Вы находитесь на странице: 1из 21

APJHR 46_2_Lawton.

qxd 11/06/2008 10:33 PM Page 220

220 Asia Pacific Journal of Human Resources 2008 46(2)

Examining determinants of employee benefit preferences:


Joint effects of personality, work values, and demographics

Katie E. Lawton
University of Canterbury, New Zealand

Oleksandr S. Chernyshenko
Nanyang Technological University, Singapore

This investigation focused on identifying the determinants of employee benefit


preferences in a large New Zealand retail organisation (N = 229). The study
delved beyond previous research by examining the role of individual difference
variables in explaining employee benefit preferences in addition to commonly
studied demographic variables. Specifically, it was hypothesised that work
values and the five-factor model traits of personality would explain incremental
variance in employee benefit preferences, beyond the effects of demographics.
This hypothesis was supported for education-supportive benefit preferences,
which were predicted by status and independence work values, extroversion,
conscientiousness, age, education, and job category. However, family-supportive
benefit preferences appeared to be driven solely by demographics (marital status,
number of dependants and employment status). The implications of these findings
and possible applied uses are discussed.

Keywords: employee preferences, flexible benefit plans, personality, work values

In the labour market of today, organisations are in constant competition to


attract and retain employees. The unemployment rate in New Zealand (NZ)
is currently at a record low, at 3.8 per cent (Statistics NZ 2007); thus vacant
positions must be attractive enough to ideal candidates to initiate thoughts of
leaving their current job. The make-up of the compensation package can play
a major role in making one position more desirable than another. Therefore,

Correspondence to: Associate Professor Oleksandr Chernyshenko, S3-B2B-58, Nanyang


Business School, Nanyang Technological University, Nanyang Avenue, Singapore 639798;
fax: +65 6792 4217:e-mail: chernyshenko@ntu.edu.sg
Asia Pacific Journal of Human Resources. Published by SAGE Publications (Los Angeles, London, New Delhi and
Singapore; www.sagepublications.com) on behalf of the Australian Human Resources Institute. Copyright © 2008
Australian Human Resources Institute. Volume 46(2): 220–240. [1038-4111] DOI: 10.1177/1038411108091759.
APJHR 46_2_Lawton.qxd 11/06/2008 10:33 PM Page 221

Determinants of employee benefit preferences 221

if the package is designed in a manner that meets the needs and desires of
employees, organisations can gain a competitive advantage by attracting and
retaining labour of a high quality. Moreover, even within a single organisa-
tion, compensation packages may need to vary in order to the meet individual
employee needs and desires (Lawler 1990).
Traditionally, organisations have offered individuals little choice in terms
of compensation; employees were told their base pay and the benefits they
would receive (Lawler 1990). Research has shown that such conventional, fixed
benefits packages are often cost-inefficient as they are valued less by employees
than their actual cost to the organisation (Nealey 1964). Another downfall of
fixed benefit plans is that they are usually targeted to the needs of a ‘nuclear’
family, which has been becoming less common over the last few decades
(Barber, Dunham and Formisano 1992). Moreover, the composition of the
labour force has also changed dramatically, resulting in a wider range
of demands for employee benefits. For example, single parents may have
increased needs for childcare and parental leave, or conversely, families
with two working parents may end up with certain benefits duplicated
(Meisenheimer and Wiatroeski 1989).
A solution to meeting this vast range of benefit needs is to give employees
the option of choosing the benefits they want (Lawler 1990). Such flexible
benefits plans (FBPs) allow employees to tailor their benefits package to meet
their own specific needs (Meisenheimer and Wiatroeski 1989). This type of
plan is beneficial to both the employees and the organisation. The employees
get to select the benefits they value the most and forgo benefits of lesser impor-
tance, without increasing benefits expenses for the organisation. Other advan-
tageous organisational outcomes of FBPs include increased employee benefit
satisfaction and job satisfaction (Barber, Dunham and Formisano 1992) as
well as enhanced applicant attraction (Cable and Judge 1994), organisational
commitment (Meyer, Allen and Smith 1993), and culture (Lawler 1990).

Rationale for current study

In recent times, employee benefits have become a defining feature of most


compensation packages of New Zealand’s employers of choice. However,
there is a distinct lack of research in the area of employee benefit preferences,
particularly using New Zealand samples. It is well documented that different
benefits are desirable to employees at certain stages in their lives and careers
(Haslinger and Sheerin 1994) and it has been suggested that numerous
personal dimensions affect the attractiveness of benefits options (Barringer and
Milkovich 1996). However, past research has primarily focused on socio-demo-
graphic factors as antecedents of employee benefit preferences with little or no
consideration given to other areas in which individuals may differ (e.g. person-
ality, interests, and values). The major aim of this study, therefore, is to fill this
APJHR 46_2_Lawton.qxd 11/06/2008 10:33 PM Page 222

222 Asia Pacific Journal of Human Resources 2008 46(2)

gap and examine whether the personality traits of extroversion, agreeableness,


neuroticism, conscientiousness and openness (a.k.a. ‘Big Five’) in conjunction
with work values explain any incremental variance in employee benefit pref-
erences, beyond the effect of demographic variables. It is hoped that results
would be of interest not only to Asia Pacific audiences, but also generalise to
other markets.

