20
453 Journal of Occupational and Organizational Psychology (2002), 75, 453—471 <Q'i 2002 The British Psychological Society www.bps.org.uk Underemployment and relative deprivation among re-ennployed executives Daniel C. Feldman *, Carrie R. Leana and Mark C. Bolino 'University of South Carolina, USA ^University of Pittsburgh, USA ^University of Notre Dame, USA Using a sample of 517 executives who lose their jobs as a result of downsizing, this study exainined underemployment among managers in replacement jobs taken after their layoffs. Laid-off executives who were re-employed in jobs which paid less, were at lower levels of organizational hierarchies and which did not fully utilize their skills had consistently lower job attitudes. In addition, the results suggest that relative deprivation is an important mediator in explaining how underemployment leads to poorer psychological well-being in those replacement jobs. The article concludes with directions for future research on underemployment and relative deprivation in the aftermath of layoffs. Over the past several years, there has been considerable attention given to the short-term effects of job loss on individuals who have become unemployed as a result of downsizing. That research has consistently found that layoffs have immediate, and often strong, negative consequences for employees' psychological and physiological well-heing. What has received considerably less research attention, however, has heen the long-term effects of layoffs on the careers of downsized employees. As Leana and Feldman (1995) note, most of the research on joh loss has taken as its end point the attainment of re-employment; that is, when laid-off workers get new jobs, they typically cease to be the focus of future research. Nevertheless, in those studies that have examined the job histories of employees who have been downsized out of jobs, the results suggest that the quality of replace- ment jobs - and not just the fact of re-employment - is a major determinant of laid-off employees' subsequent career trajectories. Indeed, many downsized employees end up underemployed' in jobs which require less education and experience than they possess, in positions which pay considerably less than the jobs from which they were laid off, or in positions at lower levels of the organizational hierarchy (Feldman, 1996; Kaufman, 1982; Kuttner, 1994). The present study builds upon previous research on underemployment in three ways. First, it systematically examines the effects of underemployment on the job attitudes and work attachment of employees. Initial studies of underemployment sug- gest that laid-off workers re-employed in lower quality jobs have more negative job ^Requests for reprints should be addressed to Professor Daniel C. Feldman, University of South Carolina, The Daria Moore School of Business, Deportment of Management, Columbia, SC 29208, USA (e-mail: [email protected]).

Underemployment and relative deprivation among re-employed executives

Embed Size (px)

Citation preview

453

Journal of Occupational and Organizational Psychology (2002), 75, 453—471

<Q'i 2002 The British Psychological Society

www.bps.org.uk

Underemployment and relative deprivationamong re-ennployed executives

Daniel C. Feldman *, Carrie R. Leana and Mark C. Bolino'University of South Carolina, USA^University of Pittsburgh, USA^University of Notre Dame, USA

Using a sample of 517 executives who lose their jobs as a result of downsizing, thisstudy exainined underemployment among managers in replacement jobs taken aftertheir layoffs. Laid-off executives who were re-employed in jobs which paid less,were at lower levels of organizational hierarchies and which did not fully utilize theirskills had consistently lower job attitudes. In addition, the results suggest thatrelative deprivation is an important mediator in explaining how underemploymentleads to poorer psychological well-being in those replacement jobs. The articleconcludes with directions for future research on underemployment and relativedeprivation in the aftermath of layoffs.

Over the past several years, there has been considerable attention given to theshort-term effects of job loss on individuals who have become unemployed as a resultof downsizing. That research has consistently found that layoffs have immediate, andoften strong, negative consequences for employees' psychological and physiologicalwell-heing. What has received considerably less research attention, however, has heenthe long-term effects of layoffs on the careers of downsized employees. As Leana andFeldman (1995) note, most of the research on joh loss has taken as its end pointthe attainment of re-employment; that is, when laid-off workers get new jobs, theytypically cease to be the focus of future research.

Nevertheless, in those studies that have examined the job histories of employeeswho have been downsized out of jobs, the results suggest that the quality of replace-ment jobs - and not just the fact of re-employment - is a major determinant of laid-offemployees' subsequent career trajectories. Indeed, many downsized employees end upunderemployed' in jobs which require less education and experience than they

possess, in positions which pay considerably less than the jobs from which they werelaid off, or in positions at lower levels of the organizational hierarchy (Feldman, 1996;Kaufman, 1982; Kuttner, 1994).

The present study builds upon previous research on underemployment in threeways. First, it systematically examines the effects of underemployment on the jobattitudes and work attachment of employees. Initial studies of underemployment sug-gest that laid-off workers re-employed in lower quality jobs have more negative job

^Requests for reprints should be addressed to Professor Daniel C. Feldman, University of South Carolina, The DariaMoore School of Business, Deportment of Management, Columbia, SC 29208, USA (e-mail: [email protected]).

454 Dar}iel C. Fddman et al.

attitudes towards their new employers, invest less energy in their new jobs and aremore likely to keep searching for different jobs even after accepting those new pos-itions. In fact, in some previous research, the negative effects of underemploymenthave been found to be as detrimental to psychological well-being as unemploymentitself (Kuhnert. 1989; Leana & Feldman, 1995; O'Brien, 1986; O'Brien & Feather,1990),

Second, the present study examines relative deprivation as a potential mechanismfor understanding how underemployment leads to negative jofi attitudes. Althoughunderemployment has often been found to he negatively correlated with variousmeasures of psychological well-being, the underlying reasons for these relationshipshave not been explored fully. Here, we suggest that underemployment may lead tonegative joh and career attitudes because laid-off workers both desire and feel entitledto have better jobs than those they occupy at this point in their careers. In turn, thisdesire for, and sense of entitlement to, better jobs may directly lead to more negativejob attitudes (Cxosby, 1976; Martin, 1981).

Third, this research uses a large, heterogeneous sample of downsized executives toexplore these issues of underemployment. To date, most of the research on under-employment has focused on two groups of workers: laid-off blue-collar workers (eg.Leana & Feldman, 1995: Liem & Liem, 1988) and underemployed high school orcollege graduates (eg. Feldman & Tumley, 1995; Winefield & Tiggemann, 1990;Winefield, Winefield, Tiggemann, & Goldney, 1991). These workers have been fre-quently investigated because of the harsh financial difficulties underemployment islikely to impose upon them. The present sample allows for a more thorough investi-gation of the psychological as well as the financial prohlems associated with under-employment.

