| HOME | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|---|
| ||||||||||||||||||||||||||||||||
RESEARCH ARTICLE |
a Department of Gerontology, University of Nebraska at Omaha
b Department of Sociology, University of Kansas, Lawrence
c Department of Social Sciences, University of Maryland Eastern Shore, Princess Anne
Karl Kosloski, Department of Gerontology, University of Nebraska at Omaha, Omaha, NE 68182 E-mail: kkosloski{at}mail.unomaha.edu.
Decision Editor: Margie E. Lachman, PhD
 |
Abstract |
|---|
 TOP Abstract Retirement and Job...Other Factors Related to...Retirement PlanningHypothesesMethodsResultsDiscussionReferences |
|---|
) to evaluate whether certain job-related gratifications might reduce retirement planning. Three definitions of retirement planning were evaluated and then regressed separately on a set of variables that included 3 types of job-related satisfactions (intrinsic gratification, positive social relations, and ascendance in the workplace) and 7 covariates: education, age, sex, health, marital status, race, and pension eligibility. Findings indicated that jobs high in ascendance were related to an increase in certain types of retirement planning, but jobs high in intrinsic rewards and positive social relations were related to less planning, regardless of how planning was defined. The findings suggest that information about work-related rewards may be useful in targeting individuals who might benefit from retirement planning programs, in developing planning programs to help workers realize more complex retirement plans, and in assisting employers who hope to retain older workers. THE approaching retirement of the Baby Boom cohort, changes in eligibility ages for full Social Security benefits, how best to invest any Social Security surplus, and the extent to which individuals should manage their public and private retirement accounts are just a few of the issues that have served to refocus attention on the importance of individual retirement planning. Given the prominence of these issues in the media, one might expect older workers to be developing specific plans for their own retirement.
Recent evidence suggests that retirement anticipation is normative among older workers (
Ekerdt, Kosloski, and DeViney 2000) and that the great majority of American workers can cite an age at which they expect to retire (
National Public Radio 1999). Even so, those retirement ages are increasingly variable (
Han and Moen 1999), and two fifths of workers in their 50s report that they have no specific plan for retirement or have not given it much thought (
Ekerdt, DeViney, and Kosloski 1996). The picture that emerges is one of a cohort of workers aware that they will probably retire but inconsistent in their plans and intentions.
To explain the variability of both plans and behavior toward retirement, researchers have had recourse to demographic, financial, and health characteristics. Less attention has been devoted to the nature of the job as a motive for retirement readiness. One of the earliest and most persisting assumptions about retirement, however, posits that for researchers to understand what retirement means to people, it is necessary to understand what work has meant to them (
Friedmann and Havighurst 1954). Our purpose in the present study was to evaluate the relevance of certain work-related rewards (i.e., intrinsic enjoyment, positive social relations, and opportunities for ascendance in the workplace) for retirement planning, conceived here as a broad set of cognitions and behaviors toward the event of withdrawal from work. We expected to find an association between the extent of work-related rewards and reduced levels of retirement planning.