Benefit preference grouping in this investigation

The current research was conducted in conjunction with an external organ-


isation to examine predictors of employee benefit preferences in a real-world
setting. A focus group was carried out with employees of the company to
discover what types of benefits would potentially be valued by different
employee groups, and thus should be included in the study. Based on the
results of this preliminary review, an array of 14 different benefits was
generated (the list of the benefits can be found in table 1). They covered a wide
range in order to anticipate the needs of the organisation’s diverse workforce.
It was expected that benefits with a common theme would be highly corre-
lated and, hence, form natural benefit preference groupings. For example, it is
probable that if an employee rates preference for childcare benefits highly, they
will also prefer similar benefits such as extra parental leave.
On this premise, it was envisaged that the benefits on offer would fall into
four distinct groups: flexible-lifestyle, family-friendly, training and develop-
ment, and financial benefits. It is proposed that flexible-lifestyle benefits will
be comprised of benefits to do with extra vacation time and a relaxed work
schedule, such as extra annual leave and flexible work hours. Family-friendly
benefits will consist of benefits to do with looking after and supporting
a family, such as childcare, extra parental/family leave, medical and life
insurance. Training and development benefits will contain benefits related to
current or future study needs, such as paid leave to attend courses and funded
study if relevant to current career path, and paid study leave for purposes
outside of current role. Finally, financial benefits will be made up of benefits
options that could help to improve one’s personal finance situation. Studying
benefit groupings rather than individual benefits allowed this study’s findings
to be generalised across different organisational settings, where specific benefits
offered may be different but the underlying benefit groupings will remain
consistent.

Demographics as a predictor of benefit preference

Past research has shown that the needs and desires of employees for different
benefits options are largely dictated by their personal demographic character-
APJHR 46_2_Lawton.qxd 11/06/2008 10:33 PM Page 223

Determinants of employee benefit preferences 223

istics. Older workers are reported to prefer time off, schedule changes and
assistance in skill development over cash compensation (Hedge, Borman and
Lammlein 2006). Therefore, it is predicted that age will be related to prefer-
ences for flexible-lifestyle and training and development benefits. In addition,
White (1983) concluded that employees lower in the job hierarchy assigned
more value to education and training opportunities. It is expected that a similar
relationship will be found in the current study, with job category being related
to preference for training and development benefits.
The Metlife Incorporated study of employee benefits trends (2006)
reported differences in the job attributes valued by employees based on their
life stage. ‘Opportunity for financial advancement’ and ‘work–life balance’
emerged as being the most valued job traits for single employees. Therefore,
it is proposed that individuals without a partner will prefer financial and
flexible-lifestyle benefits. In contrast, individuals with young families rated
‘opportunities for skill building and professional growth’ as being of greatest
value to them. Thus, individuals who are living with a partner are more likely
to value training and development benefits. Moreover, age of youngest child
has been shown to predict perceived importance of family-supportive
programs (Frone and Yardley 1996). Thus, it is proposed that number of
dependants will relate to preferences for family-friendly benefits.

Individual difference predictors of benefit preference

Work values
It has been suggested that values are likely to influence the relative personal
attractiveness of a reward system (Furnham 1992), making it probable that
work values will affect benefit preferences. In this study, work values are
conceptualised as the importance placed on various job dimensions. Specific
relationships were hypothesised based on the content of each work values
factor being measured. ‘Comfort and security’ work values refer to the impor-
tance placed on job characteristics such as comfortable working conditions and
job security. This desire to comfortably support a family’s general well-being
and financial security is likely to be related to preferences for family-friendly
and financial benefits. Comparable findings are also expected for flexible-
lifestyle benefits, as this work values factor also asks about the importance of
ample leisure time off the job.
‘Competence and growth’ work values items include the importance of
job characteristics such as intellectual stimulation, continued development of
skills, and a feeling of accomplishment. Such desire for continued education
and learning is likely to result in preferences for training and development
benefits. Similarly, ‘status and independence’ work values items ask about the
importance of job features such as the opportunity to work independently,
advance to high responsibility and take risks. Individuals with these work
APJHR 46_2_Lawton.qxd 11/06/2008 10:33 PM Page 224

224 Asia Pacific Journal of Human Resources 2008 46(2)

values are also more likely to prefer training and development benefits as they
are beneficial to career progression.

Personality
Trait-dependent preferences are another factor that has been suggested to
affect the personal attractiveness of compensation packages (Furnham 1992).
Family-friendly benefits items encompass benefits designed for an employee to
be able to take care of their family and aid their physical well-being. Neurotic
individuals may prefer this benefit group, as they tend to be nervous, anxious,
and have an inability to deal with stress (Goldberg 1992). Therefore the
insurance aspects of this benefits group may be appealing as they allow indi-
viduals to feel more at ease about the future. Conscientiousness is also likely to
be related to this group of benefits. The industriousness, responsibility and
virtue facets of conscientiousness (Roberts et al. 2005) may reflect a conven-
tional desire and duty to support a family and make sure that their needs are
sufficiently met. Furthermore, agreeableness is likely to be related to family-
friendly benefit preferences, as highly agreeable individuals possess a selfless
concern for others (Costa and McCrae 2003), and are likely to put their family’s
needs before their own.
The financial benefits group is comprised of benefits designed to aid an
employee’s financial well-being. It is anticipated that neuroticism and consci-
entiousness will be positively related to preferences for this benefits group.
Neurotic individuals may worry about supporting themselves or others finan-
cially in the future. Undertaking personal finance training or participating in
a savings scheme may help to ease this concern. Individuals who are consci-
entious, however, are likely to make a committed effort to saving or paying
off debt. Choosing benefits such as personal finance training or a high interest
savings account would help achieve these goals.
Training and development benefits comprise options that would enable
employees to enhance their professional development and career, through
education. Previous research has shown that the personality traits of conscien-
tiousness, openness and extroversion are positively related to motivation to
learn (Major, Turner and Fletcher 2006), a desire likely to be reflected in pref-
erence for training and development benefits. Similar results are expected for
the current study, with conscientiousness, openness and extroversion being
related to preferences for training and development benefits.
Finally, the flexible-lifestyle benefits group contains leave-taking items,
allowing employees to have more free time. Extroverted individuals enjoy
lively activity and social interaction (Costa and McCrae 2003), and are more
likely to have a hectic social schedule. The option of flexible working hours
would allow these social activities to be easily integrated into a working week.
Furthermore, openness to experience refers to an individual’s receptiveness to
new ideas, approaches and experiences (Costa and McCrae 2003). Open indi-
APJHR 46_2_Lawton.qxd 11/06/2008 10:33 PM Page 225

Determinants of employee benefit preferences 225

viduals are likely to choose flexible-lifestyle benefits as they are more likely to
want to try new innovations such as flexible work hours.