Theory

Construct of underemploymentAs Feldman (1996, p. 387) notes in his review of the underemployment literature, theconstnict of underemployment has been defined in a variety of ways across academicdisciplines. All definitions of the term, however, have conceptualized underemploy-ment as jobs which are lower in quality in some 'way. In the organizational sciences,three conceptuaHzations of underemployment have received the most attention.

First, underemployment has often been defined in terms of the hierarchical level ofthe new position. Commonly, employees laid off from 'permanent' full-time jobs findthemselves working in part-time or temporary johs or, particularly in the case ofmanagers, at lower hierarchical levels of organizations (Buss & Redburn, 1983; Gordus,Jarley, & Fcrman, 1981).

Seeond, underemployment has frequently been defined in terms of loss of wages.For example, Zvonkovic (1988) defines underemployment as 'current earnings at least20% less than earnings in the previous job . Reflecting the severity' of economicconditions, earlier research on re-employment problems among workers during theDepression often used a 33% rate of income loss as the standard for underemployment(Elder, 1974).

Third, underemployment has often been conceptualized in terms of skill utilization(eg. Clogg & Shockey, 1984; Ciogg, Sullivan, & Mulchler, 1986; Humphr>'s & OBrien,1986), For instance, in work on teenaged 'school leavers' (Feather & O'Brien, 1986;Winefield et al., 1991) and college graduates (Feldman & Turnley, 1995), researchers

Underemployment and relative deprivation 455

have focused on the extent to which individuals have jobs which do not fully utilizethe skills and abilities they learned in school. Similarly, in their research on under-employment among contingent workers, Feldman and Doerpinghaus (1992) askedindividuals to report whether their jobs could be performed adequately by people withconsiderably less education and work experience than they themselves possessed.

As the preceding discussion illustrates, there is little consensus on how under-employment is defined and operationalized. These various conceptualizations haveoffered an expansive description of the underemployment experience, but theyhave not led to entirely consistent findings in terms of the consequences ofunderemployment.

Consequences of underemploymentUnderemployment has been linked with a variety of negative job attitudes and indi-cators of poor psychological well-being. The job attitude most commonly associatedwith underemployment is job dissatisfaction (Borgen, Amundson, & Harder, 1988;Khan & Morrow, 1991). Dissatisfaction with the work itself has a strong impact onglobal job dissatisfaction (Hackman & Oldham, 1980). Moreover, disappointmentswith pay and promotional opportunities are also likely to be greater in underemploy-ment situations (Khan & Morrow, 1991).

Underemployment is also frequently associated with lower organizational commit-ment. Wben individuals find themselves in positions where they perceive a majordiscrepancy between the rewards they receive and the rewards they are used toreceiving, they are likely to reduce that inequity by psychologically distancing them-selves from their employers and lowering their contributions to their organizations.Borgen etal. (1988) and Leana and Feldman (1995), for instance, found that under-employed workers felt they had less reason to be committed to their jobs or tocontribute over and above the call of duty.

In general, trust has been defined as an individual's generalized positive expec-tations of another's goodwill, willingness to reciprocate, and honouring of commit-ments. Many employees losing jobs tbrougb downsizing feel that organizations violatethe implicit agreements they have with them about job security and procedural fair-ness (Rousseau, 1995). Consequently, underemployed executives in replacement jobs'may be particularly sensitive to unmet expectations in tbeir new jobs, may reactmore negatively to those unmet expectations, and may have less faith in their newemployers in general (Rousseau. Sitkin, Burt, & Camerer, 1998).

Along the same lines, underemployment has consistently been found to beassociated with greater job-searcbing behaviour. Burris (1983) found that under-employed workers were less likely to give tbeir jobs one year' to improve beforeleaving, whereas Leana and Feldman (1995) found tbat laid-off workers who wereunderemployed in their new jobs were more likely to continue to job bunt even afterbecoming re-employed. Robinson, Kraatz, and Rousseau (1994) also found that recentMBA graduates wbo felt their psychological contracts had been violated were lesslikely to stay two years with their employers and were less likely to give advance noticebefore leaving. For underemployed workers, then, finding replacement jobs does notend tbe search for satisfactory re-employment.

Finally, underemployment may have negative spillover effects on individuals' long-term attitudes towards work and tbeir careers more generally. For example, Borgenet al. (1988) found that underemployed college graduates were more disillusioned

456 Daniel C. Feidmar) et ai.

with their job situations, more frustrated by lack of opportunities for advancement and'more worried about being stuck'. Leana and Feidman (1992) found that downsizedworkers had less sense of excitement about their careers and lower career investmenteven after they had found replacement jobs. In a study of MBA graduates, Rousseau(1990) found that individuals who had their psychological contracts violated weremore likely to be cynical about the relationship between hard work and career suc-cess. As Feidman (1996) suggests, then, underemployment may lead to more careeristattitudes toward ^vork' and an increased reliance on nonperformance-based tactics(such as networking and impression management) to get ahead.

In terms of the main effects of underemployment on outcome variables, then, fivehypotheses can he proposed:

HI : Underemployment will be negatively related to job satisfaction.H2: Underemployment will be negatively related to organizational commitment.H3: Underemployment will be negatively related to trust in the organization.H4: Underemployment will be positively related to careerist attitudes towards work.H5: Underemployment will be positively related to continued job searching.

