Katou Budwhar Patel 2014
Katou Budwhar Patel 2014
Katou Budwhar Patel 2014
PROCESS IN
THE HRM-PERFORMANCE
RELATIONSHIP: AN EMPIRICAL
EXAMINATION
A N A S TA S I A A . K AT O U , PA W A N S . B U D H W A R ,
A N D C H A R M I PAT E L
O
ver the last 20 years or so, a sub- (Paauwe, 2009), and the specific ways in
stantial volume of research has which HRM affects organizational perfor-
been devoted to examine the link mance are not clear (Buller & McEvoy, 2012).
between human resource (HR) This is because empirical research differs
practices and organizational per- widely with respect to the definitions and
formance. Although empirical research sug- measures of HR practices and performance,
gests that human resource management levels of analysis, and theoretical founda-
(HRM) positively impacts organizational per- tions (Guest, 2011; Paauwe, 2009).
formance (e.g., Boselie, Dietz, & Boon, 2005; Two fundamental approaches have
Combs, Liu, Hall, & Ketchen, 2006), the been followed in the HRM–organizational
Human Resource Management, July-August 2014, Vol. 53, No. 4. Pp. 527–544
© 2014 Wiley Periodicals, Inc.
Published online in Wiley Online Library (wileyonlinelibrary.com).
DOI:10.1002/hrm.21606
528 HUMAN RESOURCE MANAGEMENT, JULY–AUGUST 2014
performance debate: the best practices (Pfeffer, deliver HR practices in order to achieve a spe-
1994), supporting that context and HR prac- cific purpose (e.g., improved organizational
tices are mutually independent in improv- performance), the employees perceived the
ing organizational performance (Huselid meaning of these HR practices according to
& Becker, 1996), and the best fit (Schuler & their individual dispositions. But if the mean-
Jackson, 1987), supporting that HR practices ing of the HRM system is not shared among
should be consistent with a given context in employees, the shared perceptions of employ-
maximizing business performance (Schuler ees may be weak, and as a result, the specific
& Jackson, 1987; Wright & Snell, 1998). In purpose intended by employers may not be
either approach, the leading trend in research properly achieved (Nishii et al., 2008). To
on the HRM-performance relationship has address this question, HRM researchers have
moved from a focus on individual HR prac- called for more attention to be placed on
tices to a set, or HRM system of practices. The investigating the relationship between HRM
individual practices that make up the HRM and performance through employee shared
system constitute the content of perceptions in reactions to HRM (Wright &
the HRM system. Thus, the con- Van De Voorde, 2007). This is because the
If the meaning of tent of the HRM system refers to same HRM system could lead to broadly
the HRM system is a set of HR practices and policies varying employee reactions, resulting in dif-
through which organizations can ferent levels of organizational performance
not shared among improve the acquisition, develop- (Takeuchi, Chen, & Lepak, 2009; Takeuchi,
ment, retention, and utilization Lepak, Wang, & Takeuchi, 2007).
employees, the of their human capital in order to Further, although empirical studies dem-
achieve the strategic goals of the onstrate the validity of the process-based
shared perceptions
organization such as organiza- approach in advancing the understand-
of employees may tional performance (Boselie et al., ing in the HRM-performance relationship
2005). The underlying assump- (Frenkel, Li, & Restubog, 2012; Li et al., 2011;
be weak, and as a tion in these approaches is that Nishii et al., 2008; Sanders, Dorenbosch, &
HRM systems may affect organi- de Reuver, 2008; Strumpf, Doh, & Tymor,
result, the specific
zational performance through its 2010; Takeuchi et al., 2007), these studies
purpose intended impact on employee attitudes and have primarily concentrated on the process-
behavior (Combs et al., 2006). based approach while considering HRM con-
by employers may Over the last decade or so, the tent as a given. This eclectic treatment of the
literature has paid more atten- HRM system may produce erroneous results
not be properly
tion to the HRM process view (see due to misspecification regarding the guid-
achieved. Bowen & Ostroff, 2004; Nishii, ing factors of the HRM-performance rela-
Lepak, & Schneider, 2008). HRM tionship. Bowen and Ostroff (2004, p. 206)
process refers to “the way HR propose “HRM content and process must be
policies and practices are communicated to integrated effectively in order for prescriptive
employees” (Li, Frenkel, & Sanders, 2011, p. models of strategic HRM actually to link to
1826), or alternatively, “to the features of an firm performance.”
HRM system that send signals to employees In this article, we extend and test this
that allow them to understand the desired and idea by examining the effects of an HRM
appropriate responses and form a collective system, which integrates both content and
sense of what is expected” (Bowen & Ostroff, process, on organizational performance, con-
2004, p. 204). This attention from HRM con- sidering that the mechanism through which
tent to process is based on the assumption collective employee reactions influence orga-
that employees may find it difficult to attach nizational performance still remains an unan-
only one kind of meaning to an HRM system swered question (Boxall & Macky, 2009).