Study hypotheses
In light of the above discussion, specific relationships were anticipated between
the predictor variables, demographics, FFM personality factors, and work
values, and the dependent variable, benefit preference. These relationships are
expressed in the following hypotheses:

Hypothesis 1: Flexible-lifestyle benefit preferences will be predicted by age,


single marital status, comfort and security work values, openness, and
extroversion.

Hypothesis 2: Family-friendly benefit preferences will be predicted by


number of dependants, comfort and security work values, neuroticism,
conscientiousness, and agreeableness.

Hypothesis 3: Training and development benefit preferences will be


predicted by age, de facto/married marital status, job category, competence
and growth, and status and independence work values, openness,
conscientiousness, and extroversion.

Hypothesis 4: Financial benefit preferences will be predicted by single


marital status, comfort and security work values, neuroticism, and
conscientiousness.

Method

Sample

Employees from a large New Zealand retail organisation made up the sample
for this study. Employees across the entire organisation were invited to take
part, from shop-floor workers to executive management. Questionnaires were
distributed accompanied by a cover sheet explaining the nature of the study
and assuring anonymity and confidentiality. The task took between 10 and 15
minutes. Participation was encouraged by the organisation but was not
mandatory, and no incentive was provided. Nevertheless, 237 of 961 employees
of the company decided to participate in the study. Of these, 229 provided
useable data after screening for careless responses, resulting in a response
rate of 25 per cent. Specific demographic characteristics of participants are
summarised below.
APJHR 46_2_Lawton.qxd 11/06/2008 10:33 PM Page 226

226 Asia Pacific Journal of Human Resources 2008 46(2)

Measures

Demographics

Demographic variables were assessed to investigate their contribution to


predicting benefit preferences. Specific items administered include sex, age,
marital status, number of dependants, education level, household income, job
category (higher scores indicated jobs with more supervisory responsibilities),
and employment status (0 = full-time, and 1 = part-time). The response format
consisted of a range of options for each question, and participants ticked the
box which best represented their own personal characteristics. A wide cross-
section of the company’s workforce participated in the study. The following is
a summary of the demographic characteristics of participants: 104 participants
were male, 125 female, with a mean age bracket of 30–39 years. A range of
family situations were reflected, with 49% of respondents being single, 51%
lived with a partner, and the mean number of children was 0.7. Furthermore,
a wide array of employment situations was also represented. The average level
of education was 7th form, which in the NZ school system is equivalent to the
last year of secondary school or year 12–13. Among respondents, 63% worked
full-time, 37% part-time; 73% of respondents held basic level jobs and 27%
were supervisors or team leaders.

Benefit preferences
The 14 benefits described earlier formed the basis of the initial section of the
questionnaire. Each benefit was accompanied by a brief description or example
in order to clarify what would be on offer. Examples of such items include
‘Study leave: Paid time per week for educational purposes outside of current
role or career path, e.g. to attend lectures’ and ‘Parental/Family Leave: 1 week
extra paid time off for caring for dependent children or elderly parents, etc.’.
Each benefit option was accompanied by the question ‘Would you choose this
benefit if it was available?’, and participants rated the likelihood that they
would choose the benefit on a 5-point Likert scale, ranging from 1 = ‘definitely
not’ to 5 = ‘definitely’. The final item in this section asked participants about
the likelihood that they would choose to participate in such a program in
exchange for a small fee. This enabled benefit preference responses to be
obtained from all participants, even if some indicated that they would opt out
of a flexible benefits program, if implemented.
The exploratory factor analysis of the responses to these benefit-prefer-
ence items (see results section for more detail) showed that they should be
grouped into two, rather than the four factors originally anticipated. Upon
inspection it was evident that benefit preferences that were originally labelled
as ‘family-friendly’ and ‘financial’ collapsed into a single factor. These included
childcare and insurance type benefits, as well as benefits related to saving, all
APJHR 46_2_Lawton.qxd 11/06/2008 10:33 PM Page 227

Determinants of employee benefit preferences 227

of which are related to supporting a family. Thus, the first factor was named
‘family-supportive’. The second factor essentially included ‘training and devel-
opment’ benefits, as well as flexible hours and debt payments. It seemed that
all of these benefits were useful for an employee undertaking study, because
flexible hours allow classes to be scheduled into a working week and debt
payments would be beneficial to an employee with a student loan. Therefore,
the second factor was labelled ‘education-supportive’.
Because the two-factor rather than the anticipated four-factor benefit
structure had emerged, it was necessary to rework our initial hypotheses to
match the new factors. Our initial hypotheses 2 and 4 were combined to relate
to the first factor, family-supportive benefits, and hypotheses 1 and 3 were
combined to relate to the second factor, education-supportive. Any conflicting
predictor variables were evaluated in a logical manner, and amended to fit the
new benefits factor structure.

Revised hypothesis 1: Family-supportive benefit preference will be predicted


by de facto/married marital status, number of dependants, comfort and
security work values, neuroticism, conscientiousness, and agreeableness.

Revised hypothesis 2: Education-supportive benefit preference will be


predicted by age, de facto/married marital status, job category, competence
and growth, and status and independence work values, openness,
conscientiousness, and extroversion.

Work values
Employee work values were assessed using the Work Values Inventory
(Manhardt 1972). This measure evaluated the importance of 21 different job
characteristics on a 5-point Likert scale, where 1 = ‘unimportant’ and 5 = ‘very
important’. These characteristics have been shown to group into three dimen-
sions: comfort and security, competence and growth, and status and inde-
pendence. Meyer, Irving and Allen (1998) assessed the scale’s reliability:
coefficient alpha values ranged from 0.63 to 0.72 for comfort and security, 0.65
to 0.80 for competence and growth, and 0.62 to 0.68 for status and independ-
ence. The current study produced similar findings, with coefficient alphas of
0.59 for comfort and security values, 0.76 for competence and growth values,
and 0.73 for status and independence values.