|
Retirement and Job Characteristics |
|---|
|
TOPAbstractRetirement and Job...Other Factors Related to...Retirement PlanningHypothesesMethodsResultsDiscussionReferences |
|---|
;
Eisdorfer 1972;
Fillenbaum 1971;
Friedmann and Havighurst 1954;
Goudy, Powers, and Keith 1975;
Miller 1965;
Shanas 1972). Generally, past views have been based on the straightforward supposition that work provides individuals with certain outcomes. To the extent that these outcomes are valued, individuals will be less likely to leave them (
Glenn and Weaver 1985;
Hayward and Hardy 1985;
McCune and Schmitt 1981). Conversely, negative attitudes toward the job provide impetus toward retirement (
Beehr 1986;
Ekerdt and DeViney 1993;
Hansson, DeKoekkoek, Neece, and Patterson 1997). It follows, then, that workers who like their jobs will be less likely to make plans for leaving them. Some support for this notion is provided by
Richardson and Kilty 1992 who found that, in a sample of African American professionals, workers who were highly committed to their work tended to avoid planning for or thinking about retirement. Other studies have demonstrated a link between attitudes toward one's job and either retirement intentions or behavior. For example,
Schmitt, Coyle, Rauschenberger, and White 1979, in a comparison of early retirees and nonretirees, found that retirees reported less autonomy, fewer opportunities to deal with others, and less intrinsic satisfaction in their jobs than nonretirees. In a study of factors associated with early retirement,
Hayward and Hardy 1985 found that certain characteristics or requirements of the job, including substantive complexity, high social skills, and low physical demands, moderated the manner by which certain background characteristics of workers (e.g., health, education, income) were related to retirement, presumably by changing the attractiveness of retirement through a shaping of the worker's opportunity structure. Similarly,
Barfield and Morgan 1969 reported that workers who were dissatisfied with their jobs were more receptive to the idea of early retirement. And in a comparison of compulsory and voluntary retirees, Palmore and his colleagues found that workers with pro-job attitudes were more likely to keep working until they were forced to retire (
Palmore, Burchett, Fillenbaum, George, and Wallman 1985). On the other hand, studies that have attempted to relate global work attitudes to global attitudes toward retirement have not documented a strong inverse relationship between the two (
Fillenbaum 1971), even among persons for whom work holds a central, organizing position in their lives (
Goudy et al. 1975). This suggests, perhaps, the need to differentiate the important functions of work and relate them to specific retirement intentions and behaviors.
Consideration of the positive functions of work has a long history in the psychology of work. Although there is presently no definitive model that details all of the outcomes associated with working, a substantial number of factor analytic studies support the view that work provides rewards in addition to the possible intrinsic enjoyment of performing the work activity itself and any economic return (
Gruneberg 1979). For example,
Friedlander 1963 identified two other work-related rewards: (a) a pleasing social and technical environment and (b) recognition through advancement. The latter factor comprised such things as job promotion, merit increases, receiving challenging assignments, and public recognition. Similarly,
Voydanoff 1978 identified a factor that she labeled self-expression. It included items such as "I am given a chance to do the things I do best" and "I have a chance to develop my own special abilities."
Kosloski, Ginsburg, and Backman 1984 identified a factor that they labeled support for self-view. It, too, focused on ascendance in the workplace and included items such as "to have a chance to lead others in activities" and "to have opportunities to demonstrate your particular skills, talents, and abilities."
Other researchers have also identified at least one factor, in addition to intrinsic and economic rewards, that appears to be related either to one's social environment or to opportunities for self-expression and ascendance in the workplace (e.g.,
House, McMichael, Wells, Kaplan, and Landerman 1979;
Mannheim 1975;
Phelan, Bromet, Schwartz, Dew, and Curtis 1993). Taking an anthropological perspective,
Goldschmidt 1995 has shown that the social function of work is important and that the opportunity to demonstrate one's competence in task performance is an important gratification of work across a range of societies. Similarly,
Etzioni 1995 argued that in addition to intrinsic and monetary rewards, many individuals work long and hard at their job because of the prestige and support for self-identity that it provides.
Although the rewards associated with working can undoubtedly be represented in myriad ways, outcomes representing intrinsic enjoyment, positive social relations, and opportunities for ascendance in the workplace are consistent with previous research. The manner in which these rewards might be related to retirement planning can be inferred directly from early models of retirement such as the substitution model (
Friedmann and Havighurst 1954; see also
Shanas 1972), the reinforcement model (
Eisdorfer 1972), and the identity crisis model (
Miller 1965). All assume that work provides a well-established avenue for attaining important rewards. For example, the substitution model contends that work potentially provides at least five valued outcomes in the life of an individual, including an income, regulation of life activities, the means for acquiring an identity, structured patterns of interpersonal interactions, and meaningful life experiences. According to Friedmann and Havighurst, "if these functions and the satisfactions they bring are lost by retirement, then retirement is an undiluted tragedy for a man" (1954, p. 189). Thus, to the extent that retirement implies the loss of valued rewards, it will be viewed negatively by the worker and adjustment to retirement will be more difficult. Extending this logic, if retirement is devalued, planning for retirement should be devalued as well. Thus, a negative relationship between work-related rewards and retirement planning can be hypothesized: As jobs increase in their abilities to provide work-related rewards, the extent of retirement planning should diminish.