Mediating effects of relative deprivationDuring the I97()s, organizational researchers became increasingly interested in theeffects of equity, broadly defined, on employees' responses to their jobs. For example,Lawler (1973) posited a discrepancy model of joh satisfaction. His research suggestedthat workers' satisfaction with their jobs was not only a function of how positiveactual job conditions were (eg. pay, work itself, supervision) but also a function ofwhat job conditions employees felt should exist. Locke (1976) also suggested thatemployees' satisfaction with their jobs was influenced not only hy objective job con-ditions but also by how well jobs fulfilled individuals' basic values and the importanceof those values. Adams's 'equity theory', too, proposed that individuals' satisfaetionwith job rewards was influenced by comparisons with co workers. In equity theory, anindividual's satisfaction with pay is determined by how the ratio of an individual s jobrewards to job inputs stacks up against the ratio of job rewards to job inputs of his/hercolleagues (Adams, 1976),

Relative deprivation theory, too, addresses the role of comparisons in shaping indi-viduals' attitudes. However, rather than coming from the organizational sciences andfocusing on individuals' assessments of specific jobs, relative deprivation came fromthe social psychology literature and focused on individuals' seuse of injustice withvarious societal conditions, Stouffer, Suchman, DeVinney, Star, and Williams (1949)first introduced the term relative deprivation' to explain why levels of satisfactionamong soldiers did not consistently coincide with their objective job conditions. Theirresults suggest that individuals' job attitudes are at least partially influenced by howobjective joh conditions match up to what individuals desire and feel entitled toreceiving from their jobs. Since its introduction, relative deprivation theory has beenused by researchers to explain social problems (such as race and gender discrimi-nation) in which people's suhjective feelings and objective circumstances do notmatch (Crosby, 1976, 1982),

More recently, researchers in the organizational sciences have used relativedeprivation theory to explain individuals' reactions to work-related problems such asinequities in pay rises and promotion decisions (Buunk &Jansst;n, 1992; Martin, 1981;

Ur\derempioymer}t and relative deprivation 457

Sweeney, McFarlin, & Inderriedeti, 1990). Relative deprivation may be a particularlyappropriate approach to examining the underemployment phenomenon becausemuch of the dissatisfaction underemployed workers experience may he the result oftheir experiences with previous employers or frustrated hopes of obtaining betteremployment in the futtire rather than from injustices at the hands of their presentemployers. For example, whereas equity theory examines how employees assess thefairness of their job rewards relative to their present colleagues, relative deprivationtheory allows us to examine the comparisons underemployed workers make to thejobs they lost and to the jobs they hope ultimately to obtain.

In short, then, whereas the construct of underemployment refers to 'objective'indices of job quality (such as lower pay, lower hierarchical level or lower skillutilization), the constritct of relative deprivation refers to individuals' 'subjective'reactions to their employment predicaments. Relative deprivation theory suggests thathow negatively employees react to underemployment will depend upon how much anindividual wants job rewards, feels entitled to those rewards, and his/her standards ofcomparison for assessing the fairness or justness of the rew ards they do receive.

For example, middle-aged executives who have lost their johs may be tnuchmore desirous of obtaining replacement jobs at equally high levels of pay, whereasexecutives in their late 60s may not have high needs for replacement johs at all. Alongsimilar lines, workers who hold graduate degrees may feel more entitled to jobs thatutilize their extensive education, whereas workers without high-sehool diplomas maynot have such high expectations of obtaining self-actualization from their jobs. Iflaid-off executives use laid-off factory workers from their own firms as a standard ofcomparison for assessing their job predicament, they may feel relatively lucky; if theyuse other executives who escaped being downsized, they may feel much worse abouttheir state of affairs.

Here, we suggest that relative deprivation mediates the relationship between under-employment and important job outcomes. Using a discrepancy approach similar toprevious research on relative deprivation, the present research suggests that indi-viduals compare their present job situations with those they want and with which theyfeel entitled. The greater the discrepancy between present job conditions and desiredjob conditions (that is, tbe greater the relative deprivation), the more negativeemployees' job attitudes will be. Thus, it is through generating relative deprivation thatunderemployment may lead to negative job and career attitudes (Feldman, Leana, &Turnley, 1997).

H6: Relative deprivation mediates the relationships between underemployment and joboutcomes.

Method

SompleThe participants in the study consisted of 517 senior managers who had been laid offfrom their jobs within the past 12 months. All of the participants were alumni' of aninternational outplacement firm headquartered in tbe northeastern United States ofAmerica. As graduates" of the outplaeemetit programme, all had found new jobs.Names and mailitig addresses were provided to the researchers by the outplacementorganization. Surveys (along with self-addressed, stamped return envelopes to theresearchers) were sent to a total of 1700 individuals who had used the services ofthe outplacement firm in the past year. The response rate was 30%.

458 Daniel C. Feldman et al.

The average age of the participants in the study was 46 (SD=7.75). The sample was74% male and 26% female. Seventy-seven percent of the respondents were married. Onaverage, respondents had worked for the organization from which they were laid offfor 12 years (SD=9.11). The average respondent received six weeks' advance notice ofhis/her termination (SD=65.12 days).

Following their layoffs, it took the participants in the study an average of five monthsto transition into their current positions (SD=5 months). The average respondent inthe study had been working for hisAier current organization for eight months at thetime of the survey (SD=6 months). On their replacement jobs, 33% earned over$100,000 per year. 50% earned between $50,000 and $99,000 and 17% earned annualsalaries of less than $50,000. ln terms of job functions, 28% were re-employed itimarketing and sales, 14% in finance and accounting, 19% in engineering, operationsand information technology, 27% in corporate and general management and 12% inother functional areas such as business law .

Although, ideally, it would have been preferable to bave a higher response rate, aresponse rate of 30% is not atypical in mail survey research, particularly when many ofthe participants who lost tbeir jobs may well have changed residences after using theoutplacement services. To check the representatives of the sample, sample characteris-tics were compared with data provided by the outplacement firm on the total popu-lation of alumni that graduated from their executive programmes during the sameperiod. The profiles were highly similar. The average age of the alumni population wasalso 46, and 75%) of the alumni were male. The reported mean salary data on tbepopulation for the replacement job was $85,000; this study asked for salary rangesrather than exact salaries, but tbe ranges (reported above) are in line witb the popu-lation average. For the population as a whole, 25%) found replacement johs in market-ing and sales, 14% in finance and accounting, 22%) in engitieering, operations and infor-mation technology and 28% in corporate and general management. Agiun, thisdistribution generally parallels the functional specialities of the sample describedabove.

MeasuresThe means, standard deviations and correlations among all the key variables in thestudy appear in Table 1. Cronbach s alpha for measures, where available, appear alongtbe diagonal. Tbe measures are described in more detail below.