Personality
The Big Five personality traits, also known as the Five-Factor Model (FFM),
were measured using a 50-item personality measure from the International
Personality Item Pool (Goldberg 1999). Participants made judgements about
how accurately the statements described themselves on a 5-point Likert scale,
APJHR 46_2_Lawton.qxd 11/06/2008 10:33 PM Page 228

228 Asia Pacific Journal of Human Resources 2008 46(2)

ranging from 1 = ‘very inaccurate’ to 5 = ‘very accurate’. Goldberg (1999)


reported the coefficient alpha for each of the five subscales (10 items each) as
ranging from 0.77 to 0.86, indicating an acceptable level of internal consistency.
The data from the current study support this, with alphas for the scales of
extroversion, agreeableness, conscientiousness, neuroticism, and openness of
0.84, 0.77, 0.78, 0.83 and 0.75 respectively. Further details about psychometric
properties of the 50-item Goldberg’s measure can be found in Guenole and
Chernyshenko (2005).

Results and discussion


Benefit preference statistics

Results presented in table 1 include a summary of the distribution of responses


for each benefit on offer, with the percentage of responses in each category
illustrated, as well as the means and standard deviations of each benefit option.
As can be seen in table 1, many of the benefits items appear to be negatively
skewed, with the majority of responses between 3 = ‘maybe’ and 5 = ‘defin-
itely’. Such items include: course leave, extra annual leave, saving for the

Table 1 Distribution of responses and descriptive statistics for each benefit option

Response option (%)

1 2 3 4 5 Descriptives
Benefit Definitely Probably
option not not Maybe Probably Definitely Mean SD

Childcare 41 12 17 12 18 2.52 1.55


Savings account 11 15 31 21 23 3.30 1.27
Course leave 3 3 14 31 48 4.17 1.03
Extra annual leave 0 1 3 9 87 4.81 0.57
Life insurance 10 14 29 22 25 3.37 1.28
Debt payments 18 12 29 15 27 3.21 1.42
Saving for the future 4 5 21 24 45 4.00 1.13
Medical insurance subsidy 7 10 23 25 36 3.73 1.23
Full medical insurance 5 6 16 25 48 4.04 1.17
Funded study 3 5 20 28 44 4.07 1.04
Flexible hours 4 7 12 31 46 4.07 1.12
Parental/ family leave 17 6 16 19 42 3.65 1.48
Personal finance training 12 16 25 26 21 3.28 1.29
Study leave 7 11 20 17 44 3.79 1.31
APJHR 46_2_Lawton.qxd 11/06/2008 10:33 PM Page 229

Determinants of employee benefit preferences 229

future, full medical insurance, funded study, flexible hours, parental/family


leave, and study leave. These benefits all have high means, and over 40 per
cent of respondents rating ‘5’, they would definitely choose the benefit if it was
available. However, all but one of the 14 benefits had high standard deviations,
indicating considerable variability among employees. The ‘extra annual leave’
item was discarded at this stage of the analysis as it had extremely low variance
(SD = 0.57); 96 per cent of respondents rated that they would probably or
definitely choose the benefit if it was available. This benefit option did not
contain enough variation in responses to be useful in remaining analyses.
Overall, the likelihood that each of the employee benefits would be chosen
was generally high. This is not surprising for traditional-type benefits designed
to protect an employee’s financial and personal well-being. Hospital insurance
benefits have been cited as the most highly preferred by employees, especially
in the Unites States (Nealey 1964). The desire for such benefits seems to be
universal, with the need to support a family’s financial security and health tran-
scending time and place, as medical and life insurance preferences were rated
highly in the current study. Nealey’s (1964) research found that a shorter
working week was the least preferred employee benefit. However, the current
findings show that flexible work hours are preferred by a large proportion of
employees. This contrast may illustrate the changing needs and desires of the
workforce of today compared to that of a few decades ago.
The other non-traditional benefits were also highly preferred, cementing
the proposition that modern employees have a large range of benefits needs
which would not be met by a traditional benefits program. For instance, a large
proportion of the workforce rated that they would definitely choose the
benefits extra annual leave and funded study, if available. These preferences
illustrate the desire for work–life balance and opportunities for skill building
and professional growth, two of the top five job attributes valued by workers
of today (Metlife Incorporated 2006).

Benefit structure
An exploratory factor analysis was conducted on the remaining benefit pref-
erence items in order to summarise the underlying patterns of correlations and
thus group closely related benefits. The extraction method used was Principal
Axis Factoring with Direct Oblimin rotation, allowing factors to be correlated.
The Kaiser criterion was used for the initial stage of the analysis, and two
factors emerged with eigenvalues greater than 1. The scree plot, illustrated in
figure 1, showed a solution consistent with this interpretation. The analysis
was then rerun, specifying a two-factor solution.
The factor loadings in the pattern matrix were then examined to
determine which items were defining parts of factors and should be used in
the remainder of the analysis. These loadings are reported in table 2. The
criteria for inclusion were a factor loading of 0.30 or higher and loading only
APJHR 46_2_Lawton.qxd 11/06/2008 10:33 PM Page 230