|
Other Factors Related to Retirement Plans and Decisions |
|---|
|
TOPAbstractRetirement and Job...Other Factors Related to...Retirement PlanningHypothesesMethodsResultsDiscussionReferences |
|---|
;
Ekerdt et al. 1996;
Palmore et al. 1985;
Taylor and Shore 1995). Gender has also been shown to be important, because women tend to have more complex work histories than men (
DeViney and O'Rand 1988) and are more likely to report having no specific plan for retirement relative to changing jobs, reducing work hours, or stopping altogether. Poor health is related to both increased retirement planning (
Ekerdt et al. 1996;
Taylor and Shore 1995) and earlier exits (
Hayward and Hardy 1985). Marital status is an important covariate because married couples tend to view retirement as a joint venture (
Henkens 1999;
Vinick and Ekerdt 1991), and perhaps as a result, married workers are more likely to express certainty regarding their retirement plans (
Ekerdt et al. 1996). Race is related to a number of factors that may limit the opportunity structure for older workers (
Belgrave 1988;
Gibson 1987). For example, non-Whites are more likely to indicate a plan to stop working altogether or to change jobs in retirement rather than express uncertainty about their future (
Ekerdt et al. 1996). Pension eligibility has been shown repeatedly to exert a profound influence on retirement behavior (e.g.,
Gustman, Mitchell, and Steinmeier 1993;
Quinn, Burkhauser, and Myers 1990), including reducing uncertainty about retirement plans (
Ekerdt et al. 1996). Education also represents an important covariate in the present study. Higher education is likely to be associated with both a greater opportunity for economic rewards from work that increase retirement options (
Campbell and Henretta 1976) and increased opportunities for noneconomic rewards (e.g., intrinsic enjoyment and opportunities for ascendance) that may affect one's desire to remain in the workforce.
The relevance of these demographic and background factors leads to an important caveat: The function and meaning of work and retirement change over time. For example, in the case of work,
Bell 1973 has argued that modern America is completing a value revolution in which the Protestant work ethic is being replaced by a new system of hedonistic values that emphasize leisure and consumption. Under such a scenario, work may become a less efficient vehicle for gratification attainment than for past generations. Even with less dramatic shifts in values, it is clear that the relationship of individuals to their work roles is constantly changing (
Morris and Bass 1988). Similarly, retirement has had multiple functions throughout the last century (
Graebner 1980). As in the case of work, retirement has been shaped by changing employer-employee relationships, by the ability of the economy to support a leisure class at any point in time, and by norms that promote the individual accumulation of wealth necessary to support a retired lifestyle. As a result, the relationships between work, retirement, and the foregoing demographic and background factors are historically situated and thereby limited in their generalizability.
|
Retirement Planning |
|---|
|
TOPAbstractRetirement and Job...Other Factors Related to...Retirement PlanningHypothesesMethodsResultsDiscussionReferences |
|---|
). In the present study, retirement planning was defined in three different ways. At its simplest level, retirement planning can be represented as an intention toward the form retirement will take (e.g., full retirement, partial retirement, or second career;
Ekerdt et al. 1996).
A second approach is to model workers' expectations for future behavior.
Juster 1997 has noted that measures of subjective probabilities have performed well in empirical models predicting future behavior. Using data from the Health and Retirement Study (HRS),
Holden, McBride, and Perozek 1997 recently demonstrated that respondents' expectations concerning future nursing home use were very similar to known probabilities of lifetime use. The HRS also asked workers about the likelihood that they will be working past the ages of 62 and 65. It is anticipated that researchers can use these subjective probabilities to gain similar insights into workers' plans for retirement.
Perhaps the most direct measure of retirement planning is anticipatory rehearsal. That is, to what extent are workers thinking about retirement and discussing it with their friends and coworkers? Such behaviors are rich in content validity and valid on their face. Moreover, such activities serve as an anchor by which to evaluate and add meaning to the other measures of retirement planning (i.e., plans and expectations).