because individuals may not uniquely inter- Particularly, we address the linkage between
pret the same HR practices. This means that HRM and organizational performance at the
although the employers were intending to macro level by considering that this linkage
can be explained through the aggregation of (Becker & Huselid, 2006), the purpose of this
individual-level mediating actions of employ- article is to investigate whether the HRM
ees (Minbaeva, Foss, & Snell, 2009). In doing system, expressed by integrating both con-
so, the aggregation possibility of individual tent—reflected by a system of perceived
survey data on perceptions of HRM and HR practices (Nishii & Wright, 2008)—and
employee reactions into meaningful organi- process—reflected by distinctiveness, con-
zational-level constructs is accordingly evalu- sistency, and consensus of an HRM system
ated (Ostroff & Bowen, 2000). (Bowen & Ostroff, 2004)—influ-
The data for this study is obtained by an ence organizational performance
The purpose of
employee survey in both public and private through collective employee reac-
sectors of Greece. Although the literature tions. Considering that generally this article is to
highlights studies that investigate various there is no empirical HRM research
aspects of the HRM-performance relation- to address this purpose (Höglund, investigate whether
ship in the Greek context (Dimitriades, 2007; 2011; Wright & Van De Voorde,
the HRM system,
Giannikis & Mihail, 2011), most of them 2007), the HRM-performance rela-
examining the HR practices–organizational tionship in this article is investi- expressed by
performance relationship have been con- gated by means of the following
ducted in the United States and, to a lesser research question: integrating both
extent, in the United Kingdom (for reviews,
To what extent can an HRM sys- content—reflected
see Boselie et al., 2005; Combs et al., 2006;
tem expressed by both content
Wright, Gardner, Moynihan, & Allen, 2005; by a system of
and process contribute through
Subramony, 2009). Furthermore, the majority
collective employee reactions
of strategic HRM research has been conducted perceived HR
toward organizational perfor-
in manufacturing environments, because the
mance? practices—and
positive relationship between HRM systems
and organizational performance is more The Research Framework process—reflected
stable in this sector (Combs et al., 2006),
and Hypotheses
neglecting the considerable presence of other by distinctiveness,
sectors (Liao, Toya, Lepak, & Hong, 2009). Based on the seminal work of
This study extends the investigation of the Huselid (1995), arguing that HRM consistency,
HRM systems–organizational performance has an impact on financial perfor-
and consensus
relationship in the manufacturing, services, mance, over the last 20 years or so
and trade sectors in the Greek context, where a large body of research has dem- of an HRM
manufacturing accounts for 18.1 percent of onstrated a consistent relation-
the GDP, while services and trade account for ship between HR practices and system—influence
46.1 percent and 33.7 percent, respectively, of organizational performance
organizational
the GDP (Eurostat, 2011). (Combs et al., 2006). This research,
Greece is a peripheral country in the utilizing the resource-based view performance
European Union that has been heavily of the firm (where the organiza-
affected by the 2008 economic and financial tion is defined by the resources it through collective
crisis. It would be interesting to extend the controls), was concentrating on a
debate on the role of HRM systems in improv- macro-framework, where both employee reactions.
ing organizational performance in countries HRM and performance were con-
that are under severe financial and economic sidered at the firm level. These theories were
crisis, such as Greece, and it can be extended criticized because they were concentrating on
to countries predicted to go down a similar the two endpoints of the relationship, and
route, such as Cyprus, Portugal, Spain, and thus they were neglecting to investigate the
Ireland. mediating mechanisms through which HR
In summary, based on attribution the- practices were hypothesized to affect organi-
ory (Kelley, 1967) and further assuming zational performance (Purcell & Kinnie,
that employees may be largely idiosyncratic 2007).
of the event-effect relationship. An HRM sys- and employee reactions. In other words, it
tem may be considered to be a strong one if it is argued that perceived HR practices shape
satisfies the features of distinctiveness, con- desired employee reactions in stronger situa-
sistency, and consensus. A strong HRM sys- tions compared to weaker situations (Li et al.,
tem produces a shared meaning about HRM 2011; Nishii & Wright, 2008; Sanders et al.,
among the employees, thus shaping com- 2008). Hence, it is hypothesized that:
mon attitudes and behavior, which influence
organizational goals. Hypothesis 2: Features of HR practic-
Bowen and Ostroff (2004, p. 207) wrote, es (distinctiveness, consistency, and A strong HRM
“In a strong situation employees develop a consensus) moderate the relationship
system produces
shared interpretation of the organization’s between perceived HR practices and
policies, practices, procedures, and goals employee reactions. a shared meaning
and develop shared perceptions about what
behaviors are expected and rewarded in Based on Hypotheses 1 and about HRM among
the organization.” This is because in strong 2, Figure 1 presents the opera-
tional model of this study. the employees,
situations individuals perceive events the
same way, and accordingly make uniform Particularly, this model supports thus shaping
expectations about the most appropriate the view that perceived HR prac-
responses (Ostroff & Bowen, 2000; Schneider, tices (i.e., HRM content) have an common attitudes
Salvaggio, & Subirats, 2002). On the contrary, effect on organizational perfor-
in weak situations employees experience a mance through the mediating and behavior,
high degree of ambiguity regarding what the mechanism of employee reac-
which influence
appropriate responses should be (González- tions, and HR features (i.e., HRM
Romá, Peiró, & Tordera, 2002). This is because process) moderate the perceived organizational goals.
in weak situations individuals do not perceive HR practices–employee reactions
events the same way, and accordingly do not relationship.
make uniform expectations about the most
appropriate responses (Schneider et al., 2002). Method
Considering that the features of distinctive- Sample
ness, consistency, and consensus determine
the strength of the situation, it is supported Data for this research was collected in
that these features may moderate the rela- October–December 2011 by help of a ques-
tionship between perceived HR practices tionnaire survey, which was distributed to the
HR Practices
Features
Distinctiveness
Consistency
Consensus
employees of public and private organiza- percent belonged to the public and 81.2 per-
tions in the manufacturing, services, and cent to the private sector; 24.1 percent were
trade sectors covering the whole of Greece. from the manufacturing sector, 42.9 percent
The questionnaires were administered by 100 were from the services sector, and 33.1 per-
individuals (samplers) pursuing management cent were from the trade sector. Of the sample
degrees at a Greek business school. Con- of 1,250 respondents, 60.5 percent were male
sidering that these individuals were coming and 39.5 percent were female. Their average
from all over Greece, they helped to collect age was 38.21 (± 9.81) years old, and the aver-
the data from organizations where they were age seniority was 10.39 (± 8.03) years. Finally,
working and also from their contact organiza- 22.2 percent of the respondents were senior
tions, thus ensuring the randomness of the managers, 23.4 percent were middle manag-
sample. The survey instrument was distrib- ers, and 54.4 percent belonged to the other
uted to 400 organizations with more than 20 category.
employees. The questionnaires were com-
pleted with the help from the people who Measures
administered them, after the latter had
attended seminars for ensuring that they all Most measures were directly taken from the
follow the same protocol for selecting the cited research, and very few were modified
respondents and for understanding the mean- from prior research, as indicated later.