230 Asia Pacific Journal of Human Resources 2008 46(2)

Figure 1 Scree plot for EFA of benefit preference ratings

3
Eigenvalue

1 2 3 4 5 6 7 8 9 10 11 12 13
Factor number

on a single factor. On this basis, the benefits option ‘personal finance training’
was discarded as it cross-loaded on factors 1 and 2, with loadings of 0.32 and
0.46 respectively. All of the other benefits items were retained as they loaded
cleanly onto a single factor with a factor loading of over 0.30.
On the basis of the findings from the exploratory factor analysis, two
composite benefits variables were formed by combining scores of the indi-
vidual benefits items which loaded onto each factor. That is, full and subsidised
medical insurance, life insurance, saving for the future, childcare, parental
/family leave, and savings account were combined to form the composite
variable, family-supportive benefits. On the other hand, item scores on the
study leave, funded study, course leave, flexible hours, and debt payments
benefits were combined to form the education-supportive benefits composite.
Such aggregation of individual benefit preferences was advantageous from
a statistical point of view, because it is known to decrease the amount of
random error variance in the preference score and, thus, improve predictability
(see Rushton, Brainerd and Pressley 1983). Both composite variables had
reasonably high levels of internal consistency reliability, with α = 0.76 for
family-supportive benefits, and α = 0.66 for education-supportive benefits.
Importantly, the two composites were only moderately correlated (r = 0.25,
APJHR 46_2_Lawton.qxd 11/06/2008 10:33 PM Page 231

Determinants of employee benefit preferences 231

Table 2 Factor loadings for benefits factors

Factor 1 Factor 2 Factor 1 Factor 2

Family- Education- Family- Education-


Item supportive supportive Item supportive supportive

Medical insurance Personal finance


subsidy .92 –.11 training .46 .32
Full medical insurance .80 –.17 Study leave –.11 .64
Life insurance .64 .02 Funded study .20 .62
Saving for the future .42 .00 Course leave .19 .52
Childcare .40 .10 Flexible hours –.06 .46
Parental/family leave .39 .18 Debt payments –.02 .44
Savings account .37 .23
Note: Items which load on a single factor with a loading of .30 or greater are shown in bold.

p < 0.05) indicating that the two composites were not part of some overall pref-
erence for benefits.
Relationship between demographics and benefit preference
Results reported in table 3 include intercorrelations between demographic
variables and the two composite benefits variables: family-supportive and
education-supportive.

Table 3 Intercorrelations among demographic variables and benefit preference composites

Intercorrelations

N 1 2 3 4 5 6 7 8 9

1. Sex 229 –
2. Age 227 –.06 –
3. Number of
dependants 227 .03 .34* –
4. Education 222 .10 .03 –.06 –
5. Job category 228 .05 .25* .06 .15* –
6. Employment
status 229 –.04 –.34* –.08 –.08 –.32* –
7. Marital status 208 –.16* .56* .40* .13 .21* –.24* –
8. Family-
supportive 224 –.01 .32* .36* .01 .07 –.23* .39* –
9. Education-
supportive 224 –.12 –.28* –.13 .11 –.18* .13 –.07 .25* –
* Significant at the p < .05 level.
APJHR 46_2_Lawton.qxd 11/06/2008 10:33 PM Page 232

232 Asia Pacific Journal of Human Resources 2008 46(2)

It was hypothesised that marital status and number of dependants would


be related to family-supportive benefit preferences. In addition, there were
significant relationships with age and employment status, which is explained
by the fact that these two demographic variables positively correlate with the
number of dependants (i.e. people with more dependants tend to be older and
hold full-time jobs).
Education-supportive benefit preference had significant negative rel-
ationships with age and job category. As expected, young employees with
simpler jobs were more likely to desire education-supportive benefits than
their older counterparts.

Relationship between individual difference variables and benefit preference


Results presented in table 4 include means, standard deviations and correl-
ations between each of the three work values factors: comfort and security,
competence and growth, and status and independence, and the two composite
benefits variables: family-supportive and education-supportive.
It was hypothesised that preferences for education-supportive benefits
would be related to competence and growth values and status and independ-
ence values. These relationships were supported by the data. However,
contrary to expectations, comfort and security work values were not signifi-
cantly related to preferences for family-supportive benefits. Other significant
relationships that emerged included significant positive correlations between
each of the work values factors. These relationships ranged from 0.48 to 0.62
in magnitude, suggesting that they may not be completely conceptually distinct
from one another.
Correlations between the FFM personality traits and the two composite
benefits variables are presented in table 5, as well as their means and standard
deviations. It was predicted that family-supportive benefit preference would be

Table 4 Means, standard deviations and intercorrelations among work values factors
and benefit preference composites

Intercorrelations

M SD 1 2 3 4 5

1. Comfort and security 21.09 2.75 –


2. Competence and growth 24.02 3.43 .53* –
3. Status and independence 27.90 4.10 .48* .62* –
4. Family-supportive 24.59 5.83 .12 .20* .21* –
5. Education-supportive 24.09 4.04 –.05 .17* .20* .25* –
* Significant at the p < .05 level.
APJHR 46_2_Lawton.qxd 11/06/2008 10:33 PM Page 233

Determinants of employee benefit preferences 233

Table 5 Means, standard deviations and intercorrelations among FFM traits and benefit
preference composites

Intercorrelations

M SD 1 2 3 4 5 6 7

1. Extroversion 33.70 6.80 –


2. Agreeableness 39.42 5.11 .23* –
3. Conscientiousness 38.04 5.37 .09 .32* –
4. Neuroticism 24.91 6.41 –.19* –.23* –.20* –
5. Openness 36.40 5.42 .32* .14* .10 –.01 –
6. Family-supportive 24.59 5.83 .05 .16* .08 –.14* –.13 –
7. Education-supportive 24.09 4.04 .22* .11 –.15* .01 .23* .25* –
* Significant at the p < .05 level.

positively related to neuroticism, conscientiousness and agreeableness. However,


only the relationship with agreeableness was supported by data. Furthermore,
a negative relationship was found between neuroticism and family-supportive
benefit preference; however, it was in the opposite direction to what was antici-
pated. One could make an argument that more neurotic individuals are likely
to have already taken care of their insurance and childcare needs, and thus no
longer desire family-supportive benefits, but more research is clearly needed to
address the observed contradiction. The relationship between conscientiousness
and family-supportive benefit preference was non-significant.
Education-supportive benefit preference was found to be significantly
related to extroversion, conscientiousness and openness. However, the rel-
ationship between conscientiousness and education-supportive benefit prefer-
ence was in the opposite direction to what was expected. While one may expect
conscientious individuals to desire more educational benefits, in fact the
opposite appears to be true.

Predicting benefit choices

Predictors of family-supportive benefits


An initial simple multiple regression was run including all of the measured
work values, FFM traits and demographic characteristics as independent
variables to investigate their contribution to predicting family-supportive
benefit preferences, when calculated simultaneously. It was hypothesised that
family-supportive benefit preferences would be predicted by comfort and
security work values, marital status, number of dependants, neuroticism,
conscientiousness and openness. However, the only variables that emerged as
significant predictors were the demographic characteristics of number of
APJHR 46_2_Lawton.qxd 11/06/2008 10:33 PM Page 234

234 Asia Pacific Journal of Human Resources 2008 46(2)

dependants, marital status and employment status; neither work values nor
personality variables had significant effects on preference for family-supportive
benefits once demographics were taken into account.
The three significant demographic variables (marital status, number of
dependants and employment status) accounted for just under one-quarter of
the variance in family-supportive benefit preference, R2 = 0. 23, F(6, 215) =
10.8, p < 0.05. Marital status accounted for 3% of the variance, FΔ(3, 215) =
3.14, p < 0.05, with the married/de facto subgroup preferring family-supportive
benefits more than the single reference group, β = 0.25. Employment status
accounted for 4% of the variance, FΔ (2, 215) = 5.03, p < 0.05, with part-time
employees choosing family-supportive benefits less frequently than those who
worked full-time, β = –0.20. Finally, the number of dependants explained 6%
of the variance, FΔ(1, 215) = 16.6, p < 0.05, as the greater number of depen-
dants an employee had, the more likely they were to choose family-supportive
benefits, β = 0.27.
The relationships of marital status and number of dependants with
family-supportive benefit preferences were consistent with our hypotheses.
They also align with the findings of Nealey (1964), who reported that marital
status and number of dependants dramatically increased employees’ prefer-
ences for hospital insurance, citing family responsibility as a preference deter-
minant. Similar reasoning can explain the results from the current study, with
the duty to support a family driving preference for family-supportive benefits,
designed to take care of a family’s well-being. Employment type was an addi-
tional significant predictor found for family-supportive benefit preferences.
This could be attributed to the desire for the childcare and parental leave
features of this benefit group, as employees who work full-time have less time
available to tend to childcare needs. Together, these findings suggest that pref-
erences for family-supportive benefits are dictated more heavily by necessity,
due to demographic characteristics than by personal individual difference
variables.

Predictors of education-supportive benefits


It was hypothesised that education-supportive benefit preferences would be
predicted by competence and growth, and status and independence work
values, age, marital status, job category, extroversion, openness, and conscien-
tiousness. An initial simple multiple regression was run including all of the
measured work values, FFM traits and demographic characteristics as inde-
pendent variables to investigate their contribution to predicting education-
supportive benefit preferences, when calculated simultaneously.
Significant predictors included variables from each of the groups of inde-
pendents, that is, work values, FFM traits and demographic characteristics;
specifically, status and independence work values, extroversion, conscien-
tiousness, age, job category and education. These six independent variables
APJHR 46_2_Lawton.qxd 11/06/2008 10:33 PM Page 235

Determinants of employee benefit preferences 235

were significant at the p < 0.05 level, and accounted for just over one-quarter
of the variance in education-supportive benefit preferences, R2 = 0. 26, F(8,
209) = 8.9, p < 0.05.
Consistent with the hypothesis, and the findings of White (1983), job
category was a negative predictor of education-supportive benefit preference,
β = –0.21. This can be attributed to lower level employees feeling they had a
greater chance of advancement with additional education. However, education
level was found to be a positive predictor of education-supportive benefit pref-
erence, β = 0.15, because, in New Zealand, more education means more
student loans. Age was also negatively related to education-supportive benefit
preference, β = –0.24. A plausible explanation for this finding is that younger
employees put more value on opportunities for education as they have their
whole careers ahead of them, whereas older employees are unlikely to reap the
full benefits of additional education in terms of their career advancement, and
thus do not value it at as high a level. Note, however, this finding conflicts with
what was found by Hedge, Borman and Lammlein (2006) who suggested that
older workers valued assistance in skill development.
As expected, individuals with higher levels of status and independence
work values showed higher preferences for education-supportive benefits, β
= 0.34. This work values factor pertains to the importance placed on the career
progression and advancement aspects of a job. Therefore, it follows that indi-
viduals with such values would choose employee benefits that allow them the
opportunity to increase their skills and personal development. Competence
and growth work values did not emerge as a significant predictor of education-
supportive benefit preference.
The results suggest that an individual’s personality (extroversion) is also
important in predicting education-supportive benefits, β = 0.17. This was
consistent with the hypothesis and the findings of Major, Turner and Fletcher
(2006), who reported that extroversion was positively related to motivation to
learn. Extroverts require social contact, attention and fun (Costa and McCrae
2003), needs which may be met through learning and education. However,
openness did not significantly predict education-supportive benefit preference,
as expected. This finding was inconsistent with that of Major, Turner and
Fletcher (2006).
Contrary to what was predicted, conscientiousness had a negative rel-
ationship with education-supportive benefit preference, β = –0.14. The
proposed rationale centred on the idea that individuals high on conscientious-
ness would be industrious and achievement-driven, making them more
inclined to choose education-supportive benefits for career amelioration
purposes. Furthermore, conscientiousness is positively associated with moti-
vation to learn, suggesting that conscientious individuals would choose to
undertake opportunities for knowledge and education (Major, Turner and
Fletcher 2006). However, there are a number of possible explanations for this
finding. For example, highly industrious, task-oriented employees may see
APJHR 46_2_Lawton.qxd 11/06/2008 10:33 PM Page 236

236 Asia Pacific Journal of Human Resources 2008 46(2)

paid leave to attend courses as an unnecessary distraction from their work,


when their main interest is the job at hand. Furthermore, such individuals may
not value educational benefits as they do not require them. Conscientious indi-
viduals are likely to have already taken control of their educational needs and
career plan, leading to lower desire for education-supportive benefits.
Another possibility raised by reviewers of this paper is the presence
moderators. It may be the case that people with management level responsi-
bilities (who already have some education) may show a different strength of
association between conscientiousness and education benefits than at the
those at lower levels who have fewer responsibilities and perhaps are not
as distracted by their jobs. Hence, there may be an interaction between job
category and conscientiousness when it comes to evaluating education-
supportive benefits. To investigate this, we conducted a moderated regression
analysis, but found interaction to be insignificant. Finding significant moder-
ator effects, however, is notoriously difficult due to the general lack of power
(see Aguinis et al. 2005). In our study, job category samples may be just too
small to have the power to detect moderation. Inspection of simple bi-variate
correlations between conscientiousness and education supportive benefits for
basic and supervisory job subgroup revealed that there may be possible moder-
ation. Respondents holding basic category jobs (i.e., sales representative) had a
–0.18 correlation between conscientiousness and education benefit preferences,
while for respondents with supervisory jobs this correlation was 0.07. Although
the difference was too small to conclude that job category was a significant
moderator, future research involving larger samples may need to investigate
potential moderator effects between demographics and personality/values
variables.

Conclusion

The current research endeavoured to expand on the US-based literature in the


domain of flexible compensation benefits. First, it gathered much-needed data
from a sample of employees in the Asia Pacific region (New Zealand) and
showed that there was a considerable variation in benefit preferences. Second,
the study investigated whether personality traits and work values factors added
incremental validity to demographic variables when predicting employee
benefit preferences. This hypothesis was supported for education-supportive
benefit preferences, which were predicted by status and independence work
values, extroversion, conscientiousness, age, education, and job category.
However, family-supportive benefit preferences appeared to be driven solely
by demographics, that is, marital status, number of dependants and employ-
ment status. These findings add to the current literature by increasing our
knowledge of the factors that affect compensation preferences.
APJHR 46_2_Lawton.qxd 11/06/2008 10:33 PM Page 237

Determinants of employee benefit preferences 237

Implications and future directions

The results from this study have utility for organisations interested in the
design of a flexible benefits program. If an organisation’s workforce demo-
graphics are similar to those in the current study, our study provides a
benchmark with regard to the types and quantities of benefits employees may
desire. Having even preliminary estimates would be useful as it would ensure
that the benefits program would meet the needs and desires of employees, and
thus be more effective than a generic approach. The data would be of further
use in negotiating plan prices with external companies.
Even if a formal FBP is not prescribed, it would still be beneficial to vary
employee benefits to match the needs and wishes of different individuals. For
example, employee benefits could be offered as an incentive, or as a bargaining
tool in wage negotiations. In this instance, it would be advantageous to under-
stand the personal factors which influence employee preferences for different
types of benefits. For example, based on the findings of this study, family-
supportive benefits would be desirable to an employee who worked full-time
and had dependent children, whereas an extroverted low-level employee who
desired to advance their career would prefer education-supportive benefits.
Therefore, if human resources professionals possessed an understanding of an
employee’s personal demographic characteristics, personality, and work values,
they could design personal benefits schemes based on their knowledge of
employee preferences. This would create unique value for employees, leading
to a strategic competitive advantage in the labour market.
Another way in which the results of this study could be utilised is to
increase the number of employees who choose certain benefits. High-quality
benefit communication has been linked to increased employee benefit use and
knowledge (Sinclair, Leo and Wright 2005), and is therefore of utmost impor-
tance to an organisation when implementing a benefits program. Such
effective communication could be accomplished through marketing strategies
devised to target the specific groups that have the propensity to prefer each
type of benefit. For example, family-supportive benefits could be promoted by
informing employees about their potential to make family life easier and to
enhance their financial security, as preferences for these benefits are dictated
by demographic necessity. Conversely, education-supportive benefits could be
marketed by tapping into status and independence work values by raising
awareness about the need for education in high-level jobs.

Limitations
There are, of course, some concerns about how well the present study gener-
alises to the actual flexible-benefits scenarios. In this study, we only studied
respondents’ preferences for benefits (i.e. the likelihood that they would choose
each individual benefit on a Likert scale). Another approach is to ask respon-
APJHR 46_2_Lawton.qxd 11/06/2008 10:33 PM Page 238

238 Asia Pacific Journal of Human Resources 2008 46(2)

dents to make trade-off decisions about the benefits with respect to one
another. Previous studies have dealt with this issue using methods such as point
allocation (White 1983) or paired comparisons (Nealey 1964). These tech-
niques, however, are not without limitations as they are time-consuming,
demand higher cognitive load for participants than a Likert scale rating, and
yield rank order data. Future research should also investigate how relation-
ships between personality and value predictors generalise to the actual benefit
usage. Although it is reasonable to expect respondents acting in accordance
with their preferences, some loss of predictive power would be expected.
Another limitation affecting the strength of conclusions is the fact that
our sample was from a single organisation. Although a wide range of job cate-
gories and family types were represented, the participants may have shared
some common interests which affected their compensation preferences.
Furthermore, because this research was conducted in a real-world organisa-
tion, some participants may have intentionally overrated the benefits as they
thought it may make it more likely that the program would get implemented.
Future research attempting to replicate or expand on our findings would
benefit from utilising a sample representing multiple organisations. If sample
sizes allow, particular attention should be paid to interplay between demo-
graphic variables, personality and values. Job category, experience, or education
level may act as a moderator of relationships between personality variables and
benefit preferences.
In conclusion, the current investigation has shown that the individual
difference variables of work values and personality affect employee benefit
preferences, over and above the effects of demographic variables. This finding
has potential implications for compensation research, especially for the area of
strategic pay. For example, compensation packages which closely match the
preferences of employees may lead to increased motivation, and in turn,
performance. Future investigation should also investigate the generalisability
of these findings with respect to actual benefit usage. It is hoped that the
current analysis will drive future research in the area, not only to validate the
existing findings, but also to explore other compensation initiatives that could
give employers a leading edge in the labour market.

