity of experienced seniors is also two to four times greater than that of the inexperienced reference group (Odds Ratios 0.227 and 0.205 or 0.457 and 0.276). Altogether, therefore, experience effects predominate, while retirement effects are of rather minor significance. The analyses of the German panel data thus essentially confirm the results from the US (Mutchler et al. 2003).
5. Conclusions
The one-sided perception of the ageing of society as a growing ‘age burden’ is short-sighted. It is well known that a high percentage of people over 50 are productive by performing voluntary work today, and that this is increasing prosperity in society as a whole. The share of (older) volunteers is especially high in the Scandinavian countries and The Netherlands. Based on European and US survey data, Erlinghagen (2008: 250) reports 36.9 percent of the older population (50+) in Norway, 33.2 percent in The Netherlands and 33.0 percent in Sweden doing some formal volunteer work in 2003. The respective activity rate in Germany is lower (24.9 percent) and shows quite similar values to the UK (26.2 percent) and the US (24.8 percent) (see also Erlinghagen and Hank 2006 for similar findings based on other survey data). However, such international comparisons do also show that there are significant differences in volunteer activities of the older population (mean activity rate: 17.1 percent). Since, for example, more than every third Norwegian senior works as a volunteer, the activity rate in Poland (5.2 percent), Spain (4.9 percent) and Italy (4.5 percent) is much lower. Hank and Stuck (2009) have provided evidence that these differences in Europe are not only caused by social structural differences but also by differences in the institutional context. They show a positive correlation between, for example, social expenditure and the share of volunteer workers. In line with other results (see for example Chambre 1989; Chaves et al. 2004; Motel-Klingebiel et al. 2005) there seems to be no danger that the welfare state crowds out volunteering but that welfare state activities could encourage volunteer activities by providing an adequate infrastructure, for example, in terms of public gyms or professional supervising staff.
But even if the participation of older people in this work is relatively high, and has even increased in recent years (cf. Hank and Erlinghagen 2010), it may still appear meaningful to activate productive potential that is still unutilised because voluntary work can improve the well-being of the participants themselves while performing important and valuable work for society as a whole.
So far this activation has been seen primarily as a question of time use. In that sense the transition to retirement would appear to be an ideal point in time to persuade an increasing number of older people to engage in voluntary work, as they may be assumed to have a lot of free time, and after leaving working life many are looking for a new direction. However, these analyses have clearly shown that the role of retirement is often exaggerated. Previous experience of voluntary work is, by contrast, of major importance in encouraging seniors to engage and continue in these activities. At the same time the analyses have again shown the great importance of health and education as major resources, also and especially for participation in voluntary work during ageing.
Although our analysis provides new insight into the dynamics of volunteering there are some limitations that should be taken into account when interpreting the results and should be addressed by future research. The major weakness of the presented investigations lies in the lack of continuous information about volunteer activities. Because of data restrictions explained in section 3.3 we are only able to measure volunteer dynamics by comparing activities in the years 2001 and 2005. Thus, we know nothing about individual volunteer activities in between. It is definitely possible that individuals who were active in 2001 and 2005 and who are, therefore, regarded as ‘stable active’ could have stopped and then resumed their volunteer work in the meantime. Thus, to further increase our understanding of the life course dynamics of volunteer work (more) continuous information on individual volunteer histories is needed. It should also be mentioned that we have no information about the intensity of volunteer work in our data set. Thus, it is possible that we define individual volunteer activity as stable over time, although there has been a significant decline in the amount of hours provided by a volunteer worker. Further analyses should, therefore, make some effort to investigate in the impact of retiring and past volunteering activities not only on volunteer activities per se but especially on volunteer hours.
The presented results are also limited because our investigations are restricted to Germany. Although the findings are in line with results for the US we cannot be sure that in other countries the ‘experience effect’ is similarly important to understand older people's participation in volunteer work. As mentioned above there is strong evidence that country-specific welfare state arrangements have an important impact on individuals’ decision to volunteer. In this respect the interplay between welfare state institutions, the life course and the dynamics of volunteering has to be addressed by future research (cf. Hank and Erlinghagen 2010).