Although each of these outcomes taps a separate domain of retirement planning, they are likely to be related. That is, if workers do not have a plan for the form of retirement, they are less likely to state with any degree of certainty that they will not be working at ages 62 or 65 and less likely to be thinking about retirement or discussing it with friends. In contrast, workers who express a high degree of certainty that they will be fully retired at ages 62 or 65 will be more likely to be thinking about retirement and discussing it with friends.
|
Hypotheses |
|---|
|
TOPAbstractRetirement and Job...Other Factors Related to...Retirement PlanningHypothesesMethodsResultsDiscussionReferences |
|---|
|
Methods |
|---|
|
TOPAbstractRetirement and Job...Other Factors Related to...Retirement PlanningHypothesesMethodsResultsDiscussionReferences |
|---|
) that includes an oversampling of Florida residents. The sample for the present study included all persons aged 51–61 from the core sample, excluding the Black and Hispanic oversamples. From these 8,017 cases, the 2,945 persons who stated that they were not current workers or were completely retired were dropped. Because questions about employers and coworkers were asked only of those individuals who stated that they were working for someone else, self-employed individuals also were excluded, leaving 4,237 persons for the final analyses. In addition, the two items we used to assess the extent to which workers (a) thought about retirement or (b) discussed it with friends and coworkers were not asked of individuals who indicated on an earlier question that they never planned to retire. Consequently, in analyses involving the dependent variable formed from these two items, the sample was further reduced by 429 cases. We attempted to maximize the sample size for each analysis, depending on the available data. Effective sample sizes ranged from 3,725 (when all variables were considered simultaneously) to 4,545 (for confirmatory factor analysis of the dependent variables). The proportion of missing data on individual items in the survey not due to skip patterns in the questionnaire was extremely small (2%) and deleted listwise.
The HRS employs a household design that includes both primary respondents and their spouses. Of the 4,237 non-self-employed respondents, 1,335 (31.5%) were spouses of primary respondents. Treating observations that are clustered within family dyads as though they were independent has the potential to bias results by underestimating the total variation between individuals. The potential for this distortion is typically expressed as the design effect: 1 + (c - 1)p where c represents the average cluster size and p represents the intraclass correlation. Monte Carlo studies suggest that design effects less than 2.0 do not result in overrejections of models at the conventional (.05) alpha level in normal theory methods (Table 2 in
Muthen and Satorra 1995). Design effects for all of the outcome variables in the present study are comfortably below this limit, because of the small cluster size.
|
Measurement of Variables
Work satisfaction.
One drawback to use of archived data is that, because the data were collected for other purposes, direct measures of constructs of interest are not always available. To some extent, that was the case in the present study. Nonetheless, a number of questions were asked in the HRS that appeared to have at least some relevance to the constructs of intrinsic rewards, positive social relations, or opportunities for ascendance. With respect to intrinsic gratifications, two items seemed relevant. First, respondents were asked the extent to which they agreed with the following statement: "Even if I didn't need the money, I would probably keep on working." Responses were coded on a scale from 1 (strongly disagree) to 4 (strongly agree). Later in the interview, respondents were also asked, "Do you think of work as important mainly because of the money, or of money as less important than the work itself?" Responses were coded on a scale ranging from 1 (work mainly for the money) to 3 (work itself is most important). People endorsing the pros and cons of both were scored 2. The distributional characteristics of responses to all of the questions used in the present study are shown in Table 1 .
|
We used three items to represent the factor that has been variously described as representing self-expression (
Voydanoff 1978), recognition through advancement (
Friedlander 1963), support for self-view (
Kosloski et al. 1984), or competence (
Goldschmidt 1995) and labeled here as ascendance in the workplace: (a) "My job requires skill in dealing with other people," (b) "My job requires me to keep up with the pace set by others," and (c) "My job requires that I learn new things." Each of the items appears to tap the challenge of maintaining one's competence in the organization relative to others (e.g., "skill" in dealing with others, keeping up the "pace" set by others, and learning "new things" to stay abreast of developments). Each of these items was scored on a four-point scale from 1 (all or almost all of the time) to 4 (none or almost none of the time) and then reverse coded so that high scores reflected greater opportunity for ascendance and self-enhancement in the workplace.