UrideremphymentIn the present study, we examined the three most common ways that underemploy-ment has been assessed. In all three measures, the higher tbe score the greater thedegree of underemployment.

First, underemployment was assessed by measuring how the hierarchical level ofrespondents' current jobs compared with that of the jobs from which they wereoutplaced. Responses ranged from (1) current job is at a much higlier level than theone I had' to (5) current joh is at a much lower level than the one I had'.

Second, underemployment was assessed by measuring how great a pay differencethere is between respondents' current jobs and those from whicb they wereoutplaced. Responses ranged from (1) current job pays 20%- or more' to (5) currentjob entails a pay cut of 20% or more'.

Tbird, underemployment was measured by examining bow respondents' skillutilization on tbeir current jobs compared witb that on the jobs from which they were

Underemployment and relative deprivation 459

-o•a

I I

p p p p p p p OI I I I II

O L r i o O — — O — O O O — O

(N — O C O — — u n f O O O O ^ r S• ^ ^ ^ p^ QN O CO -"" 00 ^~ * ^ Ci ro^ ^ *! I* ^ ' f**J ro oJ ^( ro m fo fS fn

§ G £ "S~ <U .?^ .N

= ^ ^ ?«< I- U I 0 , 0

460 Daniel C. Feldman et al.

outplactd. Respondents were asked about their present skill utilization compared withtheir skill utilization in their previous jobs in nine areas: supervisory skills (managingpeople), administrative skills (managing projects), industry knowledge, technical/functional skills, knowledge of markets and competitors, communication skills,negotiation skills, complex decision-making skills, and financial and budgeting skills.Responses ranged from (1) 'use much more on current job tban on the job I lost' to (5)use much less on current job than on the job 1 lost'.

Relative deprivation

Whereas Crosby's original conceptualization of relative deprivation consisted of sixcomponents, more recent work suggests that two components (wanting more andfeeling entitled to more) account for most of the variance in relative deprivation(Feldman et al.. 1997; Olson & Hafer, 1996). In this study, 10 items from Olson. Roese,Meen, and Robertson (1995) were used to assess relative deprivation.

Five items from Olson et al. (1995) were used to measure the extent to whichrespondents desired better job situations. Tbe alpba for tbese 5 items w as .88. Asample item is; 'Would you like a job situation tbat is better in terms of job challenge?'Five items from Olson et al. (1995) were also used to measure the extent to whichrespondents feel entitled to better job situations. Tbe alpba for tbese five items was.92. A sample Item is: Do you tbink you ought to have a job situation tbat is better interms of job responsibility?' Responses ranged from (1) not at all' to (7) to a greatextent'.

JobJob satisfaction was measured using 5 items from Taylor and Bowers (1972). Respond-ents were asked how satisfied they were with variou.s aspects of tbeir current jobs(e.g. tbeir supervisors and tbeir working conditions). Responses ranged from (1) 'verydissatisfied' to (5) 'verj' satisfied'.

Orgar)izQtiQr)al comm\tmer)t

Organizational commitment was measured using Meyer and Allen's (1984) 8-itemaffective commitment scale. Tbe scale taps the extent to whicb individuals feelemotionally attached to their organizations and feel that their organizations have agreat deal of personal meaning for them. Responses ranged from (1) 'strongly disagree'to (5) strongly agree'.

TrustRobinson's (1996) 7-item scale was used to measure respondents" trust in tbeir currentemployers. For instance, respondents were asked to agree or disagree with statementsthat their employers were honest, truthful, consistent and predictable. Responsesranged from (1) 'strongly disagree' to (5) strongly agree.

Careerist attitudes toward workCarccrism was measured using 7 items from Feldman and Weitz (1991). This scalemeasures the extent to which individuals agree that advancement in organizations isbased more on image management and personal connections than on competence.Responses ranged from (1) strongly disagree' to (5) strongly agree'.

Underemployment and relative deprivation 461

job searching

Individuals' job searcb bebaviours were measured using 5 items from Blau (1993). Thescale taps tbe extent to which respondents are circulating tbeir resumes at othercompanies and currently interviewing for otber jobs. Responses ranged from (1)strongly disagree' to (5) 'strongly agree\

Control variables

Four control variables were also included in the study: gender, amount of advancenotice of layoffs, amount of time required to obtain the current job, and tenure in tbecurrent organization. Tbe variables were measured by close-ended items; descriptivedata on these variables appear in the Sample section above.

Results

Main effects of underemploymentAs the results in Table I suggest, Hypotheses 1-5 are generally supported. Whenmeasuring underemployment in terms of hierarchical level and skill utilization, under-emplo>TTient is significantly related to all five outcome variables and in the predicteddirection. Re-employed managers who liad taken larger pay cuts were significantly lessjob satisfied, less organizationally committed and more likely to be job searching. Paycuts are negatively related to trust and positively related to careerism, as predicted, butonly at tbe .10 significance level.

Subsequently, five regression models were nm to determine the joitit effects of thethree underemployment indicators (skill utilization, pay difference and hierarchicallevel) on each of tbe five outcome variables (job satisfaction, commitment, trust, jobsearching and careerism). In each case, tbe skill utilization operationalization wassignificantly related to the outcome variable at the .01 or .001 level. In contrast,the pay difference variable was not significantly related to any of the five outcomevariables. Hierarchical level was only significantly related to the organizationalcommitment variable.

Mediating effects of relative deprivationIn order to test the mediating effects of relative deprivation (H6), the regressionprocedures recommended by Baron and Kenny (1986) were employed. Baron andKenny (1986) indicate tbat three conditions are necessary to demonstrate mediation:(1) the independent variable must be significantly related to tbe dependent variable;(2) tbe independent variable must be significantly related to the mediator variable; and(3) tbe mediating variable must be significantly related to the dependent variable.

hi order to conclude that tbere is full mediation, the independent variable has tohave no significant effect on the dependent variable when controlling for the mediat-ing variable. If the effect of the independent variable on the dependent variable is lesswben controlling for the mediating variable than when not controlling for it, partialmediation is said to be demonstrated.