ing of the questionnaire. Following Gerhart, The HR features construct comprised three
Wright, McMahan, and Snell (2000), who subscales—distinctiveness, consistency, and
suggest that the reliability of HR measures consensus (Bowen & Ostroff, 2004). Each sub-
will be increased by using five to ten respon- scale included a number of sub-subscales. For
dents per firm, the samplers were asked to example, distinctiveness comprised the four
concentrate on up to eight respondents from sub-subscales of visibility, understandability,
each organization. Considering that Greek legitimacy, and relevance. The items for each
organizations are rather small, the individu- sub-subscale were taken from Delmotte, de
als who administered the questionnaires were Winne, Gilbert, and Sels (2007). However,
asked to concentrate on two respondents for two items were transformed from reverse cod-
each firm at the senior-management level ing into forward coding for being consistent
(i.e., the heads of the Finance/HRM/Personnel with all other items. For example, the origi-
Department, for answering questions with nal item of “The HR practices implemented
respect to organizational performance), two in this organization sound good in theory
respondents at the middle-management level but do not function in practice (R)” was
(i.e., line managers, for answering questions transformed into the item of “The HR prac-
with respect to the HR practices features), and tices implemented in this organization sound
four respondents at other levels (i.e., adminis- good in theory and function in practice.” The
trative and other workers who were randomly items for each sub-subscale were measured
selected, for answering questions with respect on a scale ranging from 1 = totally disagree
to HR practices as experienced and employee to 5 = totally agree. Examples of the items
reactions). According to this protocol, the included “I was attracted to this company
samplers were asked to distribute a total of because of its good HR policies and practices”
3,200 questionnaires. A total of 1,250 usable (visibility) and “The HR department under-
questionnaires were returned from the takes exactly those actions that meet our
employees in 133 organizations, a response needs” (relevance).
rate of 33.25 percent at the organization level, The perceptions of HR practices construct
and 39.06 percent at the employee level. was measured via four subscales—resourc-
Of the sample of 133 organizations, 51.9 ing, development, reward, and relations.
percent had 20 to 100 employees, 27.8 per- Considering that perceptions of HR prac-
cent had 101 to 200 employees, and 20.3 tices require that attention is focused on
percent had more than 200 employees; 18.8 how employees experience and then judge
the level of satisfaction derived from the HR good. Example items included “How would
practices applied to them (Purcell & Kinnie, you rate effectiveness (i.e., if the organization
2007), the items of each subscale were taken meets its objectives) in your organization?”
from Kinnie et al. (2005), and they were mea- and “How would you rate development (i.e., if
sured on a scale ranging from 1 = very little the organization is developing in its capacity
to 5 = very much. Example items included to meet future opportunities and challenges)
“How satisfied do you feel with the level of in your organization?” For all questions refer-
training you receive in your current job?” and ring to organizational performance dimen-
“How satisfied do you feel with your pay?” sions, a specific definition was assigned to
The employee reactions construct com- produce a better focus in responses.
prised four subscales—motivation (Lockwood, Several additional organizational vari-
2010), organizational commitment (Allen & ables were controlled in order to avoid empir-
Meyer, 1990), work engagement (Schaufeli, ical results of the analysis being erroneous
Salanova, Gonzalez-Roma, & Bakker, 2002), (Boselie et al., 2005). Specifically, the controls
and organizational citizenship behavior, or used are the sector of production where the
OCB (Niehoff & Moorman, 1993). Each sub- organizations are activated, the ownership of
scale included a number of sub-subscales. For organizations indicating whether the orga-
example, organizational commitment was nization is public or private, and the size of
composed of three sub-subscales—affective the organization in terms of the number of
commitment, continuance commitment, people employed.
and normative commitment. Similarly, OCB
comprised five sub-subscales—altruism, cour- Consistency of the Survey
tesy, sportsmanship, conscientiousness, and
Instrument and Data Aggregation
civic virtue. The items for each sub-subscale
were taken from the cited research and were To avoid the disadvantages of using one indi-
measured on a scale ranging from 1 = totally vidual as the key informant for practices and
disagree to 5 = totally agree. Examples of the outcomes, data were collected using multiple
items included “I feel proud to tell people employees per organization as respondents.
who I work for” (affective commitment) and This means that employee responses within
“I am helping others who have very heavy an organization should be aggregated at the
work loads” (altruism). organizational level to reflect the collective
Organizational performance is a multifac- perception of the HRM system, reactions, and
eted concept, which is usually indicated by performance (Takeuchi et al., 2007, 2009).
the respondent’s perceptions measured by the Assuming that the dataset refers to employees
help of three subscales—productivity, growth, being at three different levels in the organiza-
and creativity (see Delaney & Huselid, 1996; tional hierarchy (i.e., senior managers, mid-
Yang, Huang, & Hsu, 2014). The productiv- dle managers, and other employees), for
ity subscale includes the items of effectiveness aggregating the data at organizational level
(i.e., if the organization meets its objec- we followed a two-phase methodology.
tives), and efficiency (i.e., if the organization In the first phase, composite scales at a
uses the fewest possible resources to meet respondent level were created following a
its objectives). The growth subscale includes three-step approach. First, low-level compos-
the items of development (if the organization ite scales were developed as averages of all
is developing in its capacity to meet future its items comprising the scale. For example,
opportunities and challenges) and satisfaction the composite scale of visibility was devel-
of all participants (stakeholders, employees, oped using the four relevant items of the
customers). The creativity subscale includes questionnaire. Second, medium-level com-
the items of innovation (for products and pro- posite scales were developed as averages of all
cesses) and quality (percentage of products of its low-level composite scales already devel-
high quality). The items were measured on a oped in step one. For example, the composite
scale ranging from 1 = very bad to 5 = very scale of distinctiveness was developed using
Ownership Size
1
ability, legitimacy, and relevance developed
in step one. Finally, higher-level composite
scales were developed as averages of all its
–0.193*
medium-level composite scales already devel-
1
oped in step two. For example, the composite
scale of HR features was developed using the
Performance Sector
medium-level scales of distinctiveness, consis-
–0.104
0.109
tency, and consensus developed in step two.