One obvious reinforcer from working is money, and its absence as a work-related reward in the present study warrants explanation. Two items were available that addressed employees' perceptions of the economic aspects of their jobs: (a) "My pay is fair considering what other people in this line of work are paid," and (b) "How much pay I receive depends almost entirely on how well I do my job." The correlation between these two items was extremely low (.21), suggesting that they tapped different factors and precluding their use as multiple indicators of a single construct. The zero-order correlations between these two items and the three retirement planning variables used in this study were also trivial and ranged from -.04 to +.04. When each item was added as a last step to the multivariate models, its contribution was statistically nonsignificant. More importantly, perhaps, adding these items did not affect the relationship of any of the other work-related rewards to retirement planning—a finding consistent with past research indicating that the social and financial rewards are largely unrelated (
Mannheim 1975). Because the goal of the present study was to determine if any work-related rewards were related to retirement planning rather than to represent all possible work-related rewards, and because the impact of financial rewards could not be fairly evaluated given the lack of appropriate items, financial rewards were not included in the set of work-related rewards. Pension status (coded "yes" or "no") was included as a control variable in all of the multivariate models, but it carried no evaluative judgment on the part of the respondent as to its sufficiency. It merely conveyed whether the person was currently in a job covered by a pension plan.
Retirement planning.
Three measures of planning activity were used. The first was a single indicator taken from an earlier study using the HRS that categorized respondents according to their intentions to (a) stop work altogether, (b) work fewer hours, (c) change the kind of work they did, and (d) never stop working. A final category comprised workers who had not given it much thought or had no current plans (
Ekerdt et al. 1996). This final category was of primary theoretical interest in the present study, because people may fail to form a plan because of the nature of the rewards associated with their jobs. Thus, a dichotomous variable was constructed that contrasted individuals in the first four categories who had definite retirement plans, including plans never to retire (coded 1), with those in the last category who had no definite plans (0).
The second measure was composed of two items that tapped respondents' expectation to remain at work in the future: (a) "What are the chances that you will be working full time after you reach age 62?" and (b) "What are the chances that you will be working full time after you reach age 65?" Responses were originally coded 0 (no chance) to 10 (certainty) but were reverse coded in the present study so that higher scores reflected a higher subjective probability or intention to be retired at each respective age.
The third measure of planning also involved two items: (a) "How much have you thought about retirement?" and (b) "How much have you discussed retirement with your friends or coworkers?" Responses were scored on a metric from 1 (a lot) to 4 (hardly at all) and reverse coded so that higher scores reflected high levels of retirement planning.
Other factors related to retirement plans and decisions.
Seven covariates were included in the analyses because they have been shown to be part of the relevant opportunity structure for holding retirement plans. These variables were education, age, female gender, health, marital status, White race, and availability of a pension plan. All of the covariates were single-item indicators. Respondents' ages were coded directly. Gender was coded 0 (male) or 1 (female). Health was assessed with the question, "Would you say your health is poor (1), fair (2), good (3), very good (4), or excellent (5)." Marital status was represented by a binary variable: married (1) versus all others (0). Seventy-seven percent of the sample (3,251 of 4,237) were coded 1 (see Table 1 ). Of these, 41% (1,335 of 3,251) were spouses of other respondents in the study. Race was similarly represented: White (1) versus all others (0). Whether the respondent was eligible for a pension plan was also coded as a binary variable: yes (1) or no (0). Years of education were coded directly, ranging from 0 to 17 or more.
Analysis Plan
Hypotheses 1 and 2 each propose a measurement model. In the case of Hypothesis 1, it is proposed that the three distinct forms of job-related rewards—intrinsic rewards, positive social relations, and opportunities for ascendance in the workplace—will conform to a three-factor structure with multiple manifest indicators to define each factor (Table 1 ). With respect to Hypothesis 2, three factors are also hypothesized in the model of retirement planning. Two of the factors are hypothesized as latent factors, each with two indicators (i.e., probability of retirement at the conventional age(s), and engagement in planning activities). The third factor is represented as a single, manifest, binary indicator (i.e., has a plan for retirement vs. not).
We used confirmatory factor analysis to test Hypotheses 1 and 2 using maximum likelihood (ML) estimation with EQS software. In the case of Hypothesis 1, we replicated the analysis using LISREL software to create matrices of polychoric correlations and asymptotic variances/covariances and then to re-estimate the model using weighted least squares (WLS) estimation.