As tbe results in Table 1 suggest, witb only two exceptions (tbe effects of paydifference on trust and careerism), all the independent variables are significantlyrelated to all the dependent variables. The first condition of Baron and Kenny (1986),then, is generally met. The results in Table 1 also suggest tbat tbe second condition isclearly met; all tbree operationalizations of underemployment are significantly related

462 Daniel C. Feldman et al.

Table 2. Decreases in betas when controlling for mediator effects of relative deprivation

Dependent variables

Independent variable: Hierarchical levelJob satisfactionOrganizational commitmentTrustCareerismJob search

Independent variable: Pay differenceJob satisfactionOrganizational commitmentTrustCareerismJob search

Independent variable: Skill utilizationJob satisfactionOrganizational commitmentTrustCareerismJob search

BetaEqn 1"

_ 26***_ 27***_ 19***

13**19***

- .20**-*_ 13**

- .08.08.14**

_ 33***

- .40***_ 23***

2 1 * * *23***

BetaEqn 2

_ I I * *_ 1 1**-.06

.00

.02

-.03.05N/AN/A.02

-.16***_ 23***- .09*

.07

.03

Mediation

PartialPartial

FullFullFull

FullFull

None (Condition 1 not met)None (Condition 1 not met)

Full

PartialPartialPartial

FullFull

Note. Coefficients are standardized regression coefficients."Equation I: Effect of underemployment on outcome variable.''Equation 2: Effect of underemployment on outcome variable when controlling for the effects of the mediator.

to the mediating variable (relative deprivation). Moreover, the third condition is alsomet; relative deprivation (the mediator) is significantly related to all five outcomevariables (cf. Table I).

Regression analysis was then used to determine whether relative deprivation (themediator) is significantly related to the dependent variables while controlling Ior theindependent variables (underemployment measures). The results directly pertaining tothis analysi.s are shown in Table 2.

Table 2 compares the beta of underemployment on outcome variables (Column 1)with the beta of underemployment on outcome variables when controlling for theeffects of relative deprivation (Column 2). Column 3 indicates whether the mediatingeffect was full, partial or nonexistent.

In almost every case, relative deprivation fully or partially mediates the relationshipbetween underemployment and outcome variables, generally supporting Hypothesis 6.For example, when underemployment is operationahzed as a drop in hierarchicallevel, relative deprivation fully mediates the relationships between underemploymentand organizational tnist, careerism, and job .search. In these three cases, the beta ofunderemployment in Equation 1 is statistically significant; however, when theeffects of relative deprivation are controlled for, the beta in Equation 2 is no longerstatistically significant. Similarly, when underemployment is operationalized as a dropin hierarchical level, relative deprivation partially mediates the relationship between

Underemployment and relative deprivation 463

underemployment and job satisfaction; wben relative deprivation is added in as amediator in Equation 2, the beta for underemployment substantially drops inmagnitude but is still statistically significant.

Structural equation modelsAs an alternative means to testing mediation, a structural equation modelling approachwas employed. 'Whereas the results of structural equation models are often consistentwith multiple regression analyses, these two approaches look at our research questionin ways that are fiindamentally different. The regression analyses allow us to assessmediation while considering the relationships between each type of underemploy-ment (i.e. bierarcbical level, pay difference and skill utilization) and each specific joboutcome (satisfaction, commitment, tnist, careerism and job search) independently. Incontrast, the stmctural equation models let us assess mediation using more globalmeasures of underemployment and job outcomes, thereby allowing us to examinethe overall relationships among underemployment variables, relative deprivationvariables, and job outcome variables at the same time and within one analysis.

Models tested

Two models were evaluated in order to determine the nature of the mediating role ofrelative deprivation. Model 1 is a fully mediated model. In this model, under-employment only influences job outcomes through the intervening effects of relativedeprivation. iModel 2 is a partially mediated model. In this model, underemploymentinfluences job outcomes both indirectly (through relative deprivation) and directly.Thus, Model 2 is identical to Model 1 except for the direct path that has been addedbetween underemployment and job outcomes. In both models, there are paths fromthe four control variables (i.e. gender, advance notice, transition time and currentorganizational tenure) to both relative deprivation and the job outcome variables.

Latent and indicator variables

The first step in preparing the structural equation models was identifying the latentand indicator variables. The underemployment latent variable was represented interms of the three forms of underemployment examined here, namely, hierarchicallevel, pay cut and skill utilization. Two indicator variables, wanting and entitlement,were used to assess the latent variable of relative deprivation. One of the indicatorvariables included the 5 items tapping 'wanting a better job'; tbe second indicatorvariable included the 5 items tapping feeling entitled to a better job\

The flve measures of the job outcomes discussed earlier (e.g. job satisfaction,organizational commitment, etc.) were used as indicator variables for the job out-comes' latent variable. Because job-searching behaviours and careerist attitudes sbouldbe negatively associated with the other outcome variahles (e.g. job satisfaction isnegatively related to job searcb), it is expected that the loadings of these two indicatorvariables on to this latent constnict will be negative.

Measuremer^t model

Before testing the goodness of fit of these two models, a measurement model with thetbree latent variables (i.e. underemployment, relative deprivation and job outcomes)was evaluated. The measurement model fit the data well (x^=139.57, df=32; GFI=.95;AGn=.91; CFI=.96; TLI=.94; RMSEA=.O8).

464 Daniel C. Feldman et al.

Confirmatory factor analysis

Also before testing the goodness of fit of these two models, a confirmatory factoranalysis was conducted to determine whether the five dependent variahles discussedahovc (job satisfaction, commitment, trust, job hunting and careerism) representeddistinct constructs. We compared the fit of a five-factor model with that of a one-factormodel. The fit of the five-factor model was good (%'= 1308.76. df=454; GFI=.85;CFI = .9O; TLI = .62; RMSEA=.()6). In contrast, the fit of the one-factor model was poor(%^=3474.00, df=464, GF1=.6O; CFI=.64; TLI=.62; RMSEA=.12). Moreover, the differ-ence in the overall fit of the five-factor model vs. the one-factor model was statisticallysignificant (change in %" = 2l65, df= 10). These data further support the position thatthe five-factor model is superior to the one-factor model.