If the data base permits, besides the analysis presented here of the commencement and ending of voluntary work, future research should also take account of the extension or reduction of the engagement over time. Moreover, the analysis of the retirement and experience effects should also include the interdependence between the various forms of activity. Evidently an individual's engagement in different areas of voluntary work can be complementary, but it can also be a substitution (Burr et al. 2005; Choi et al. 2007), insofar as it is an open question in how far formal voluntary work in retirement, for example, is influenced by past experience in other areas (like the care of relatives). That also makes obvious connections to questions of the importance of different experience with socialisation that could not be considered in this paper. A distinction must be drawn here between individual and collective socialisation experience. In the first case it must in future be examined whether beside the ‘long arm of the job’ (Wilson and Musick 1997b) there is also the ‘long arm of the family’, to influence the extent and individual dynamic of voluntary work in old age. In the second case the relation between cohort effects (Goss 1999; Putnam 2000; Jennings and Stoker 2004; Rotolo and Wilson 2004) on the one side and retirement and experience effects on the other should be analysed more exactly.
However, although these questions remain open for future research, this paper has improved knowledge of the influence of past experience on older people's participation in voluntary work. That is not only of scientific interest, it also provides important new insight for social policy in practice. Regarding the results presented here it seems very doubtful whether activation programmes that are targeted directly at the group of seniors are likely to be very successful. Paradoxically, if the productive potential of the older generation is to be stimulated and maintained in the medium and long term it would appear to be more appropriate first to win over young adults to engage in voluntary work, because that will greatly increase the chances that they will continue or resume such activities when they are much older. A sustained education and health policy is just as important, as this is essential for people to remain as productive as possible when they are older. That applies equally to participation in working life and to voluntary work.
Footnotes
On possible working time effects Wilson (2000: 221) comments: ‘Rather than hours worked, what might be important is the individual's control over those hours. The self-employed and people with flexible work schedules are the most likely to volunteer’ (see also Wilson and Musick 1997b). The analyses by Freeman (1997) actually show that the involvement and in some cases the amount of time spent do not decrease as working time increases; in some cases they actually rise again.
By ‘older people’ we mean persons aged over 50. This relatively early definition is partly for pragmatic reasons (the lower age threshold in many data sets used in social science research into age is 50 or even lower). At the same time it also takes account of the fact that ‘ageing’ is not a state but a process, the development of which needs to be traced. We also use the term ‘seniors’ as a synonym for ‘older people’, independent of whether they would regard themselves as such.
Attending cultural or sports events, active in sports, fine arts and other artistic activities, social activities and church attendance.
In the year 2005 1,153 (12.4 percent) of the respondents aged 50 years or older took part in formal volunteering sporadically (less than once a month) and 1,897 (20.4 percent) were regularly (at least once a month) engaged. In contrast, 4,290 (46.2 percent) performed informal volunteer work sporadically whereas 3,325 (35.8 percent) report regular informal volunteer activities during the preceding 12 months.
The two years 2001 and 2005 were chosen because in 2003 participants in the survey were asked to report formal voluntary work, but not informal help. Moreover, in 2003 the questions were slightly different.
In 2000 the survey did not asked about voluntary work, and no information on informal help is available for 1998. In principle, information from even earlier years could have been used; however, that would clearly have reduced the number of cases that could be analysed, and that is why the retrospective period has been limited to five years here.
Control variables that change over time would hardly help the results here, as sufficiently precise information is not available on the timing of explanatory and dependent events.
It should be quite clear that our data restrictions might result in a biased sample. However, a comparison of the distribution of our control as well as dependent variables on the one hand in our analysis sample and one the other hand in the original SOEP sample for the year 2001 has shown no serious bias. The overall distribution of the variables in our analysis is quite congruent.
References
Marcel Erlinghagen, Dr. in Sociology, works as a research associate at the Institut Arbeit and Qualification (IAQ) of the University of Duisburg-Essen. He is also a research affiliate of the German Institute for Economic Research (DIW), Berlin. Research interests: job security; labour market dynamics; volunteer work. Recent publications include articles in Ageing & Society and European Sociological Review.