Hypothesis 3 is a general hypothesis that posits an inverse relationship between the extent of work-related rewards and retirement planning, controlling for a set of seven covariates. We evaluated this general proposition as a set of three specific hypotheses, a, b, and c, with one hypothesis used to represent each retirement planning outcome variable. The test of each hypothesis involved regressing the relevant retirement planning outcome variable on the set of seven covariates and the set of three work-related rewards. In the case of the binary outcome (i.e., has a plan for retirement vs not), we did this using SPSS software to estimate a logistic regression model. For the other two outcomes, both latent variables, we tested the hypotheses using EQS software with ML estimation.
|
Results |
|---|
|
TOPAbstractRetirement and Job...Other Factors Related to...Retirement PlanningHypothesesMethodsResultsDiscussionReferences |
|---|
), even when only four measurement categories are used (
Green, Akey, Fleming, Hershberger, and Marquis 1997). The results indicated acceptable fit of the three-factor model (GFI = .99; AGFI = .98; CFI = .94; NFI = .93; NNFI = .90, SRMR < .06) with both conventional criteria (
Hoyle and Panter 1995) and emerging criteria (
Hu and Bentler 1999).
|
). The findings were highly similar to the ML results, with item factor loadings slightly larger, as would be expected, with the polychoric correlations. ML estimates of the correlations among the factors were intrinsic/social = .23; intrinsic/ascendance = .31; and social/ascendance = .05. Using WLS, we found that these estimates were .27, .30, and .10, respectively. Regardless of estimation technique, the correlations among the three factors were relatively low, supporting the view that these factors represented different types of work-related rewards.
Hypothesis 2: Model of Retirement Planning
To evaluate the hypothesis that the three planning variables represented a family of discrete, loosely related constructs (
Hanisch 1995), we performed a confirmatory factor analysis using ML estimation (see
Fig. 2). Model fit was good, 
2 = 2.5, with subjective fit indices nearly perfect. In addition, the correlations among the three different definitions of retirement planning were all in the hypothesized direction. The correlation between the two latent variables representing the intention of not working past 62 or 65 and engagement in planning acts was .35. The correlations between the single-item (has plan) variable and the two latent variables representing (a) the intention of not working past 62 or 65 and (b) engagement in planning acts were also positive, .18 and .38, respectively. The size of the correlations between factors suggests that these factors represent relatively distinct types of retirement planning behaviors.
|
2(10, n = 4,147) = 203; p < .001. As hypothesized, workers who reported higher levels of intrinsic gratification or positive social relations from work were significantly less likely to report having a plan for retirement. Contrary to expectations, however, workers with jobs high in opportunities for ascendance were significantly more likely to report having a plan for retirement. In agreement with prior research (
Ekerdt et al. 1996), higher education, being older, being male, and having a pension were all significantly related to having a plan for retirement.
Intention of not working past ages 62 or 65.
The results of the regression of the second retirement planning variable, the subjective probability of being retired at ages 62 or 65, on the set of 10 predictors, are also shown in Table 2 . We estimated the model using structural equations with maximum likelihood. We performed the tests of structural parameters using the Satorra-Bentler Robust procedure, which adjusts standard errors for the degree of multivariate kurtosis (
West, Finch, and Curran 1995).
Subjective indicators of overall model fit were acceptably high (GFI = .98; AGFI = .96; CFI = .92; NFI = .91; NNFI = .85; SRMR < .05). Only the non-normed fit index fell short of the conventional criterion of .9. The reason appeared to be that the NNFI extracted a penalty for excessive parameterization of a model (
Tanaka 1993). When nonsignificant covariances among the exogenous variables (which served primarily as statistical controls) were dropped from the model, the value for the NNFI became more consistent with the other indicators of model fit.
As hypothesized, the intention not to work past 62 or 65, that is, the intention to retire, was negatively related to intrinsic gratifications and social relations. Specifically, as intrinsic or social rewards increased, the intention to retire decreased. The relationship between the work-related gratification of ascendance and the intention of not working past 62 or 65, however, was nonsignificant. In addition, being female, experiencing poor health, being married, and having a pension were all significantly related to the intention of not working past the ages of 62 or 65.