1
The development of these composite scales
Organizational
was possible considering that all Cronbach’s
alphas are higher than 0.70 (Nunnally, 1978),
[0.852]
0.037
–0.072
–0.074
establishing survey instrument construct
internal consistency, and that the percentage
of all total variance explained per dimension
obtained by applying confirmatory factor
Employee
Reactions
analysis (CFA) with Varimax rotation and
0.247**
Means, Standard Deviations, and Correlation Coefficients of the Constructs Used in the Study
[0.695]
the eigenvalue greater than one criterion are
–0.090
–0.015
0.016
higher than 50.0 percent, indicating accept-
able survey instrument construct validity
(Hair, Anderson, Tatham, & Black, 2008).
Perceived HR
In the second phase, aggregative scales at
Practices
0.378**
0.710**
[0.848]
an organizational level were developed using
0.024
0.041
–0.020
the composite scales developed in the first
phase (Landis, Beal, & Tesluk, 2000). To jus-
tify the aggregation of individual-level survey
HR Practices
–0.051
0.001
0.015
3.522 (0.449)
3.156 (0.623)
3.887 (0.714)
Deviations)
2
+ ∑ 1 – λi
∑λi
Employee reactions
Size
that the degree of construct reliability is large samples and high correlations, to indi-
acceptable, since all reliability scores exceed cate excellent, good, or mediocre fit, respec-
0.90. Construct discriminant validity was tively); the goodness of fit index, or GFI (with
assessed by examining whether the square a critical level not lower than 0.80, or 0.70
root of each factor’s AVE is larger than its cor- for complex models); the normed fit index,
relations with other factors. Table I presents or NFI (with a critical level not lower than
the correlation coefficients of all constructs 0.90); the comparative fit index, or CFI (with
used in the study. It is seen that the correla- a critical level not lower than 0.90); and the
tion coefficients are smaller than the square root mean squared error of approximation,
root of each factor’s AVE, thus providing evi- or RMSEA (with critical levels not more than
dence for separate constructs. 0.05, 0.08, or 0.10 to indicate excellent, good,
To reduce the common method bias or mediocre fit, respectively; for details see
threat in the survey design, we asked multi- Hair et al., 2008).
ple respondents at three levels in the organi-
zation to answer the questions in the survey
Results
(Lindell & Whitney, 2001). However, taking
into consideration that some correlation coef- Results in Table I show strong, positive, and
ficients were rather high, Harman’s (1967) significant correlations between all structural
single-factor test was also used to examine constructs, thus supporting the hypotheses of
the likelihood of a common method bias the study. However, results based on correla-
threat. According to this test, the simultane- tions, although interesting, may be mislead-
ous loading of all items in a factor analysis ing due to the interactions between several
revealed five factors and not just one, with variables. Therefore, in order to isolate the
the first factor covering only 27.53 percent possible links between the variables involved
of the total variance explained, thus indicat- in the research framework presented in
ing that the common method bias in the data Figure 1, two models have been estimated.
was rather limited. The first model reflects HRM content and
refers only to mediation (not including the
moderating effect of HR practices features).
Statistical Analysis
The second model reflects both HRM content
To test the hypotheses developed for the pro- and process, refers to both the mediation and
posed framework, the methodology of struc- moderation dimensions, and refers to the
tural equation modeling (SEM) was used via complete model presented in Figure 1.
LISREL, with bootstrapping estimation (see The goodness-of-fit indices (chi-square
Jöreskog & Sörbom, 2004). SEM is effective = 876.40, df = 361, p = 0.000; normed-chi-
when testing models that are path analytic square = 2.428, RMSEA = 0.104, NFI = 0.90,
with mediating variables, and includes latent CFI = 0.93, GFI = 0.73) relatively confirmed
constructs that are being measured with mul- the validity of the estimated HRM content
tiple items. We used bootstrapping because model presented in Figure 2. The goodness-of-
this method is considered to be the most fit indices (chi-square = 701.38, df = 359, p =
appropriate for testing mediation due to the 0.000; normed-chi-square = 1.954, RMSEA =
fact that it does not require the normality 0.085, NFI = 0.92, CFI = 0.94, GFI = 0.75)
assumption to be met (Shrout & Bolger, 2002). confirmed the validity of the complete model
We assessed the overall model fit fol- presented in Figure 3. However, it must be
lowing Bollen’s (1989) recommendation to noted that considering that chi-square statis-
examine multiple indices, since it is possible tics may be inflated by high sample sizes and/
for a model to be adequate on one fit index or high correlations, the value of the normed-
but inadequate on many others. We used the chi-square (i.e., value of chi-square/degrees of
chi-square test (with a critical significant level freedom) was used instead. In our cases, these
p > 0.05) and the normed-chi-square ratio values are less than 3, confirming the validity
(with critical levels 1–3, 3–5, and 5–7 for very of the two models.