Oleksandr S. Chernyshenko (PhD, Illinois Urbana-Champaign) is currently an associate professor of


strategy, management, and organization at Nanyang Business School in Singapore. Previously, he
taught at the psychology department at the University of Canterbury, New Zealand. His research
focuses mainly on issues related to personnel selection and performance management.

Katie Lawton (MSc (AppPsych), Univ of Canterbury) works as an HR consultant at the University of
Canterbury, supporting such functions as job analysis and evaluation, information systems, and
recruitment and selection.
APJHR 46_2_Lawton.qxd 11/06/2008 10:33 PM Page 239

Determinants of employee benefit preferences 239

References
Aguinis, H., J.C. Beaty, R.J. Boik, and C.A. Pierce. 2005. Effect size and power in assessing
moderating effects of categorical variables using multiple regression: A 30-year review.
Journal of Applied Psychology 90(1): 94–107.
Barber, A.E., R.B. Dunham, and R.A. Formisano. 1992. The impact of flexible benefits on
employee satisfaction: A field study. Personnel Psychology 45(1): 55–76.
Barringer, M.W., and G.T. Milkovich. 1996. Employee health insurance decisions in a flexible
benefits environment. Human Resource Management 35(3): 293–315.
Cable, D.M., and T.A. Judge. 1994. Pay preferences and job search decisions: A person–
organisation fit perspective. Personnel Psychology 47(2): 317–48.
Costa, P.T., and R.R. McCrae. 2003. Personality in adulthood: A five-factor theory perspective.
2nd edn. New York: Guildford Press.
Frone, M.R., and J.K Yardley. 1996. Workplace family-supportive programmes: Predictors of
employed parents’ importance ratings. Journal of Occupational and Organizational
Psychology 69(4): 351–66.
Furnham, A. 1992. Personality at work: The role of individual differences in the workplace.
London: Routledge.
Goldberg, L.R. 1992. The development of markers for the big-five factor structure.
Psychological Assessment 4(1): 26–42.
Goldberg, L.R. 1999. A broad-bandwidth, public-domain, personality inventory measuring
the lower-level facets of several five-factor models. In Personality psychology in Europe,
eds I. Mervielde, I. Deary, F. De Fruyt, and F. Ostendorf, vol. 7, 7–28. Tilburg, The
Netherlands: Tilburg University Press.
Guenole, N., and O.S. Chernyshenko. 2005. The suitability of Goldberg’s Big Five personality
markers in New Zealand: A dimensionality, bias, and criterion validity evaluation. New
Zealand Journal of Psychology 34(2): 85–95.
Haslinger, J.A., and D. Sheerin. 1994. Employee input: The key to successful benefits
programs. Compensation and Benefits Review 26(3): 61–70.
Hedge, J.W., W.C. Borman, and S.E. Lammlein. 2006. The aging workforce. Washington DC:
American Psychological Association.
Lawler, E.E. 1990. Strategic pay: Aligning organizational strategies and pay systems. San
Francisco: Jossey-Bass.
Major, D.A., J.E. Turner, and T.D. Fletcher. 2006. Linking proactive personality and the big
five to motivation to learn and development activity. Journal of Applied Psychology 91(4):
927–35.
Manhardt, P.J. 1972. Job orientation of male and female college graduates in business.
Personnel Psychology 25(2): 361–8.
Meisenheimer, J.R., and W.J. Wiatroeski. 1989. Flexible benefits plans: Employees who have a
choice. Monthly Labor Review 112(12): 17–23.
Metlife Incorporated. 2006. The Metlife study of employee benefits trends: Findings from the 2005
national survey of employees and employers. New York: Metropolitan Life Insurance
Company.
Meyer, J.P., N.J. Allen, and C.A. Smith. 1993. Commitment to organisations and occupations:
Extension and test of a three-component conceptualisation. Journal of Applied Psychology
78(4): 538–51.
Meyer, J.P., P.G. Irving, and N.J. Allen. 1998. Examination of the combined effects of work
values and early work experiences on organisational commitment. Journal of
Organisational Behaviour 19 (1): 29–52.
APJHR 46_2_Lawton.qxd 11/06/2008 10:33 PM Page 240

240 Asia Pacific Journal of Human Resources 2008 46(2)

Nealey, S.M. 1964. Determining worker preferences among employee benefit programs.
Journal of Applied Psychology 48(1): 7–12.
Roberts, B.W., O.S. Chernyshenko, S. Stark, and L.R. Goldberg. 2005. The structure of
conscientiousness: An empirical investigation based on seven major personality
questionnaires. Personnel Psychology 58(1): 103–39.
Rushton, P.J., C.J. Brainerd, and M. Pressley. 1983. Behavioral development and construct
validity: The principle of aggregation. Psychological Bulletin 94(1): 18–38.
Sinclair, R.R., M.C. Leo, and C. Wright. 2005. Benefit system effects on employees’ benefit
knowledge, use, and organisational commitment. Journal of Business and Psychology
20(1): 3–29.
Statistics NZ. 2007. Main indicators. Statistics New Zealand website accessed 24 January 2007
from http://www.statsnz.govt.nz.
White, R.A. 1983. Employee preferences for non-taxable compensation offered in a cafeteria
compensation plan: An empirical study. Accounting Review 58(3): 539–60.

Вам также может понравиться