To assess the contribution of the set of work-related rewards relative to the set of covariates, we also examined the two sets of variables hierarchically. A model containing only the seven covariates explained 5% of the variance in the intention to retire by age 62 or 65. When the set of work-related rewards was added to the model, the proportion of explained variance increased to 14%, Fchange (3, 4121) = 144.2; p < .01.
Thinking about retirement and discussing it with friends
The regression of planning acts, defined as (a) thinking about retirement and (b) discussing retirement with friends, on the set of 10 predictors is shown in the last column of Table 2 . Model fit was acceptable (GFI = .98; AGFI = .97; CFI = .92; NFI = .91; NNFI = .85; SRMR < .05). Once again, intrinsic and social rewards held significant, negative relationships to planning. That is, as these gratifications increased, the worker was less likely to think about retirement or discuss it with friends. Again, contrary to expectations, individuals with jobs high in opportunities for ascendance were significantly more likely to engage in retirement planning. In addition, higher education, increasing age, being male, experiencing poor health, being married, being White, and having a pension were all associated with greater planning for retirement.
To examine the relative contribution of the work-related rewards, we again examined the sets of variables hierarchically. A model containing only the seven covariates explained 10% of the variance in planning activities. When the set of work-related rewards was added to the model, the proportion of explained variance increased to 18%, Fchange (3, 3714) = 121.3; p <.01.
|
Discussion |
|---|
|
TOPAbstractRetirement and Job...Other Factors Related to...Retirement PlanningHypothesesMethodsResultsDiscussionReferences |
|---|
Past research indicates that workers who like their jobs are more likely to remain at them (
Glenn and Weaver 1985). The results of the present study suggest that this relationship between work and retirement is manifest upstream, at the stage where older workers would be anticipating retirement. Workers who like their jobs are also less likely to make plans for leaving them. This is an important elaboration because the impact of work gratifications on planning for retirement may be consequential for the individual, not only by delaying retirement but by forestalling and limiting options for financial, residential, and leisure arrangements. At the same time, the tendency for work gratifications to hinder retirement planning is evidence that employers, if so inclined, can hope to retain older workers if jobs are attractive enough (
Reich 1993).
In contrast, jobs high in opportunities for ascendance appear to promote certain types of retirement planning. Specifically, workers with jobs high in ascendance are more likely to think about retirement and discuss it with others and more likely to report having a plan for retirement. The mechanism underlying these relationships remains unknown. One explanation focuses on the nature of the job, such that persons occupying jobs high in opportunities for ascendance are more likely to be aware (or even required to be aware) of retirement options within the organization. Thus, there may be a fashioning effect operating on individuals in these positions that promotes retirement planning. That is, once inside a particular occupational role, the occupant comes to take on the characteristics of that role. In support of this explanation is the observation that persons with jobs high in ascendance are more likely to be engaged in the process of planning (i.e., to have a retirement plan and to think about retirement and discuss it with others), but they are not more likely to anticipate working past normative retirement age (62–65). Another possibility is that persons occupying jobs that provide opportunities for ascendance are also high in personality traits that promote planning. From this perspective, there is a selection effect operating to attract individuals who are more likely to be planners to these positions. Selection and fashioning effects are not mutually exclusive and could be operating jointly. A third possible explanation focuses on the nature of retirement planning for persons with jobs high in opportunities for ascendance. In analyses not shown here, we examined the specific forms of retirement intentions held by these workers. There was a clear tendency for those with jobs high in ascendance to favor partial retirement or changing jobs. This suggests that workers with jobs high in opportunities for ascendance may conceive retirement paths of a more complex kind that promote continued labor force participation of some form. Thus, in a manner consistent with the findings of
Hayward and Hardy 1985, occupations that involve high social skills and substantive complexity may create opportunity structures for continued involvement in the workplace. But this continued involvement requires more complex forms of planning. Seen from this perspective, the effects of intrinsic enjoyment, positive social relationships, and opportunities for ascendance on retirement planning are not so different in kind. All promote continued participation in the labor force.