Training 0.92
Perceived HR 0.75 Employee 0.54 Organizational 0.88
Growth
Practices Reactions Performance
Rewards
0.88
-0.08*
Creativity
0.87
Relations 0.90 Sector
Work
Motivation Commitment OCB
Resourcing 0.83 engagement
0.69 0.91 0.84 0.51
Training 0.92
Perceived HR 0.22
Practices
Rewards 0.88 0.88 Productivity
0.29
Relations 0.90 0.60
Employee Organizational 0.88
Growth
Reactions Performance
0.45
Distinctiveness 0.92 0.08* -0.07* Creativity
0.87
0.93 HR Practices Interactions
Consistency Practices × Features Sector
Features
Consensus 0.92
Comparing the goodness-of-fit indices of devoted to the results of this model. The cir-
the two models, we see that the fit statistics of cles in Figures 2 and 3 represent the related
the second model are better than those of the latent variables, the bold arrows indicate the
first model. Thus, we conclude that the com- structural relationships between the corre-
plete moderation-mediation model is more sponding variables, and the numbers that are
preferred than only the mediation model, assigned to each arrow show the estimated
and consequently, the rest of the study is standardized coefficients. All coefficients are
significant at p < 0.01, except the starred coef- HR practices and employee reactions, thus
ficients that are significant at p < 0.05. Finally, supporting Hypothesis 2. Specifically, we see
we must note here that although we tried that HR practices features directly influence
all possibilities connecting controls with all employee reactions. However, as suggested by
the other constructs (Paauwe & Richardson, Bowen and Ostroff (2004, p. 215), “it is likely
1997), the only significant results obtained that some features are more critical than oth-
are those reported in Figures 2 and 3. ers in creating a strong situation. For example,
without consistent HRM messages, distinc-
tiveness and consensus may lose impact.” In
Findings
Figure 3, we see that the standardized coef-
The major findings of this study are now ficient of consistency is larger than the coeffi-
summarized. First, employee reactions posi- cients of distinctiveness and consensus, thus
tively and partially mediate (Baron & Kenny, verifying this suggestion. Furthermore, the
1986) the relationship between perceived HR total impact of HRM content on organiza-
practices and organizational performance, tional performance is equal to 0.394, which
thus supporting Hypothesis 1. Specifically, is stronger than the corresponding impact
considering the levels of the standardized of HRM process, which is equal
coefficients in Figure 3, it is seen that to 0.270. This finding is impor-
Employee training
employee training and relations influence tant because it indicates that the
employee commitment and work engage- proper content of high-perfor- and relations
ment the most, which in turn have a positive mance HR practices may improve
impact on organizational performance. This organizational performance, spe- influence employee
result is important, because although it indi- cifically in periods of economic
commitment and
cates that training and development may crisis, where organizations might
help organizations to improve their perfor- benefit by putting in more efforts work engagement
mance, on the contrary, in periods of eco- to communicate to employees the
nomic crises the first action that businesses specific features of their HR prac- the most, which
take is to press down costs by reducing train- tices. In other words, it may be
in turn have a
ing and development expenditures. more important to put effort in
Additionally, we see that good relations finding appropriate HR practices positive impact
between all people involved in an organiza- according to context rather than
tion seem to be important, because they risk failing to properly implement on organizational
develop a firm’s social capital (Nahapiet & these HR practices (Guest, 2011).
Ghoshal, 1998), which is defined as “the fea- Additionally, considering the performance.
tures of social life—networks, norms and interaction coefficients between
trust—that enable participants to act together perceived HR practices and HR practices fea-
more effectively to pursue shared objectives” tures, it is seen that the interaction of HR
(Putnam, 1996). However, considering the practices features with employee training
rather low standardized coefficients of moti- influences the most employee reactions, and
vation and OCB, this study is not firmly sup- thus organizational performance. Besides,
porting the findings of Jiang, Lepak, Hu, and using the data in Table I and the results in
Baer (2012) for improved employee motiva- Figure 3, and applying the ModGraph soft-
tion and also that of Mossholder, Richardson, ware (Jose, 2008), Figure 4 presents the mod-
and Settoon (2011) for improved OCB. We eration effects of HR practices features in the
believe that this finding is important because relationship between perceived HR practices
it may reflect the discouragement of employ- and employee reactions. The findings show
ees working in the unstable and insecure that the higher the strength of the HRM
Greek business environment. system, the higher (i.e., upward shift) the
Second, HR practices features positively positive relationship between perceived HR
and quasi-moderate (Sharma, Durand, & Gur- practices and employee reactions, although
Arie, 1981) the relationship between perceived the strength (i.e., slopes) of this relationship
Fourth, this study extends the literature that social capital in Greece may be charac-
referring to the HRM-performance relation- terized as weak (Jones, Malesios, Iosifides, &
ship in a non-US/UK context (i.e., Greece), Sophoulis, 2008). Therefore, where an HRM
reconfirming that context matters (Combs system triggers systematic processing of HR
et al., 2006). Considering that Greece is practices signals resulting in stable interpreta-
under a severe financial and economic cri- tions that are shared within an organization,
sis, this investigation is uniquely context- managers need to develop skills for designing
specific. On the whole, the analysis indicates and implementing HRM systems that lead
that although employees are discouraged and to desired employee reactions. Open com-
feeling insecure in cases where organizations munication, for example, between managers
face severe financial problems, their reactions and their employees during an economic cri-
appear to still function as a mediating mecha- sis period may facilitate desired actions that
nism between HRM and performance. will make employees commit-
ted from the first day of joining On the whole, the
the organization. Additionally,
Implications for Practice
frequent professional dialogue analysis indicates
A core message of the HRM content theoriza- between managers and their
tion is that employee training and relations employees may reduce possible that although
are two major dimensions of a high-perfor- negative outcomes associated
employees are
mance HR practices system as it improves with HR practices experienced dif-
organizational performance, even in cases ferently by individuals (Bowen & discouraged and
where the organization is operating under an Ostroff, 2004; Nishii et al., 2008).
economic crisis environment. Additionally, a Alternatively, in cases of economic feeling insecure
core message of the HRM process theoriza- crisis, managers may concentrate
in cases where
tion is that consistency is the principal fea- on those employees who produce
ture of a strong HR practices system, as it also the highest value to the organiza- organizations face
influences organizational performance, even tion. However, this approach risks
in an economic crisis environment. These the likelihood of undesirable reac- severe financial
two core messages indicate that the design tion by other employees who per-
problems, their
and implementation of appropriate HR prac- ceive unfair treatment (also see Li
tices systems may enhance human capital et al., 2011). reactions appear
(through training) and social capital (through Finally, recognizing that in
relations), making individuals in the organi- today’s global economy human to still function
zation share common goals. Consequently, resources constitute a competitive
this study contributes to the recognition that advantage asset, the improvement as a mediating
the content and process characteristics of an of just human capital (i.e., what mechanism
HRM system influence organizational perfor- you know) and social capital (i.e.,
mance (Bowen & Ostroff, 2004). who you know) is not enough. between HRM and
The findings of the study also suggest Specifically, in cases of severe eco-
that practitioners should make their HRM nomic crisis such as in Greece, performance.