GaodnesS'Offit of models

The goodness-of-fit index for Model 1 is .95 (x^=196.18, df=67; AGFI = .91; CFI=.95;TLI = .93; RMSEA=.O6). Overall, the fit indices provide strong support for the model.Also, support for full mediation is shown by the significant paths between under-employment and relative deprivation and between relative deprivation and job out-comes. None of the control variables had a statistically significant effect on relativedeprivation or the outcome variables (Fig. 1).

For Model 2, the goodness-of-fit index is .93 (x^= 193.37, df=:66; AGF1 = .91; CFI = .95;TLI=.93; RMSEA = .O6). Here, too, none of the control variahles had a statisticallysignificant effect on relative deprivation or the outcome variables (Fig. 2).

The indices suggest that this model is a good fit for the data as well. However, inModel 2, the direct path between underemployment and job outcomes is not statisti-cally significant. In addition, the change in x" between Model 1 and Model 2 was only2.81 (df=l); this change was not statistically significant. Taken together, thesefindings provide greater support for the fully mediated model (Model 1) in whichunderemployment affects job outcomes through its effect on relative deprivation.

job tenureBecause it could be argued that the effects of relative deprivation and underemploy-ment might change over time, further post hoc analyses were conducted on therelationships between job tenure and relative deprivation and underemployment inparticular.

The zero-order correlations of job tenure with relative deprivation and the threeindicators of underemployment are all below .09. Wlien tenure is dichotomized using amedian split, again there are no significant differences in relative deprivation or under-employment. For example, the mean on skill utilization for the low-tenure group is2.87, whereas the mean on skill utilization for the high-tenure group is 2.75. The meanon relative deprivation for the low tenure group is 4.19; the mean for the high-tenuregroup is 4.10. The effects of job tenure as a trichotomized variable on relativeredcprivation and underemployment yield similarly non-significant results. Overall,then, the evidence suggests that job tenure does not explain a significant amount ofvariance in the present results.

Moderator analysesFinally, additional post-hoc analyses were conducted to determine if relative depri-vation moderated (rather than mediated) the relationship between underemploymentand joh outcomes. Whereas the relative deprivation literature has typically viewed

Underemployment and relative deprivation 465

relative deprivation as a mediating variable which precedes affective outcomes(cf. Crosby, 1982, 1984), the underemployment literature has typically viewed relativedeprivation as a moderating variable between underemployment and job outcomes(cf. Feldman, 1996; Feldman et al., 1997). Consequently, examining the potentialmoderating effects of relative deprivation as well as its potential mediating effects inthe context of underemployment miglit prove illuminating in terms of fiiture theorydevelopment.

Moderated regression analysis was used to test for these potential moderatingeffects. Fifteen models were estimated: the three forms of underemployment eachinteracting with relative deprivation on the five outcome variables. The results failed tosupport the idea that relative deprivation moderates the relationship between under-employment and job outcomes. Indeed, of the 15 interaction terms examined, only 1was significant (skill underutilization x relative deprivation on job satisfaction). Relativedeprivation, then, appears to play a mediating role rather than a moderating role inexplaining reactions to underemployment.

DiscussionThe results of this study suggest several important conclusions. First, underemployment -regardless of how it is operationalized - is associated with a number of negativeattitudes toward jobs and careers. Moreover, these fmdings are consistent across thethree different measures of underemployment examined in this research. Thus, thenegative consequences of underemployment for job attitudes are robust and not likelyto be tbe result of measurement artifacts alone.

Second, the results of this research suggest that declines in skill utilization, ratherthan pay cuts or demotions, play the greatest role in negative reactions to under-employment. Unfortunately, this component of underemployment has received theleast attention from academics and managers alike (Bolino & Feldman, 2000). Foracademics, the results here suggest that skill underutilization needs to he assessed asfully as pay cuts (which have historically dominated the literature on underemploy-ment) in understanding managers' reactions to replacement jobs. For outplacementcounsellors and for downsized managers who are looking for replacement positionsthemselves, the results here suggest that skill utilization is as important, if not a moreimportant criterion for assessing the appropriate potential of new jobs. Ironically, evenat the major outplacement firm where this research was conducted, there has been notracking of skill utilization in replacement jobs; rather, the traditional measures ofsalary and hierarchical level have been used instead.

Third, the effects of underemployment on outcomes are significantly mediated byrelative deprivation. Both the mediated regression analyses and the stmctural equationmodels suggest that underemployment generates feelings of relative deprivationwhich, in turn, adversely affect individuals" attitudes toward both their presentjobs and their careers more generally. Thus, relative deprivation can be a powerfulexplanatory mechanism in understanding the effects of underemployment.

Fourth, the results of the present research suggest that underemployment is aproblem that affects executives as well as lower paid employees. Not surprisingly,most of the previous research on underemployment has focused on low-paidemployees because of the financial hardship unemployment - and replacement jobsthat pay even less - can have on this population. However, even with bigbly paid man-agers, underemployment can have adverse consequences for job attitudes and job

466 Dome) C. Feldman et al.

jT

a.

II

, GFI

^=6

3;

II

•o

o

^qua

re-

1

u

-.07.

w

RM

S

of

if

.94:

TL

I

\\

o

o

ath:

a.

N

Stan

dard

its

are

c

' o

ath

coef

Cu

<

craa

cOC

Non

-;.0

5.at

p<

i

gnif

ic

(U1- .

ra

lid l

ines

oinra

C/l

d lin

esh

eby

da

rao

are

indi

raa.

Underemployment and relative deprivation 467

iio

oll'

a

OSl l '

;GFI

=

fN

w

IcrCO

O

.=.0

7.

00

Pi

OS

jf

OS

o1/1

ra

•T3tuN

ndar

di

C.2io

coef

f

a.

<

ican

t

'c2?'iflC

Z

o

at p

<

crao

gnif

i

'En<Um

line

s;so

l:

ra

;d l

ine

C/3

ra•o>-,

X)X)

raf 1

: ind

ii

rain

a.Ul

468 Daniel C. Feldman et al.

performance. The dominant models of job satisfaction, even those including equityconsiderations, have generally suggested that employees' reactions to their jobs arelargely a function of their experiences with tht:-\r present employers. The results here.suggest that the job attitudes of laidolf managers may be due as much to theirexperiences with their yM.sA employers as to any ill behaviour on the part of currentemployers.