The other covariates in the models are generally related to planning in intuitively straightforward ways and consistent with earlier findings (e.g., see
Ekerdt et al. 1996). Workers with higher education are more likely to report having a plan for retirement and to be thinking about it and discussing it more frequently with friends and coworkers. The same is true for older workers. Women are more likely to report having no plans and less likely to be thinking about retirement or discussing it. In addition, they reported a lower probability of working past normative retirement age. Because the work histories of women tend to be more complicated than for men because of discontinuities associated with childrearing, retirement as a normative transition may be less relevant to them. Workers in good health are less likely to think or talk about retirement and more likely to work past the normative retirement age. In this regard, health operates similarly to intrinsic rewards and positive social relations: When things are going well, workers are less likely to plan for contingencies. There is no relationship between health and whether the worker had a specific plan for retirement. Married workers are more likely to have a specific plan, more likely to be retired by the normative retirement age, and more likely to be thinking about retirement and discussing it with others. Perhaps the necessity of coordinating one's retirement with one's spouse forces an individual to be more planful (
Henkens 1999). The only racial difference is that White workers are more likely to report thinking about retirement and discussing it with others. And, not surprisingly, workers with pension plans are more likely to have a specific plan for retirement, more likely to anticipate being retired by the normative retirement age, and more likely to be thinking about retirement and discussing it with others.
Several shortcomings of the present study suggest the need for further research. First, retirement planning in the present study is linked to fairly specific behaviors and intentions. Each of these dependent variables is intended to offer a slightly different way of conceptualizing retirement planning and, it is hoped, a more robust look at the effects of work-related gratifications on workers' retirement plans. Nonetheless, what constitutes retirement planning is not confined to these specific behaviors or planning strategies. Second, the types of work-related rewards examined in this study were limited to those reasonably represented by questions in Wave 1 of the HRS. Clearly, an examination of the relationships between other types of work-related rewards and planning is both desirable and feasible. Third, how best to conceptualize job rewards is still unclear. Whether job-related rewards represent characteristic features of jobs, reflect dispositional characteristics of the job holders, or are best seen as interactions between the two remains an important issue in the classification and measurement of job satisfaction (
Stone 1992). Fourth, there is potential for sample selection bias in the present findings to the extent that the probability of being represented in the sample is not independent of the outcomes under study (i.e., retirement intentions and planning). Workers retiring before the age of 61 were not included. Because early retirement is likely to require both intention and planning, such selection effects are likely to exist in the present data to some unknown extent. Thus, the present findings are necessarily limited to the range of workers' ages studied. Fifth, generalization is further limited by the fact that the nature and meaning of work and retirement change over time. As a consequence, the results of the present study are necessarily temporally bounded and are relevant primarily to the age cohort of workers studied.
In sum, the present findings suggest that rewards obtained from working are differentially, and significantly, related to retirement planning. One of the most frequently researched topics in organizational behavior is work satisfaction, with a major focus on its antecedents and consequences (
Roberts 1994). Research has shown that satisfied workers are more likely to remain in their jobs and to have a greater commitment to organizational goals (
Mathis and Jackson 2000). Not surprisingly, then, many organizations routinely collect information on employee satisfaction (
Staples and Higgins 1998). The present findings suggest that such information can also be used as an important tool in understanding the conditions under which individuals plan for retirement. In the case of older workers, information about work-related rewards may be useful in targeting individuals who might benefit from programs designed to promote retirement planning, especially in extra-financial areas of planning. For program planners, knowledge about the manner in which work-related rewards affect retirement planning may facilitate the development of retirement planning programs that assist workers in realizing more complex retirement plans (e.g., reducing hours, finding part-time jobs that afford similar outcomes) that would allow them to continue to reap the benefits gained from working for as long as possible. And finally, for employers, knowledge of the factors that postpone retirement among older workers is useful if their goal is to retain these workers. If that is not their goal, then they can use this information to assist their older workers in post-retirement outplacement that would facilitate a smooth, orderly transition from the workplace.
|
Acknowledgments |
|---|
Received for publication March 8, 2000. Accepted for publication November 6, 2000.
|
References |
|---|
|
TOPAbstractRetirement and Job...Other Factors Related to...Retirement PlanningHypothesesMethodsResultsDiscussionReferences |
|---|
This article has been cited by other articles:
 |
|
 |
|
  D. J. Ekerdt and J. K. Hackney Workers' Ignorance of Retirement Benefits Gerontologist, August 1, 2002; 42(4): 543 - 551. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||
| HOME | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|---|