systems attractive to employees and com- where managers and employees
municate them consistently and frequently, are discouraged and feel unsafe, practitioners
so that employees appreciate their value should try to improve the psychological capital
(Burton, Lauridsen, & Obel, 2004; Li et al., (i.e., who you are) in their organizations, by
2011; Ngo, Lau, & Foley, 2008). However, developing confidence, hope, optimism, and
under an economic crisis context, the con- resilience, aiming at improving both individ-
ceptualization of HRM as a signaling system ual and organizational performance (Luthans,
that enables employees to develop strong Luthans, & Luthans, 2004). Accordingly, in
individual and collective perceptions of periods of instability where downsizing is par-
expected attitudes and behaviors is a difficult ticularly important for organizations that rely
task. This is much more difficult considering on human and social capital for competitive
advantage, organizations should adopt HR process of an HRM system, the present study
practices that increase employee education, has implications for future research, in con-
knowledge, and skill, and interpersonal rela- texts similar to Greece that are under severe
tions in their work, which, ultimately, have a economic and financial crisis.
positive effect on commitment to the organi- In terms of the methodology followed,
zation and motivation to perform (Iverson & this study applied three criteria (two intra-
Zatzick, 2011). class correlation coefficients and an inter-rater
agreement measure) for assessing the aggrega-
tion possibilities of individual survey data on
Limitations and Further Research
perceptions of HRM and employee reactions
This study has four main limitations. First, the into meaningful organizational-level con-
data were collected using a questionnaire at a structs. Although individual survey data are
single point in time. As a result, the study does frequently aggregated to scores at higher lev-
not allow for dynamic causal inferences els within the organization, in this study the
(Cavanaugh & Noe, 1999). Future research questions were answered by multiple respon-
would benefit by employing longitudinal data. dents and multiple agents (senior managers,
Second, as the study applied a cross-sectional middle managers, and other employees) for
design, we could not rule out the possibility of appropriately reporting and interpreting mean
spurious correlations between the dependent scores and linkages between them. Therefore,
and independent variables. In this regard, lon- the study confirmed linkages between HRM
gitudinal research designs would be highly and organizational performance at the group
valuable for future similar investigations. level and HRM and employee reactions at the
Third, in aggregating the data of senior man- individual level, contributing to empirical
agers, middle managers, and other employees, tests of Bowen and Ostroff’s (2004) integra-
the sample sizes used were rather small. Future tive HRM content and process–organizational
research should aim for larger sample sizes for performance theorization.
this purpose. Fourth, considering that Greece Finally, the study explicitly addressed
is experiencing a severe financial crisis, the the “mediating mechanism hypotheses” of
findings from this unique context may not HRM content on organizational performance
generalize across borders (de Jong, Schalk, & and the “mediating-moderating mechanism
de Cuyper, 2009). Future research should con- hypotheses” of HRM content and process on
sider including other countries such as Cyprus, organizational performance, both through
Portugal, Spain, and Ireland that are experi- employee reactions. The study concluded
encing similar financial crises. that the HRM system that integrates both
Despite these limitations, this study pro- content and process presents a comprehen-
vided a theoretical and empirical test of the sive picture of the HRM-performance rela-
underlying assumption in the HRM literature tionship. Additionally, it concluded that the
that organizational HR practices can enhance HRM content effect is higher than HRM pro-
organizational performance, through their cess effect on organizational performance.
impact on various employee reactions (Combs Therefore, this study provided a useful start-
et al., 2006). Particularly, beyond extending ing point for future research that investigates
and providing one of the first studies examin- the impact of both content and process of
ing the validity of Bowen and Ostroff’s (2004) HRM systems on organizational performance
theorization in integrating the content and through employee reactions.
received a BA and an MBA from the University of Sunderland, UK, and a PhD and a
PgD from the Cardiff University in Wales, UK. She has written numerous articles that
have been published in leading academic journals, such as the International Journal of
Human Resource Management, Thunderbird International Business Review, Employee
Relations, the European Management Journal, the European Journal of International
Management, Personnel Review, and the Journal of World Business.
Bliese, P. D. (2000). Within-group agreement, non- Burton, R. M., Lauridsen, J., & Obel, B. (2004). The
independence, and reliability: Implications for impact of organizational climate and strategic fit on
data aggregation and analysis. In K. J. Klein & firm performance. Human Resource Management,
S. Kozlowski (Eds.), Multilevel theory, research, 43, 67–82.
and methods in organizations (pp. 349–382). San Cavanaugh, M. A., & Noe, R. A. (1999). Antecedents
Francisco, CA: Jossey-Bass. and consequences of relational components
of the new psychological contract. Journal of González-Romá, V., Peiró, J. M., & Tordera, N. (2002).
Organizational Behavior, 20, 323–340. An examination of the antecedents and moderator
influences of climate strength. Journal of Applied
Collins, C. J., & Smith, K. G. (2006). Knowledge
Psychology, 87, 465–473.
exchange and combination: The role of human
resource practices in the performance of high-tech- Guest, D. E. (2011). Human resource management and
nology firms. Academy of Management Journal, performance: Still searching for some answers.
49, 544–560. Human Resource Management Journal, 21, 3–13.
Combs, L., Liu, Y., Hall, A., & Ketchen, D. (2006). How Guzzo, R. A., & Nooman, K. A. (1994). Human resource
much do high-performance work practices matter? practices as communications and the psychologi-
A meta-analysis of their effects on organizational cal contract. Human Resource Management, 33,
performance. Personnel Psychology, 59, 501–528. 447–462.
de Jong, J., Schalk, R., & de Cuyper, N. (2009). Hair, F., Anderson, R., Tatham, R., & Black, W. (2008).