Limitations of the researchBecause the present study utilized self-report data, the results here may be inflated dueto common-method bias. In order to partially assess the impact of common-methodvariance in this study, Ilarman'.s one-factor test (Harman, 1967) was employed. Theresults indicate four factors with eigenvalues greater than 1.0; the first factor accountsfor only 32% of the variance. (Consequently, it does nol appear that common-methodbias overwhelms the study's results. Although the data were collected cross-sectionally, il can be reasonably assumed that the degree of underemployment thatoccurred with job loss logically precedes tbe other variables in the study. Nonetheless,conclusions about causality must remain tentative until longitudinal data are collected.

The research here used a larger and mtjre heterogeneous sample than previousstudies in the area and investigated a population whose underemployment has notreceived empirical attention. However, in future research, it would be beneficialto simultaneously sample across different t}'pes of underemployed workers (e.g.underemployed new college graduates, underemployed blue-collar workers, anddownsized executives). Such designs would help untangle the purely economicaspects of underemployment from the psychological aspects of the phenomenonOahoda, 1982).

Because all the participants in the present study had received similar levels ofoutplacement support, the role of outplacement could not be fully investigated here.Nevertheless, it is possible that the quality of outplacement assistance may influencehow quickly laid-off managers find new jobs, how underemployed they become, andhow negatively they react to that underemployment.

Directions for future researchThere are several other important variahles to explore in future research on under-employment. Iwo other potential mediating processes which might be useful toexplore here are procedural justice and psychological contract violations (Maitin,1981; Rousseau, 1995). Layoffs may spark feelings of procedural injustice toward thejirevious employer, which in turn may lead to more negative job attitudes towards.subsequent employees. Similarly, as a result of job loss and underemployment,managers may fundamentally change the ways in which they view their psychologicalcontracts and consciously decide not to invest heavily in subsequent jobs.

Other dependent variahles which might he worthy of additional study are in-role andextra-role behaviour (cf. Feldman, 1996). Dnderemploynient may not stronglyinfluence individuals' future in-role behaviours becau.se of tangible punisbments forfailure to perform adequately in regularly assigned job duties. However, underemploy-ment may have a longer-term negative impact on organizational citizenship behavioursbecause the performance of these behaviours is more under the individual discretionof employee.s themselves.

Underemployment and relative deprivation 469

Another important avenue for future research is investigating 'referent others' andtheir role in the amount of relative deprivation downsized managers experience. AsMartin (1981) notes, individuals use multiple referents in assessing the fairness oftheir job situations. Thus, underemployed managers who compare themselves with'similar up' referents (laid-off managers who obtained better replacement jobs) mayexperience greater relative deprivation than underemployed managers who comparethemselves with 'similar down' referents (laid-off managers who still had no jobs at all).Relative deprivation theory may also prove a useful framework for understanding howlaid-off workers respond to corporate assistance programmes. For example, givingdownsized workers extensive severance pay and otitplacement support may be seenas organizational 'apologies' for the layoffs, thereby mitigating against etnployeesexperiencing very high levels of relative deprivation (Olson & Hafer, 1996).

In conclusion, then, the subsequent career trajectories of employees who havelost their jobs and become underemployed in replacement positions deserve greaterattention from researchers and managers alike. Even among downsized executiveswith significant financial resources, underemployment creates feelings of relativedeprivation which do not automatically disappear when they obtain new jobs. Thus,these results call for a closer examination of the quality of employment, rather than justthe fact of it, in future studies of career transitions and adjustment to new jobs.

References

Adams, J. S. (1976). Bqiiity iheory revisited: Comments and annotated bibliography. In L.Berkowitz (Ed.), Advances in experimental social psychology (Vol. 9, pp. 45-90). New York:Academic Press.

Baron, R. M., & Kenny, D. A. (1986). The moderator-mediator variable distinction in socialpsychological research: Conceptual, strategic, and statistical considerations. Journal ofPersonality and Social Psychology, 51, 1173-1182.

Blau, G. (1993). Further exploring the relationship between job search and voluntary individualturnover. Personnel Psychology, 46, 313-33(>-

Bolino, M. C, & Feldman, D. C. (2000). Increasing the skill utilization of expatriates. HumanResource Management, 39, 367-380.

Borgen. W. A., Amundson, N. E.. & Harder, H. G. (1988). The experience of underemployment.Journal of Einploytnent Counselling, 25, 149-159.

Burris, B. H. (1983). No room at the top: Underemployment and alienation in tbe corporation.New York: Praeger Press.

Buss, T., & Redbum, F. (1983). Shutdown at Yoimgstown: Public policy for massunemployment. Albany, NY: State University of New York Press.

Buunk, B. P., & Jans.sen, P. (1992). Relative deprivation, career issues, and mental health amongmen in midlife. Journal of Vocational Bebaiinr, 40, 338-350.

Clogg, C. C, Si. Shockey, J. W. (1984). Mismatch between occupation and schooling: Aprevalence measure, recent trends, and demograpliic analysis. Demograpby, 21, 235-257.

Clogg, C. c:., Sullivan, T., & Mutchler, J. (1986). Measuring underemployment and inequality inthe work force. Social Indicators Research. 18, 375-393.

Crosby, F. (1976). A model of egoistical relative deprivation. Psychological Reriew, 83, 85-113.Crosby, F. (1982). Relative deprivation and working women. New York: Oxford University

Press.Crosby, F. (1984). Relative deprivation in organizational settings. In B. W. Staw & L. L.

Cummings (Eds.), Research in organizational behai'ior (Vol. 6, pp. 51 -93). Greenwich, CT:JAl Press.

Eider, G. (1974). Children of the Great Depression: Social change in life experiences. Chicago:University of Chicago Press.

470 Daniel C Feldman et al.

Feather, N. T., & O'Brien, G. E. (1986). A longitudinal study of the effects of employment ajidunemployment on school-ltavcrs. Journal of Occupational Psychology, 59, 121-t44.

Feidinan, D. (;. (1996). The nature, antecedents, and consequences of underemployment.Journal of Management, 22, 385-409.

Feldman, D. C, & Doerpinghaus, H. I. (t992). Patterns of part-time employment. yoMrnrt/ ofVocational Behavior, 41, 282-294.

Feldman, D. C, Leana, C. R., & Tumley, W. H. (1997). A relative deprivation approach tounderstanding underemployment. In C. L. Cooper & D. M. Rousseau (Eds.), Trends inorganizational behavior (Vol. 4, pp. 43-60). New York: Wiley.

Feldman, D. C, & Tiimley, W. H. (1995). Underemployment among recent business collegegrdduate^. Journal of Organizational Befoavior, 16, 691-706.

Feldman, D. C, & Weitz, B. A. (1991). From the invisible hand to the gladhand: Understandingthe nature of a careerist orientation to work. Human Resource Management. 30, 237-257.

Gordus, J., Jarley, P., StFerman, L. (1981). Plant closings and economic dislocation. Kalamazoo,MI: W. E. Upjohn Institute for Employment Research.

Hackman, J. R., & Oldham, G. R. (1980). Work redesign. Reading, MA: Addison-WesJey.Hamian, H. H. (1967). Modem factor analysis. Chicago: University of Chicago Press.Humphrys, P., & O'Brien, G. E. (1986). The relationship between skill utilization, professional

orientation and job satisfaction for pharmacists. Journal of Occupational Psychology, 59,315-326.

Jahoda, M. (1982). Employment and unetnploymenl: A social psychological analysis.Cambridge, UK: (^imbridge University Press.

Kaufman, H. (1982). Professionals in search of work. New York: Wiley.Khan, L. J., & Morrow, P. C. (1991). Objective and subjective underemployment relationships to

job satisfaction./f)MJTi<?/ of Business Research, 22, 211-218.Kuhnert, K, (1989). The latent and manifest consequences of work. The Journal of Psychology,

123,417-427.Kuttner, R. (1994). Where have all the good jobs gone? Business Week, August 29, 16.Lawler, E. E. UI. (1973). Motivation in ivork organizations. Monterey, C;A: Brooks/Cole.Leana, C. R., & Feldman, D. C. (1992). Coping ivith job ioss: Hoiv individuals, organizations,

and communities respond to layoffs. New York: Macmillan/Lexington Books.Leana, C. R,, & Feldman, D. C. (1995). Finding new jobs after a plant closing: Antecedents

and outcomes of the occurrence and quality of reemployment. Human Relations, 48,1381-1401.

Liem. R,, & Liem, J. H. (1988). Psychological effeets of unemployment on workers and theirfamilies. Journal of Social Issues, 44, 87-105.

Locke, E. A. (1976). The nature and causes of job satisfaction. In M. D. Dunnette (Ed.),Handbook of industrial and organizational psychology (pp. 1297-1350). Chicago:Rand-McNaUy.

Martin, J. (19B1). Relative deprivation: A theory of distributive injustice for an era of shrinkingresources. In B. M. Staw & L. L. Cummings (Eds.), Research in organizational behavior(Vol. 3, pp. 53-107). Greenwich, CT: JAI Press.

Meyer, J. P., & Allen, N. J. (1984). Testing the 'side bet theory' of organizational commitment:Some methodological considerations. Journai of Applied Psychology, 6^, 372-378.

O Brien, G. E. (1986). Psychology of work and unemployment. New York: Wiley.O'Brien, G. E., & Feather, N. T. (1990). The relative effects of unemployment and quality of

employment on the affect, work values, and personal control of adolescents. Journal ofOccupational Psychology, 63, 151-165.

Olson, J., Si Hafer, C. L. (1996). Affect, motivation, and cognition in relative deprivationresearch. In R. M. Sorrentino & E. T. Higgins (Eds.), Handbook of motivation andcognition (Vol. 3, pp. 85-117). New York: Guilford Press.

Olson, J., Roese, N., Meen, J,, & Robertson, D. (1995). The preconditions and consequences ofrelative deprivation: Two field studies. Journal of Applied Social Psychology, 25, 944-964.

Underemployment and relative deprivation A71

Robinson, S. L. (1996). Trust and breach of the psycbological contract. Administrative ScienceQuarterly, 41, 574-599.

Robinson, S. L., Kraatz, M. S., & Rousseau, D. M. (1994). Changing obhgations and tbe psycbo-logical contract: A longitudinal study. Academy of Management Journal, M, 137-152.

Rousseau, IJ. M. (1990). New hire perceptions of tbeir own and their employer's obligations:A study of psychological contncts. Journal of Organizaiional Beharior, 11, 389-400.

Rousseau, D. M. (1995). Psycbological contracts in organizations: Understanding written andunwritten agreements. Thousand Oaks, (;A: Sage.

Rousseau, D. M., Sitkin, S. B., Burt, R. S., & Camerer, C. (1998). Not so diilcrcnt after all:A cross-discipline view of trust. Academy of Management Review, 23, 393-404.

Stouffer, S. A., Suchman, E. A., DeVinney, L. C, Star, S. A., & Williams, R. M. (1949). TheAmerican soldier: Adjustments during artny life (Vol. 1). Princeton, NJ: PrincetonUniversity Press.

Sweeney, P. D., McFarlin, D. B., & Inderrieden, E. J. (1990). Using relative deprivation theory toexplain satisfaction witb income and pay level: A multistudy examination. Academy ofManagement Journal, 33, 423-436.

Taylor, J. C, & Bowers, D. G. (1972). Survey of organizations: A machine scored standardizedquestionnaire instrument. Ann Arbor, MI: Institute for Social Research, University ofMichigan.

Winefield, A., & Tiggemann, M. (1990). Rmploymcnt status and psycbological well-being: Alongitudinal study. Journal of Applied Psychology, 7.5, 455-459.

Winefield, A., Winefield, H., Tiggemann, M., & Goldney, R. (1991). A longitudinal study of thepsychological effects of unemployment and unsatisfactory employment on young adults.Journal of Applied Psychology, 76. 424-431.

Zvonkovic, A. M. (1988). Underemployment: Individual and marital adjustment to income loss.Lifestyles: Family and Economic Issues, 9, 161-178.

Received 8 tAay 2001; revised version received 25 March 2002