Balanced versus unbalanced psychological Multivariate data analysis with readings. London,
contracts in temporary and permanent employ- UK: Prentice-Hall.
ment: Associations with employee attitudes.
Harman, H. H. (1967). Modern factor analysis. Chicago,
Management and Organization Review, 5, 329–351.
IL: University of Chicago Press.
Delaney, J. T., & Huselid, M. A. (1996). The impact of
human resource management practices on percep- Höglund, M. (2011). Microfoundations of HRM effects:
tions of organizational performance. Academy of Individual and collective attitudes and perform-
Management Journal, 39, 949–969. ance. Helsinki, Finland: Publications of the Hanken
School of Economics.
Delmotte, J., de Winne, S., Gilbert, C., & Sels, L.
(2007, November). Comparing line managers’ Huselid, M. A. (1995). The impact of human resource
and trade union representatives’ assessments of management practices on turnover, productivity,
HRM strength. Paper presented at the Dutch HRM and corporate financial performance. Academy of
Network Conference, Tilburg, the Netherlands. Management Journal, 38, 635–672.
Dimitriades, Z. S. (2007). The influence of service Huselid, M. A., & Becker, B. E. (1996). Methodological
climate and job involvement on customer-oriented issues in cross-sectional and panel estimates of the
organizational citizenship behavior in Greek service human resource management–firm performance
organizations: A survey. Employee Relations, 29, link. Industrial Relations, 35, 400–422.
469–491. Iverson, R. D., & Zatzick, C. D. (2011). The effects of
Eurostat. (2011). Eurostat Database: Statistics by downsizing on labor productivity: The value of
theme. European Commission. Retrieved from showing consideration for employees’ morale and
http://epp.eurostat.ec.europa.eu welfare in high-performance work systems. Human
Resource Management, 50, 29–44.
Frenkel, S. J., Li, M., & Restubog, S. L. D. (2012).
Management, organizational justice and emotional Jiang, K., Lepak, D. P., Hu, J., & Baer, J. C. (2012).
exhaustion among Chinese migrant workers: How does human resource management influence
Evidence from two manufacturing firms. British organizational outcomes? A meta-analytic inves-
Journal of Industrial Relations, 50, 121–147. tigation of mediating mechanisms. Academy of
Management Journal, 55, 1264–1294.
Gerhart, B. (2005). Human resources and business
performance: Findings, unanswered questions, and Jones, N., Malesios, C., Iosifides, T., & Sophoulis, C. M.
an alternative approach. Management Revue, 16, (2008). Social capital in Greece: Measurement and
174–185. comparative perspectives. South European Society
and Politics, 13, 175–193.
Gerhart, B., Wright, P. M., McMahan, G. C., & Snell,
S. A. (2000). Measurement error in research on Jöreskog, K. G., & Sörbom, D. (2004). LISREL 8.7 for
human resources and firm performance: How Windows [Computer software]. Lincolnwood, IL:
much error is there and how does it influence effect Scientific Software International, Inc.
size estimates? Personnel Psychology, 53, 803–834.
Jose, P. E. (2008). ModGraph-I: A programme to com-
Giannikis, K. S., & Mihail, D. M. (2011). Modeling job pute cell means for the graphical display of mod-
satisfaction in low-level jobs: Differences between erational analyses: The internet version, Version
full-time and part-time employees in the Greek 2.0. Victoria University of Wellington, New Zealand.
retail sector. European Management Journal, 29, Retrieved from http://www.victoria.ac.nz/psyc/staff
129–143. /paul-jose-files/modgraph/modgraph.php.
Kelley, H. H. (1967). Attribution theory in social psy- Mossholder, K. W., Richardson, H. A., & Settoon, R.
chology. In D. Levine (Ed.), Nebraska symposium P. (2011). Human resource systems and helping in
on motivation (Vol. 15, pp. 192–240). Lincoln, NE: organizations: A relational perspective. Academy of
University of Nebraska Press. Management Review, 36, 33–52.
Kinnie, N., Hutchinson, S., Purcell, J., Rayton, B., & Nahapiet, J., & Ghoshal, S. (1998). Social capital,
Swart, J. (2005). Satisfaction with HR practices intellectual capital, and the organizational
and commitment to the organization: Why one advantage. Academy of Management Review, 23,
size does not fit all. Human Resource Management 242–266.
Journal, 15(4), 9–29.
Ngo, H., Lau, C. M., & Foley, S. (2008). Strategic
Klein, K. J., Bliese, P. D., Kozlowski, S. W. J., human resource management, firm performance,
Dansereau, F., Gavin, M. B., Griffin, M. A. . . . Bligh, and employee relations’ climate in China. Human
M. C. (2000). Multilevel analytical techniques: Resource Management, 47, 73–90.
Commonalities, difference, and continuing ques-
Niehoff, B. P., & Moorman, R. H. (1993). Justice as a
tions. In K. J. Klein & S. Kozlowski (Eds.), Multilevel
mediator of the relationship between methods of
theory, research, and methods in organizations (pp.
monitoring and organizational citizen behavior.
512–551). San Francisco, CA: Jossey-Bass.
Academy of Management Journal, 36, 527–556.
Landis, R. S., Beal, D. J., & Tesluk, P. E. (2000). A com-
Nishii, L. H., Lepak, D. P., & Schneider, B. (2008).
parison of approaches to forming composite meas-
Employee attributions of the “why” of HR prac-
ures in structural equation models. Organizational
tices: Their effects on employee attitudes and
Research Methods, 3, 186–207.
behaviors, and customer satisfaction. Personnel
Lengnick-Hall, M. L., & Lengnick-Hall, C. A. (2003). HR’s Psychology, 61, 503–545.
role in building relationship networks. Academy of
Nishii, L. H., & Wright, P. (2008). Variability within
Management Executive, 17(4), 53–66.
organizations: Implications for strategic human
Lepak, D. P., Liao, H., Chung, Y., & Harden, E. (2006). A resource management. In D. B. Smith (Ed.), The
conceptual review of human resource management people make the place: Dynamic linkages between
systems in strategic human resource management individuals and organizations (pp. 225–248). New
research. In J. Martocchio (Ed.), Research in York, NY: Taylor and Francis.
personnel and human resource management (pp.
217–271). Stamford, CT: JAI Press. Nunnally, J. C. (1978). Psychometric theory. New York,
NY: McGraw-Hill.
Li, X., Frenkel, S., & Sanders, K. (2011). Strategic HRM
as process: How HR system and organizational Ostroff, C., & Bowen, D. E. (2000). Moving HR to
climate strength influence Chinese employee a higher level. In K. J. Klein & S. W. Kozlowski
attitudes. International Journal of Human Resource (Eds.), Multilevel theory, research, and methods
Management, 22, 1825–1842. in organizations (pp. 211–266). San Francisco, CA:
Jossey-Bass.
Liao, H., Toya, K., Lepak, D. P., & Hong, Y. (2009). Do
they see eye to eye? Management and employee Paauwe, J. (2009). HRM and performance:
perspectives of the high-performance work sys- Achievements, methodological issues and pros-
tems and influence processes on service quality. pects. Journal of Management Studies, 46,
Journal of Applied Psychology, 94, 371–391. 129–142.
Lindell, M. K., & Whitney, D. J. (2001). Accounting Paauwe, J., & Richardson, R. (1997). Introduction:
for common method variance in cross-sectional Special issue on HRM and performance.
research designs. Journal of Applied Psychology, International Journal of Human Resource
86, 114–121. Management, 8, 257–262.
Lockwood, N. R. (2010). Motivation in today’s work- Pfeffer, J. (1994). Competitive advantage through
place: The link to performance. Research Quarterly. people: Unleashing the power of the work force.
Alexandria, VA: Society for Human Resource Boston, MA: Harvard Business School Press.
Management. Purcell, J., & Kinnie, N. (2007). HRM and business per-
Luthans, F., Luthans, K. W., & Luthans, B. C. (2004). formance. In P. Boxall, J. Purcell, & P. Wright (Eds.),
Positive psychological capital: Beyond human and The Oxford handbook of human resource manage-
social capital. Business Horizons, 47, 45–50. ment (pp. 533–551). Oxford, UK: Oxford University
Press.
Minbaeva, D., Foss, N., & Snell, S. (2009). Bringing the
knowledge perspective into HRM. Human Resource Putnam, R. (1996, March). Who killed civic America?
Management, 48, 477–483. Prospect, pp. 66–72.
Sanders, K., Dorenbosch, L., & de Reuver, R. (2008). Takeuchi, R., Chen, G., & Lepak, D. P. (2009). Through
The impact of individual and shared employee the looking glass of a social system: Cross-level
perceptions of HRM on affective commitment: effects of high performance work systems on
Considering climate strength. Personnel Review, employees’ attitudes. Personnel Psychology, 62,
37, 412–425. 1–29.
Schaufeli, W. B., Salanova, M., Gonzalez-Roma, V., & Takeuchi, R., Lepak, D. P., Wang, H., & Takeuchi, K.
Bakker, A. B. (2002). The measurement of engage- (2007). An empirical examination of the mecha-
ment and burnout: A two sample confirmatory nisms mediating between high-performance
factor analytic approach. Journal of Happiness work systems and the performance of Japanese
Studies, 3, 71–92. organizations. Journal of Applied Psychology, 92,
Schneider, B., Salvaggio, A. N., & Subirats, M. (2002). 1069–1083.
Climate strength: A new direction for climate Wood, S. (2009). HRM and organizational perform-
research. Journal of Applied Psychology, 87, 220–229. ance. In D. G. Collings & G. Wood (Eds.), Human
Schuler, R. S., & Jackson, S. E. (1987). Linking compet- resource management: A critical approach (pp.
itive strategies with human resource management 55–74). London, UK: Routledge Taylor & Francis
practices. Academy of Management Executive, 1, Group.
207–219. Wright, P. M., Gardner, T. M., Moynihan, L. M., & Allen,
Sharma, S., Durand, R. M., & Gur-Arie, O. (1981). M. R. (2005). The relationship between HR practices
Identification and analysis of moderator variables. and firm performance: Examining causal order.
Journal of Marketing Research, 18, 291–300. Personnel Psychology, 58, 409–446.
Shrout, P. E., & Bolger, N. (2002). Mediation in Wright, P. M., & Snell, S. A. (1998). Toward a unify-
experimental and non experimental studies: New ing framework for exploring fit and flexibility in
procedures and recommendations. Psychological strategic human resource management. Academy
Methodology, 7, 422–445. of Management Review, 23, 756–772.
Snell, S. A., Shadur, M. A., & Wright, P. M. (2002). Wright, P. M., & Snell, S. A. (1999). Social capital and
Human resources strategy: The era of our ways. In strategic HRM: It’s who you know. Human Resource
M. A. Hitt, R. E. Freeman, & J. S. Harrison (Eds.), Planning, 22, 62–65.
Handbook of strategic management (pp. 627–649). Wright, P. M., & Van de Voorde, K. (2007). Multi-level
Malden, MA: Blackwell. issues in international HRM: Mean differences,
Strumpf, S. A., Doh, J. P., & Tymor, W. G., Jr. (2010). explained variance, and moderated relation-
The strength of HR practices in India and their ships (CAHRS Working Paper #07-13). Ithaca, NY:
effects on employee career success, performance, Cornell University, School of Industrial and Labor
and potential. Human Resource Management, 49, Relations, Center for Advanced Human Resource
353–375. Studies.
Subramony, M. (2009). A meta-analytic investigation Yang, L. R., Huang, C. F., & Hsu, T. J. (2014). Knowledge
of the relationship between HRM bundles and firm leadership to improve project and organizational
performance. Human Resource Management, 48, performance. International Journal of Project
745–768. Management, 32, 40–53.
Note: This service is not intended for secure transactions such as banking, social media, email, or purchasing. Use at your own risk. We assume no liability whatsoever for broken pages.
Alternative Proxies: