See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/267041329

HRM Systems Between Control and Commitment: Occurrence,

Characteristics, and Effects on HR Outcomes and Firm Performance

Article in Human Resource Management Journal · December 2014

DOI: 10.1111/1748-8583.12054

CITATIONS READS

97 9,000

3 authors:

Sven Hauff Dorothea Alewell

Helmut Schmidt University Hamburg University
88 PUBLICATIONS 2,110 CITATIONS 92 PUBLICATIONS 639 CITATIONS

SEE PROFILE SEE PROFILE

Nina Katrin Hansen

Chemnitz University of Technology
23 PUBLICATIONS 349 CITATIONS

SEE PROFILE

All content following this page was uploaded by Sven Hauff on 16 July 2018.

The user has requested enhancement of the downloaded file.

 HRM Systems Between Control and Commitment:

Occurrence, Characteristics, and Effects on HRM Outcomes and

Firm Performance

Sven Hauff, Dorothea Alewell, Nina Katrin Hansen

Pre-publication version:

Hauff, S./Alewell, D./Hansen, N. K. (forthcoming): HRM Systems Between Control and

Commitment: Occurrence, Characteristics, and Effects on HR Outcomes and Firm
Performance. Human Resource Management Journal
Abstract

In the literature, it is often assumed that traditional, control-oriented HRM systems are

increasingly being replaced by commitment-based HRM systems because the latter generally

result in higher firm performance. However, an HRM system’s effectiveness may depend on

an organization’s external and internal context, and neither control nor commitment HR

systems are without disadvantages. Thus, the empirical validity of this claim is not clear ex

ante. This paper analyzes the empirical diffusion and determinants of control and commitment

HRM systems in Germany as well as their impact on HRM outcomes and firm performance.

The findings indicate that between the two extreme forms of high-control and high-

commitment HRM systems, there are two hybrid forms (long-term-oriented control system

and regulated commitment system) that combine elements of both ‘pure’ systems.

Commitment HRM systems outperform the high-control HRM system concerning many

HRM outcomes and firm performance measures. However, in direct comparison, the high and

the regulated commitment HRM systems do not show substantially different outcomes

indicating that there is no one best way.

Keywords:

HRM systems, HRM strategy, control HRM systems, commitment HRM systems
 Introduction

Human resource management (HRM) practices are usually not used in isolation. Accordingly,

research in strategic HRM focuses on the analysis of bundles or systems of HRM practices

and their effects on HRM outcomes and firm performance (Jiang et al., 2012; Jackson et al.,

2014). In this context, the distinction between control and commitment HRM systems can be

seen as the most prominent and still crucial distinction in many HRM systems approaches

(e.g., Walton, 1985; Osterman, 1987; Begin, 1991; Arthur, 1992; 1994; Lepak and Snell,

1999; 2002; Verburg et al., 2007; Toh et al., 2008).

Control HRM systems are often seen as a traditional HRM approach which are supposed to be

unappropriate in a modern economic environment (Walton, 1985; Kaufman, 2013; Waldman,

2013). Instead, some authors assume that high commitment HRM systems are universally

valid best practices which effect competitive advantages positively (Delery and Doty, 1996;

Martín-Alcázar et al., 2005). However, within the contingency perspective of strategic HRM,

authors have argued that an HRM system’s effectiveness depends on an organization’s exter-

nal and internal context, including industry specifics, labor market conditions, size, life cycle,

or business strategy (Jackson and Schuler, 1995; Jackson et al., 2014).

Context dependency does not only refer to the HRM system per se, but also to the question

which components of HRM systems are adequate in certain contexts (Boxall and Macky,

2009). For example, Baron and Kreps (1999: 195) state that “there is no single blueprint for a

high-commitment HR system; organizations will pick and choose [...] high-commitment HR

practices, according to their own needs, circumstances, and desires.” Beer et al. (1984) also

indicate that organizations might use a mixture of different HRM approaches. Thus, the exist-

ence of hybrid systems is to be expected, as a mixture between control and commitment prac-

tices may be favorable in some contexts (Su and Wright, 2012).

1
Following these ideas, we analyze the empirical diffusion and determinants of control and

commitment HRM systems in Germany as well as their impact on HRM outcomes and firm

performance. Our study contributes to the literature in several ways: First, by applying latent

class analysis (LCA) (e.g., Lazarsfeld and Henry, 1968), we use an advanced clustering pro-

cedure, which allows us to empirically identify distinct HRM systems on the control-

commitment continuum. Since the approach is open to any possible combination of control

and commitment HRM practices, it provides a more nuanced picture of existent HRM sys-

tems. Second, we demonstrate the empirical diffusion of control and commitment systems in

a leading European economy, something still lacking in literature. Third, we also shed light

on the determinants of different HRM system types. Thus, we address the rather neglected

question of why and where different HRM systems are used (Jackson et al. 2014). Finally, we

compare the outcomes of distinct HRM systems adding evidence to the yet unresolved ques-

tion which combinations of HRM practices have the greater impact on HRM outcomes and

firm performance (Guest, 2011).

Theoretical Background

Control and Commitment HRM Systems

Research on HRM systems has been thriving for some time (e.g., Lepak et al., 2006; Guest,

2011; Alewell and Hansen, 2012; Jackson et al., 2014). The basic idea in this research is that

consistent or complementary bundles of HRM practices may have large effects on the

achievement of HRM objectives and firms’ economic performance, while changes in single

HRM activities or the implementation of inconsistent HRM instruments, policies, and practic-

es have little, no positive or even negative effects on these goals (Jiang et al., 2012).

Authors from different disciplines have presented distinct typologies of HRM systems (for a

comprehensive review see Kaufman, 2013). A long standing tradition can be found in indus-

2
trial relations research, where different forms of workforce management are mostly discussed

under the heading of ‘employment systems’ (e.g., Kerr, 1954; Osterman, 1987; Begin, 1991;

Arthur, 1992; 1994). Labor process literature also distinguishes different models of labor con-

trol (e.g., Edwards, 1979). More recently, the idea of HRM systems has found its way into

management research (e.g., Beer et al., 1984; Walton, 1985; Delery and Doty, 1996; Lepak

and Snell, 1999; 2002; Verburg et al., 2007; Toh et al., 2008).

In an early review, Dyer and Reeves (1995: 658) pointed out that these typologies tend to “ar-

ray in reasonably comparable continua from, in the authors’ terms, ‘Traditional’, ‘Control’

and ‘Mass Production’ strategies on one end to ‘Innovative’, ‘Commitment’ and ‘Flexible (or

Lean) Production’ strategies on the other.” In the meantime, there has been considerable pro-

gress in the theory of HRM systems (e.g., Guest, 2011; Jackson et al., 2014), and several ty-

pologies have been presented. However, the dichotomy between control HRM systems and

commitment HRM systems still represents a mayor distinction in these approaches (Bae and

Lawler, 2000; Lepak et al., 2006; Verburg et al., 2007).

Even if there is still no agreement between researchers concerning the question which specific

HRM practices belong to control and commitment HRM systems, they can generally be char-

acterized as follows: The control HRM system roots in Taylor’s “scientific management” and

seeks to “establish order, exercise control, and achieve efficiency in the application of the

work force” (Walton, 1985: 78). It aims to increase labor efficiency and to decrease direct

labor costs by improving performance standards, on the one hand, and defining simple and

well-defined job tasks, on the other (Arthur, 1994). High labor specialization and strong divi-

sion of work with a deep hierarchy of narrow and well-defined jobs correspond to a top-down

and centralized decision-making process. Skills demands are low, there is little training, and

employees have no employment security (Walton, 1985). The required performance standards

are accomplished by the implementation of formal rules and procedures as well as close

3
monitoring by supervisors, which secures employee compliance with organizational goals.

Additionally, compensation practices are based on “measurable output criteria” (Arthur, 1994:

672). Control systems are generally designed to “minimize the impact of labor on the labor

process. Thus, workers are more commodity-like and more replaceable” (Lepak et al., 2006:

226) than in other HRM systems.

In contrast, the commitment HRM system facilitates the convergence of employee interests

and organizational goals (Arthur, 1994). Organizations that rely on a high-commitment strat-

egy aim “at getting more from workers by giving more to them” (Baron and Kreps, 1999:

189). Selective hiring procedures seek to find the “right”, broadly skilled employees. Based

on a flexible job design, employees enjoy comparatively high involvement in managerial de-

cisions. Long-term employment perspectives are combined with intensive training efforts,

team-based production systems, job rotation, and regular quality circles. Furthermore, com-

pensation systems are based on firm-wide, unit, and team performance measures, and include

efficiency wages and superior benefits, with merit as a central pay criterion for all staff

(Arthur, 1994; Baron and Kreps, 1999; Wright and Kehoe, 2008).

Does One Size Fit All?

In his seminal work, Walton (1985: 77) states that in “factory after factory there is a revolu-

tion under way in the management of work”, that is, that control HRM systems are being

comprehensively superseded by high-commitment HRM systems because the latter contribute

to a higher firm performance. This notion is supported by related universalistic approaches,

such as high-involvement HRM systems (Lawler, 1986) or high-performance HRM systems

(Huselid, 1995; Appelbaum et al., 2000), and was confirmed in several studies (Combs et al.,

2006),

However, other authors have criticized these universalistic approaches (Baron and Kreps,

1999; Godard, 2004; Nishii et al., 2008; Boxall and Macky, 2009; Guest, 2011; Atkinson and
4
Lucas, 2013) because they are associated with direct costs through training activities, higher

wages, and employment guarantees. They may also lead to longer decision processes as well

as an increased stress level.

Further criticism comes from proponents of the contingency approach of strategic HRM who

assume that the effects of HRM systems depend on the specific internal and external context

(Delery and Doty, 1996; Martín-Alcázar et al., 2005). Industry represents an important exter-

nal factor that influences a firm’s choice between HRM systems (Jackson and Schuler, 1995;

Jackson et al., 2014). In manufacturing industries, production must often follow predefined

processes and will thus depend on clear structures and routines implemented in the process.

Therefore, it can be assumed that control-oriented HRM systems are more often used in man-

ufacturing industries. A related question is whether daily business is influenced by strong or-

der fluctuations. Firms may have differing flexibility demands (Osterman 1987) and it can be

assumed that commitment HRM systems are less often used in firms with strong order fluctu-

ations.

Firm size may also have an impact on the choice of a HRM system (Sels et al., 2006; Fabi et

al., 2007; Bhatt and Reddy, 2011). As larger organizations receive more public attention, the

pressure to implement commitment-based HRM systems may be higher, because these are

interpreted as ‘high road’ strategies with better working conditions (Kalleberg, 2011). Further,

only large organizations may be capable of bearing the cost of high-commitment HRM sys-

tems (Jackson and Schuler, 1995). Related to firm size are life cycle and owner structure. Re-

garding a firm’s lifecycle, growing young firms may focus on recruitment and selection, while

more mature firms are concerned with implementing internal labor markets (Jackson and

Schuler, 1995). Owner structure might be another influencing factor. Helfen and Schuessler

(2009) have shown that family-owned small businesses have a significantly lower probability

5
of having a works council. Thus, family ownership may be associated with stronger resistance

to allow employee interference in decisions.

The role and resources of the HRM department should also shape the HRM system. As the

commitment-based HRM strategy requires high attention to implement and maintain HRM

functions, firms implementing this strategy should have some HRM professionalism. This

involves a good allocation of resources for HRM and a strategic orientation by the HRM de-

partment. Given differing characteristics of jobs and employee groups, professional HRM

should be related to the application of differentiated HRM for different groups of employees.

Another important factor is employee skills and qualifications. Lepak and Snell (1999; 2002)

point out that the choice of a HRM system depends on the strategic value and uniqueness of

human capital associated with different groups of employees. Accordingly, control-oriented

HRM systems are more appropriate for employees with human capital that is low in strategic

value and uniqueness. In contrast, if human capital is very valuable and unique, the focus

should be on commitment-based HRM systems.

Finally, collective employee representation through unions, works councils, or other forms of

institutionalized employee representation should also have an impact on the HRM system.

Unions and works councils give a voice to employees and are known to influence working

conditions for the better (Jackson and Schuler, 1995). We thus assume that they have a posi-

tive impact on transforming the traditional, control-oriented HRM system into a more com-

mitment-oriented system (Su and Wright, 2012).

In short, differing HRM systems will be applied in different contexts. Furthermore, it is not

only the choice of a basic system type that may be influenced by the context, but also the

choice of the specific components within a basic HRM system type (Baron and Kreps, 1999),

thus giving rise to hybrid forms as specific combinations of the basic control and commitment

types of HRM systems (Beer et al., 1984; Delery and Doty, 1996). This has been empirically
6
supported by Su and Wright (2012), who were able to show that in the Chinese context a hy-

brid system – including both commitment and control HRM practices – is more effective than

high-commitment HRM systems.

Building on these reflections and findings, we analyze the empirical diffusion of control and

commitment HRM systems and the impact of these different HRM systems on HRM out-

comes and firm performance. We choose a more or less explorative approach, because our

analysis has to be open to the ability of combinations of control and commitment HRM prac-

tices. However, an empirically based taxonomy can serve as a starting point for further theo-

retical considerations.

Data, Measures, and Methods

Data

The following analysis is based on data collected via highly structured computer-aided tele-

phone interviews with chief executives and human resource managers of firms in Germany.

The data collection was conducted in 2012 and aimed at firms with at least 20 employees in

the following sectors: chemicals and pharmaceuticals, mechanical engineering, banking and

insurance, and professional services (especial legal, accounting, and business consultancy

activities). We have chosen this sector distribution to have the opportunity to analyze a com-

paratively large number of firms in the service segment as well as in classical industrial seg-

ments, and to have two sectors each within service and industrial segments. The number of

randomly sampled firms in these sectors was 5,388. Contact information was drawn from the

database of the German Chamber of Industry and Commerce. Of the contacted firms, 1,175

took part in the study, which left us with a satisfying response rate (21.8%). However, a first

analysis of the data revealed that 76 firms did not meet the selection criteria (size and indus-

try) or provided invalid answers. Thus, the complete usable data is available for 1,099 firms;

7
we thereby interviewed one person in each firm. The sample was disproportionally stratified

for industry, resulting in an approximately uniform distribution (23.9% chemicals and phar-

maceuticals, 24.7% mechanical engineering, 28.0% banking and insurance, and 23.3% pro-

fessional services). However, based on the distribution in the German Chamber of Industry

and Commerce database, we are able to weight the data, resulting in the weighted data being

representative for all German firms with the aforementioned criteria. The following analyses

are estimated with weighted data only.

Measurement of Variables

A broad spectrum of HRM practices was captured in the interviews. In firms that stated that

they differentiated their HRM for different groups of employees, all the questions relating to

HRM practices refer to the group of employees most important for the firm’s economic suc-

cess (as suggested by Osterman, 1994; see also Delery and Doty, 1996). If HRM did not dif-

ferentiate between different groups of employees, questions on HRM practices were formu-

lated such that they encompassed all employees. Thus, our results apply to one HRM system

within a firm, even if this organization operates with more than one HRM system for differing

groups of employees. If a firm operates with only one HRM system, this system encompasses

all the employees. The HRM systems analyzed are comparable in that they apply to the group

of employees that is most important to the firm’s economic success.

For the following analyses, we included only HRM practices that were mentioned by various

authors as typical practices in either control or commitment HRM systems (Table 1). As far as

possible, we used items already tested in the literature (e.g., Lepak and Snell, 2002). For all

items statements such as “These employees perform jobs that are well defined” were present-

ed. Response categories ranged from 1 for does not apply at all to 5 for fully applies.

8
 ------------------------------------------------
TABLE 1 ABOUT HERE
------------------------------------------------

To analyze the impacts of different HRM systems on HRM-level and firm-level outcomes, we

further gathered data on several HRM (e.g., personnel cost, performance level) and firm out-

comes (e.g., opening up of new markets, sales growth). Again, items were measured accord-

ing to the degree of agreement to different statements (1 for does not apply at all to 5 for fully

applies).

When analyzing the empirical diffusion of control and commitment HRM systems and their

effects, the following control variables were included: firm size, industry, firm age, family-

owned (yes/no), endowment of the HRM department (such as number of employees per em-

ployee in the personnel department), strategic orientation of HRM (yes/no), differentiated

HRM for different groups of employees (yes/no), share of highly skilled employees, existence

of collective employee representation, binding legal commitment to or orientation towards

collective bargaining agreements, and strong order fluctuations (yes/no).

Methods

To identify control and commitment HRM systems, latent class analysis (LCA) (e.g.,

Lazarsfeld and Henry, 1968) was used. LCA represents a classification method that seeks to

identify unobserved subgroups. Compared to traditional clustering procedures (e.g., K-

means), the LCA approach has several advantages (Magidson and Vermunt, 2002); these in-

clude: (1) LCA is a model-based approach in which the membership probabilities are estimat-

ed by maximum likelihood methods. (2) Several diagnostics are provided (e.g., BIC statistic)

to determine the number of clusters. Thus, the choice of cluster is less arbitrary than in tradi-

tional clustering procedures. (3) LCA is more flexible regarding the scaling of the observed

variables (e.g., no need for equal variances and zero-order correlations). (4) Variables with
9
mixed scale type (continuous, ordinal, nominal, or counts) can be integrated. (5) Covariates

can be included directly into the clustering procedure.

LCA allows for the assignment of cases to different latent classes based on the highest a pos-

teriori probability. This information can be used for further analysis. We analyzed the effect

of several HRM systems on HRM outcomes and firm performance by applying ordered logit

regression analysis. Ordered logit models are used to predict ordinal variables and are thus

suitable for the HRM outcome and firm performance measures included in our data set. The

procedure is similar to logistic regression analysis, with the difference that the depended vari-

able has more than two categories, which can be brought into a certain order (Agresti, 2002).

Results

HRM Systems and their Characteristics

Within LCA several statistical criteria can be used to identify the appropriate number of latent

classes (e.g., Collins and Lanza, 2010; Nylund et al., 2007). Thereby, relative fit measures are

commonly used measures to evaluate model fit. Information criteria, such as the Aiken infor-

mation criterion (AIC) and the Bayesian information criteria (BIC), are particularly useful

tools where models with lower AICs and BICs represent a better model fit. Unfortunately,

these information criteria do not always concur in identifying the model with the best fit

(Collins and Lanza, 2010). Based on an intensive simulation study, Nylund et al. (2007) sug-

gested the BIC to identify the right number of clusters. However, it is also known that the BIC

criteria may be too conservative and the AIC too liberal in deciding the right number of clas-

ses (McLachlan and Peel, 2000).

In addition to the information criteria, the Lo-Mendell-Rubin test (LMRT) and the bootstrap

likelihood ratio test (BLRT) can also be used to assess relative model fit (Collins and Lanza,

2010; Nylund et al., 2007). Both tests provide a p value in order to decide whether a k class

10
solution represents the data better than a k-1 class solution, whereas the BLRT performs better

in identifying the correct number of latent classes (Nylund et al., 2007). Besides these statisti-

cal criteria, decisions about model fit and model selection should also rely on parsimony (i.e.,

simpler models are preferred to more complex models) as well as model interpretability

(Collins and Lanza, 2010).

Table 2 presents a summary of the statistics from the LCA. Unfortunately, the BIC and AIC

results point to different numbers of classes. The BIC points toward a three-cluster solution,

whereas the AIC suggests a five-cluster solution. As the information criteria do not yield an

unequivocal solution, parsimony and model interpretability become necessary selection crite-

ria. Thereby, we decided to use the four-cluster solution because it provides an additional in-

terpretable cluster compared to the three-cluster solution. At the same time, it is more parsi-

monious than the five-cluster solution in which no interpretable additional cluster is repre-

sented. This decision is supported by the fact that the BLRT provides a highly significant re-

sult, indicating that the four-class solution better fits the data than the three-class solution.

To analyze the robustness and validity of our cluster solution, we conducted several robust-

ness checks. A common procedure for cluster analysis is to split the sample along different

criteria (e.g., Hair et al., 2010). Therefore, we used firm size levels (20 to 99, ≥ 100 employ-

ees) as well as the question if firms differentiated their HRM for different groups of employ-

ees, and repeated the analysis. In every case, the analyses revealed very similar cluster pro-

files, which confirms the validity of our cluster solution. The cluster solution’s validity is fur-

ther supported by its predictive validity, because the different clusters are distinctively related

to HRM outcomes and firm performance (see below).

------------------------------------------------
TABLE 2 ABOUT HERE
------------------------------------------------

11
Figure 1 shows the profile data of the four-cluster solution. The profile data contain the prob-

ability that a firm in a particular cluster is characterized by a certain HRM practice. In addi-

tion, Table 3 presents the profile data regarding different firm characteristics. Based on this

information, the four clusters can be characterized as follows:

Cluster 1: High-control system

Within this cluster, we find the lowest employee discretion and also rather low communica-

tion and trust between management and employees. Neither training activities nor long-term

perspectives are provided within the firm. Thus, the first cluster comes closest to the pure

control system as described in the literature. However, this HRM system is not necessarily

only used for well-defined jobs that are governed by clear rules and procedures. Concerning

firm characteristics, firms belonging to this cluster can be found especially in small and medi-

um-sized firms in mechanical engineering, which are often family-owned. Qualification levels

in these firms are fairly low. The control system is used in firms with a comparatively low

allocation of human resources to the HRM department, with no strategic role and no differen-

tiated HRM for different groups of employees. The first cluster is also characterized by rela-

tively high order fluctuations.

------------------------------------------------
FIGURE 1 ABOUT HERE
------------------------------------------------

Cluster 2: Long-term-oriented control system

The second cluster resembles the first in terms of employee discretion, communication, and

trust. But in contrast to the first cluster, employees are continuously trained and have internal

career perspectives as well as relatively long-term employment perspectives. Jobs are rather

well defined and governed by clear rules and procedures. We have labeled this cluster long-

term-oriented control system. This HRM system is often found in larger firms, especially
12
those in banking and insurance. The firms in this cluster are also older and mostly not family-

owned. Compared to firms in the high-control system, these firms have a slightly higher per-

centage of highly skilled employees.

Cluster 3: High-commitment system

The third cluster can be characterized as a high commitment HRM system as described in the

literature. Work includes a great variety of tasks and is not governed by rules and procedures.

There is an open communication and intense exchange of information between management

and employees. Employees can implement changes in the way they perform their jobs and are

empowered to make decisions. Accordingly, there is a high level of trust between employees

and employer. Employment is fairly long-term-oriented, involving continuous training activi-

ties, internal career perspectives, and employment security. Work is often performed in teams.

However, incentive components are nonetheless based on individual criteria. The commit-

ment system is more often used in small professional service firms, which are mostly quite

young. These firms have a high share of skilled employees. HRM within these firms is strate-

gically oriented, and different groups of employees are treated differently. Legal forms of

employee participation are used comparatively less frequently.

Cluster 4: Regulated commitment system

The profile of the fourth cluster is similar to the profile of the third: Employees work in

teams, and communication is open and intensive. Relationships are built on trust and em-

ployment, which is long-term oriented. However, there are also some differences between the

two clusters because, within the forth cluster, employee discretion is much lower. And, more

remarkable, jobs are well defined and mostly governed by clear rules and procedures. There-

fore, we have labeled the fourth cluster a regulated commitment system. Another difference is

that even if employment is long-term-oriented, there are comparatively fewer training activi-

ties. The regulated commitment system can be found in smaller firms with low endowment of
13
the HRM department, especially in chemicals/pharmaceutics and mechanical engineering.

This HRM system tends to be used for less qualified employees.

------------------------------------------------
TABLE 3 ABOUT HERE
------------------------------------------------

Outcomes of Different HRM Systems

As shown in the last section, the four HRM system types can be linked to different firm char-

acteristics. Because this hints at different internal and external contexts, one may expect that

firms have chosen their HRM systems because it is more effective in their specific context. To

shed more light on this assumption, several regression analyses were conducted to determine

the effect of the different HRM systems on HRM outcomes and firm performance. To enable

a detailed comparison between the outcomes of different HRM systems, regression analyses

were repeated using different reference categories. Thus, each of the identified HRM systems

can be compared to the three other systems.

------------------------------------------------
TABLE 4 ABOUT HERE
------------------------------------------------

Table 4 contains the results for HRM outcomes. The choice of the HRM system has no effect

on personnel costs. However, affiliation to a cluster has strong effects for other HRM out-

comes: the long-term oriented control systems performs better than the high-control systems

in regard to the endowment with qualified personnel, up-to-date knowledge, performance lev-

el, cost and quality orientation of employees, motivation, job satisfaction, and commitment.

Except for costs, both commitment HRM systems outperform the high-control system for all

integrated HRM outcomes. The same applies for the comparison between the high-control

system and the long-term oriented control system. However, the differences between the regu-
14
lated commitment system and the long-term oriented control system are less pronounced. The

comparison between the two commitment HRM systems shows that the high-commitment

system only performs better in terms of motivation, job satisfaction and commitment. For all

other HRM outcomes, there are no significant differences between the high-commitment and

the regulated commitment system.

The effects on firm outcomes are depicted in Table 5. The long-term oriented control system

performs slightly better than the high-control system in terms of sales growth, productivity,

quality of products, brand strength, return on equity, and business development. Again, the

commitment HRM systems outperform the high-control system for most of the variables ana-

lyzed. The high-commitment system mostly performs better compared to the long-term-

oriented control system. However, there is one negative effect of the high-commitment sys-

tems that concerns realization of low prices in relation to competitors. Again, the differences

between the regulated commitment system and the long-term oriented control system are less

pronounced but the regulated commitment system still outperforms the long-term oriented

control system. The comparison between the two commitment HRM systems shows no signif-

icant differences.

------------------------------------------------
TABLE 5 ABOUT HERE
------------------------------------------------

15
 Discussion

By empirically identifying the high-control and the high-commitment system, our analysis

confirms the basic distinction made in many HRM systems approaches. The high-control ori-

ented system is characterized by the lowest employee discretion, low qualifications and short

term orientation – practices which have earlier been associated with the control (Walton,

1985; Arthur, 1992; 1994), secondary (Osterman, 1987), simple (Begin, 1991), market

(Delery and Doty, 1996; Verburg et al., 2007), compliance (Lepak and Snell, 1999; 2002) or

cost minimizers systems (Toh et al., 2008). In contrast, the high-commitment system refers to

tasks variety, absence of strict rules, high qualifications and long term perspectives which

corresponds to the commitment (Walton, 1985; Arthur, 1992; 1994; Lepak and Snell, 1999;

2002), salaried (Osterman, 1987), adhocracy (Begin, 1991), professional (Verburg et al.,

2007) and commitment maximizers systems (Toh et al., 2008).

In addition, we have found that there are two hybrid systems which combine elements from

both the control and commitment systems. The first hybrid system is labeled the long-term-

oriented control system, indicating that this system is basically control-oriented, but aims at

achieving long-term employment relationships with continuous training and internal career

perspectives, too. Quite similar systems have been described by Osterman (1987) as ‘industri-

al model’ and Begin (1991) as ‘machine bureaucracy’, which are control orientated but do not

necessarily exclude long-term employment perspectives (even if these are not guaranteed). In

contrast to these systems, training activities are of high importance within the long-term ori-

ented control systems, indicating that firms that apply these rather traditional systems may

nevertheless feel a strong need to keep their employees’ knowledge up to date.

The second hybrid system is labeled a regulated commitment system. It combines key ele-

ments of commitment HRM systems such as teamwork, open and intense communication, and

high trust, with control elements such as low employee discretion, well-defined jobs, and high
16
influence of rules and procedures. It thus points to potential limits of the commitment HRM

strategy: even if high trust employers are willing to hand over responsibility to employees

there may be natural constraints implemented in the work process itself. So far, this constella-

tion has not been discussed in existent HRM systems approaches.

By identifying these two hybrid HRM systems, our analyses show that the dichotomy be-

tween control and commitment HRM systems previously debated in the literature (e.g.,

Walton, 1985, Arthur, 1992; 1994) is too narrowly defined. This is not trivial, given the fact,

that most studies in strategic HRM emphasize high-commitment, high-involvement or high-

performance HRM systems (Jackson et al., 2014). The notion of these systems always im-

plies, explicitly or implicitly, its extreme counterparts, but neglects possible distinct combina-

tions along the control-commitment continuum. Herewith, our results also provide further

reasons for the critique of common measurement approaches (for a detailed critique see, e.g.,

Jiang et al., 2012): additive scales, indices or factor solutions assume that all practices load on

a single factor respectively correlate with each other. In contrast, our results show that there

are specific combinations of practices along the control-commitment continuum.

Our results show similarities with previous HRM-systems approaches. However, the great

variety of conceptual and empirical approaches (Kaufman, 2013) renders it difficult to com-

pare our findings with all existent typologies. With regard to workforce management we did

not look at how production can be organized (e.g., make-or-buy), or how labor markets or

organizations are structured (e.g., centralized, decentralized, bureaucratic), though these are

elements in some employment systems approaches (e.g., Kerr, 1954; Beer et al., 1984; Begin,

1991; Verburg et al., 2007). Furthermore, we do not distinguish ex ante between different

groups of employees. Thus, instead of asking how different groups (e.g., professionals) are

managed – as done by Lepak and Snell (1999; 2002) and implicitly also by Osterman (1987)

– we assumed that a specific combination of HRM practices is applicable for different groups

17
of employees. Finally, different authors have used different HRM practices to characterize

HRM systems. For example, compared to our approach, Toh et al. (2008) focused more on

recruitment, training and benefit practices and thus identified different HRM systems (contin-

gent motivators, competitive motivators and resource makers).

Our findings further show that the ‘revolution’ described by Walton (1985) – assuming it is

indeed underway – has not yet been completed in Germany: the two control HRM systems

together represent 61% of the firms analyzed and thus a majority of all firms. This seems sur-

prising, especially for Germany. Within the comparative capitalism literature (Hall and

Soskice, 2001), Germany is often seen as the role model for coordinated market economies,

which are characterized by a high influence of employee representatives, a high level of regu-

lation and employment protection, fostering the expectation of comparatively good work and

employment conditions and a high share of commitment-oriented HRM systems. But appar-

ently, this is not the case in our findings.

This is even more surprising as our findings indicate that commitment-oriented HRM systems

generally perform better than control-oriented HRM systems. The – given Walton’s (1985)

expectations – low rate of commitment HRM systems can thus not be explained by perfor-

mance arguments for specific contexts. Our results therefore contradict the potential explana-

tion that HRM systems’ effectiveness is context specific, and rather support universalistic

approaches which point to the general and universalistic positive influence of specific best

practices (Delery and Doty, 1996). Explanations for the relatively low share of firms with

commitment HRM systems must therefore be sought elsewhere. Baron and Kreps (1999) state

that a central obstacle to the implementation of commitment HRM systems might be the costs

connected with these systems (e.g. training, wages, and employment security). However, our

findings do not support the claim that personnel costs are necessarily higher in commitment

HRM systems. Alternative explanations may be found in organizational constraints which

18
might hinder the implementation of commitment HRM systems. Firms that depend on prede-

fined processes and structures may not be able to apply commitment-oriented HRM systems.

Another important factor is the skill and qualification structure: a prerequisite for responsibil-

ity and independent work is that employees have the necessary skills and qualifications

(Jackson and Schuler, 1995; Jackson et al., 2014). In this regard, we find that high-control

systems are more frequently used in manufacturing industries and for employees with rather

low qualification.

Other plausible causes for the comparatively low diffusion of commitment HRM systems

might be the resistance by senior management, employees or their representatives. Commit-

ment-based HRM systems require that the management is willing to give up some of its pow-

er and influence, which might not always be the case (Walton, 1985). Employees may be crit-

ical with regard to high-commitment systems, because more responsibility for individual em-

ployees may increase workload and stress (Godard, 2004). Employee representatives might

hinder the implementation of pure commitment systems, because stronger individual voice

may decrease the power bases of employee representations (Baron and Kreps, 1999). Again,

our findings provide some evidence regarding these assumptions. For example, the high per-

centage of family-owned firms using control HRM system supports the notion of management

resistance because especially owners may want to retain their prerogatives. Control systems

are more likely in firms with works councils and a certain degree of affiliation to collective

bargaining agreements, supporting the notion of a negative influence of employee representa-

tion on high-commitment HRM systems. But although there is some initial evidence, with our

data we cannot thoroughly disentangle the reasons for not implementing commitment HRM

systems.

Finally, the detailed comparison of the outcomes of all four HRM systems reveals that the two

commitment-oriented systems slightly differ in their outcomes. High-commitment systems

19
only perform better in terms of motivation, satisfaction and commitment, but we found no

differences for firm performance. Thus, different combinations of HRM practices possibly

lead to similar outcomes – a result contradicting the widespread assumption that there is only

one best way of HRM. Furthermore, it highlights again that common measurement approach-

es relying on additive scales, indices or factor solutions might be misleading as they focus on

a continuum instead of distinct combinations of HRM practices.

Our results and conclusions are derived under certain limitations. First, we have used a cross-

sectional instead of a longitudinal study design. Thus, the stated effects cannot be traced back

to causal relationships. Future studies should, therefore, use longitudinal data to study these

connections in greater detail. Second, our performance measures were perceptional measures.

Several researchers have been able to show that there are strong statistic relationships between

perceptual and hard performance measures (e.g., Delaney and Huselid, 1996; Bae and Lawler,

2000; Su and Wright, 2012). Nevertheless, future studies should also integrate objective

measures and should confront them with subjective ones. Additionally, we gathered answers

from a single management respondent. The usual caveats apply, as these subjective, single-

person, management-sided performance measures could be biased. Third, we have only ana-

lyzed the German case and did not scrutinize the international and comparative perspective on

HRM systems in the context of differing institutional contexts. Future research may thus ex-

plore whether the hybrid systems and the performance effects we have found for Germany

will appear in other countries. Fourth, we only asked about the presence of HRM practices,

but have no information about the degree of implementation of these practices in any single

firm. This might bias our results.

20
 Conclusion

With this paper, we contribute to the literature on HRM systems by analyzing the empirical

diffusion of two ideal types of HRM systems, namely the high-control HRM system, on the

one hand, and the high-commitment HRM system, on the other. Our results extend previous

studies through the identification of two hybrid systems between these ideal types, the long-

term-oriented control system and the regulated commitment system. By integrating a broad

set of external internal context variables we shed further light on the antecedences of different

HRM systems. Analyzing the impact of the four identified HRM systems on HRM outcomes

and firm performance, we found that the commitment-oriented HRM systems outperform the

high-control oriented system in most of the dimensions included. Therefore, our findings sup-

port the idea of specific HRM practices enhancing firm performance in different contexts.

However, in comparison, the two commitment systems identified did not show significant

differences regarding most of the HRM performance outcomes. As far as firm performance

measures are concerned, there were no significant differences at all. This indicates that there

is not only one best way for managing human resources.

21
 References

Agresti, A. (2002). Categorical data analysis (2. ed. ed.), Hoboken, NJ: Wiley-Interscience.
Alewell, D. and Hansen, N.K. (2012). 'Human Resource Management Systems - A Structured
Review of Research Contributions and Open Questions'. The German Journal of
Industrial Relations, 19: 2, 90–123.
Appelbaum, E., Bailey, T., Berg, P. and Kalleberg, A.L. (2000). Manufacturing advantage:
Why high performance work systems pay off, Ithaca, NY: Cornell Univ. Press.
Arthur, J.B. (1992). 'The link between business strategy and industrial relations systems in
American steel minimills'. Industrial and Labor Relations Review, 45, 488–506.
Arthur, J.B. (1994). 'Effects of Human Resource Systems on Manufacturing Performance and
Turnover'. The Academy of Management Journal, 37: 3, 670–687.
Atkinson, C. and Lucas, R. (2013). 'Worker responses to HR practice in adult social care in
England'. Human Resource Management Journal, 23: 3, 296–312.
Bae, J. and Lawler, J.J. (2000). 'Organizational and HRM Strategies in Korea: Impact on Firm
Performance in an Emerging Economy'. Academy of Management Journal, 43: 3,
502–517.
Baron, J.N. and Kreps, D.M. (1999). Strategic human resources: Frameworks for general
managers, New York: John Wiley.
Beer, M., Spector, B., Lawrence, P.R., Mills, D.Q. and Walton, R.E. (1984). Managing
human assets, New York; London: Free Press Collier Macmillan Publishers.
Begin, J.P. (1991). Strategic Employment Policy: An Organizational Systems Perspective,
Englewood Cliffs.
Bhatt, P. and Reddy, S.C. (2011). 'Exploring HRM Practices in SMEs'. International Journal
of Research in Commerce, Economics and Management, 1: 5, 2–41.
Boxall, P. and Macky, K. (2009). 'Research and theory on high-performance work systems:
Progressing the high-involvement stream'. Human Resource Management Journal, 19:
1, 3–23.
Collins, L.M. and Lanza, S.T. (2010). Latent class and latent transition analysis. With
applications in the social, behavioral, and health sciences, Hoboken, NJ: Wiley.
Combs, J., Lui, Y., Hall, A. and Ketchen, D. (2006). 'How much do High-Performance Work
Practices matter? A meta-analysis of their effects on organizational performance'.
Personnel Psychology, 59: 3, 501–528.
Delaney, J.T. and Huselid, M.A. (1996). 'The Impact of Human Resource Management
Practices on Perceptions of Organizational Performance'. Academy of Management
Journal, 39: 4, 949–969.
Delery, J.E. and Doty, D.H. (1996). 'Modes of theorizing in strategic human resource
management: tests of universalistic, contingency, and configurational performance
predictions'. Academy of Management Journal, 39, 802–835.
Dyer, L. and Reeves, T. (1995). 'Human resource strategies and firm performance: what do
we know and where do we need to go?'. The International Journal of Human
Resource Management, 6: 3, 656–670.
Edwards, R. (1979). Contested terrain. The transformation of the workplace in the twentieth
century, New York: Basic Books.
Fabi, B., Raymond, L. and Lacoursière, R. (2007). 'HRM Practice Clusters in Relation to Size
and Performance: An Empirical Investigation in Canadian Manufacturing SMEs'.
Journal of Small Business & Entrepreneurship, 20: 1, 25–39.
Godard, J. (2004). 'A Critical Assessment of the High-Performance Paradigm'. British Journal
of Industrial Relations, 42: 2, 349–378.

22
Guest, D.E. (2011). 'Human resource management and performance: still searching for some
answers'. Human Resource Management Journal, 21: 1, 3–13.
Hair, J.F., Black, W.C., Babin, B.J. and Anderson, R.E. (2010). Multivariate data analysis: A
global perspective (7. ed. ed.), Upper Saddle River, NJ: Pearson.
Hall, P.A. and Soskice, D.W. (2001). Varieties of capitalism: The institutional foundations of
comparative advantage, Oxford: Oxford University Press.
Helfen, M. and Schuessler, E.S. (2009). 'Uncovering Divergence: Management Attitudes
towards HRM Practices and Works Council Presence in German SMEs'. Economic
and Industrial Democracy, 30: 2, 207–240.
Huselid, M.A. (1995). 'The Impact of Human Resource Management Practices on Turnover,
Productivity, and Corporate Financial Performance'. The Academy of Management
Journal, 38: 3, 635–672.
Jackson, S.E. and Schuler, R.S. (1995). 'Understanding human resource management in the
context of organizations and their environments'. Annual Review of Psychology, 46: 1,
237–264.
Jackson, S.E., Schuler, R.S. and Jiang, K. (2014). 'An Aspirational Framework for Strategic
Human Resource Management'. The Academy of Management Annals, 8: 1, 1–56.
Jiang, K., Lepak, D.P., Han, K., Hong, Y., Kim, A. and Winkler, A.-L. (2012). 'Clarifying the
construct of human resource systems: Relating human resource management to
employee performance'. Human Resource Management Review, 22: 2, 73–85.
Kalleberg, A.L. (2011). Good jobs, bad jobs: The rise of polarized and precarious
employment systems in the United States, 1970s to 2000s, New York: Russell Sage
Foundation.
Kaufman, B.E. (2013). 'The economic organization of employment: systems in human
resource management and industrial relations'. In A. Grandori (Ed.), Handbook of
Economic Organization, Cheltenham: Edward Elgar.
Kerr, C. (1954). 'The Balkanization of labor markets'. In E.W. Bakke (Ed.), Labor, mobility
and economic opportunity, Cambridge, Mass.: M.I.T. Press.
Lawler, E.E. (1986). High-involvement management: Participative strategies for improving
organizational performance, San Francisco, Calif: Jossey-Bass.
Lazarsfeld, P.F. and Henry, N.W. (1968). Latent structure analysis, New York [u.a.]:
Houghton Mifflin.
Lepak, D.P., Liao, H., Chung, Y. and Harden, E.E. (2006). 'A Conceptual Review Of Human
Resource Management Systems In Strategic Human Resource Management Research'.
Research in Personnel and Human Resource Management, 25, 217–271.
Lepak, D.P. and Snell, S.A. (1999). 'The human resource architecture: toward a theory of
human capital allocation and development'. Academy of Management Review, 24, 31–
48.
Lepak, D.P. and Snell, S.A. (2002). 'Examining the Human Resource Architecture: The
Relationships Among Human Capital, Employment, and Human Resource
Configurations'. Journal of Management, 28: 4, 517–543.
Magidson, J. and Vermunt, J.K. (2002). 'Latent class models for clustering: A comparison
with K-means'. Canadian Journal of Marketing Research, 20, 37–44.
Martín-Alcázar, F., Romero-Fernández, P.M. and Sánchez-Gardey, G. (2005). 'Strategic
human resource management: integrating the universalistic, contingent,
configurational and contextual perspectives'. The International Journal of Human
Resource Management, 16: 5, 633–659.
McLachlan, G. and Peel, D. (2000). Finite Mixture Models, New York: Wiley.
Nishii, L.H., Lepak, D.P. and Schneider, B. (2008). 'Employee attributions of the “why” of
HR practices: Their effects on employee attitudes and behaviors, and customer
satisfaction'. Personnel Psychology, 61: 3, 203–545.
23
Nylund, K.L., Asparouhov, T. and Muthén, B.O. (2007). 'Deciding on the number of classes
in latent class analysis and growth mixture modeling: A Monte Carlo simulation
study'. Structural Equation Modeling: A Multidisciplinary Journal, 14: 4, 535–569.
Osterman, P. (1987). 'Choice of Employment Systems in Internal Labor Markets'. Industrial
Relations, 26: 1, 46–67.
Osterman, P. (1994). 'How Common is Workplace Transformation and Who Adopts it?'.
Industrial and Labor Relations Review, 47: 2, 173–188.
Sels, L., Winne, S., Delmotte, J., Maes, J., Faems, D. and Forrier, A. (2006). 'Linking HRM
and Small Business Performance: An Examination of the Impact of HRM Intensity on
the Productivity and Financial Performance of Small Businesses'. Small Business
Economics, 26: 1, 83–101.
Su, Z.-X. and Wright, P.M. (2012). 'The effective human resource management system in
transitional China: a hybrid of commitment and control practices'. The International
Journal of Human Resource Management, 23: 10, 2065–2086.
Toh, S.M., Morgeson, F.P. and Campion, M.A. (2008). 'Human resource configurations:
Investigating fit with the organizational context'. Journal of Applied Psychology, 93:
4, 864–882.
Verburg, R.M., Den Hartog, D.N. and Koopman, P.L. (2007). 'Configurations of human
resource management practices: a model and test of internal fit'. The International
Journal of Human Resource Management, 18: 2, 184–208.
Waldman, M. (2013). 'Theory and Evidence in Internal Labor Markets'. In R. Gibbons and J.
Roberts (Eds.), The handbook of organizational economics, Princeton: Princeton
University Press.
Walton, R.E. (1985). 'From control to commitment in the workplace'. Harvard Business
Review, 63: 2, 77–84.
Wright, P.M. and Kehoe, R.R. (2008). 'Human resource practices and organizational
commitment: A deeper examination'. Asia Pacific Journal of Human Resources, 46: 1,
6–20.

24
Table 1: Measures for Control and Commitment HRM Practices

Walton (1985) Arthur (1992; 1994) Baron Lepak and Snell Xiao and
and (1999; 2002) Björkmann
Kreps (2006)
(1999)

Commitment

Commitment

Commitment
Compliance
HC HRM
Control

Control

HCWS
Label
Job design Well-defined jobs
Wide variety of tasks rc
Work in semi-autonomous workgroups
Training Continuous training activities rc
Participation Open communication rc rc
Intensive exchange of information rc
Employees perform jobs that allow them to
routinely make changes in the way they per-
form their jobs rc
Employees perform jobs that empower them
to make decisions rc
Monitoring Work is governed by clear rules and proce-
dures
High trust relationships
Promotion Promotion from within
Compensation Extensive benefits package rc
Individual incentive component
Group-based incentive component
Employment security Long-term perspectives rc rc

Note: rc = reverse coded.

25
Table 2: Summary of LCA Statistics

Npar BIC(L²) AIC(L²)

1-cluster 17 6,253.657 1,1248.348
2-cluster 54 5,572.205 1,0384.257
3-cluster 91 5,561.461 1,0190.873
4-cluster 128 5,637.598 1,0084.372
5-cluster 165 5,746.789 1,0010.923

26
Table 3: Distribution of Clusters by Firm Characteristics

Cluster 1: Cluster 2: Cluster 3: Cluster 4:

High- Long- High- Regulated
control term- commitment commitment
system oriented system system
control
system Total
Firm size
20 to 49 38.8% 5.8% 30.0% 25.4% 100.0%
50 to 99 46.0% 17.0% 14.7% 22.3% 100.0%
100 to 499 38.5% 32.7% 15.4% 13.4% 100.0%
> 500 32.3% 36.5% 26.0% 5.2% 100.0%
Industry
Chemicals and pharmaceuticals 37.2% 25.6% 16.5% 20.7% 100.0%
Mechanical engineering 56.2% 1.9% 15.6% 26.3% 100.0%
Banking and insurance 13.8% 66.9% 16.0% 3.3% 100.0%
Professional services 13.6% 29.9% 51.0% 5.4% 100.0%
Firm age (mean, in years) 65.5 93.7 54.1 51.0 66.5
Family-owned
Yes 50.3% 4.3% 19.3% 26.1% 100.0%
No 30.3% 36.2% 22.3% 11.3% 100.0%
Endowment of HRM department
(mean) 107.6 84.4 62.2 142.0 99.5
Strategic orientation of HRM
Yes 32.9% 24.9% 25.2% 17.1% 100.0%
No 57.1% 11.8% 9.7% 21.5% 100.0%
Differentiated HRM
Yes 32.5% 25.3% 27.8% 14.3% 100.0%
No 48.7% 15.9% 12.0% 23.4% 100.0%
Share of highly skilled employees
(mean) 25.9% 35.4% 58.4% 12.1% 32.2%
Existence of a collective employee
representation
Works council and other forms 26.2% 41.7% 14.3% 17.9% 100.0%
Works council only 44.8% 24.8% 13.0% 17.4% 100.0%
Other forms only 32.1% 13.4% 32.1% 22.4% 100.0%
None 39.0% 14.2% 28.5% 18.2% 100.0%
Collective bargaining agreements
Commitment to collective bargaining
agreements 38.2% 30.2% 16.0% 15.6% 100.0%
Orientation towards collective bargaining
agreements 52.6% 6.5% 12.4% 28.4% 100.0%
None 25.4% 21.0% 42.9% 10.7% 100.0%
Strong fluctuations in order position
Yes 48.4% 13.2% 21.1% 17.4% 100.0%
No 31.8% 28.4% 20.6% 19.3% 100.0%
Cluster size 39.7% 21.2% 20.8% 18.3% 100.0%

27
Table 4: Effect of Clusters on HRM Outcomes

In reference to In reference to In reference to

Cluster 1: High-control system Cluster 2: Long-term-oriented Cluster 3: High-
R
control system commitment
system
Cluster 2: Cluster 3: Cluster 4: Cluster 3: Cluster 4: Cluster 4:
Long-term- High- Regulated High- Regulated Regulated
oriented control commitment commitment commitment commitment commitment
system system system system system system
Personnel costs are low -0.108 -0.061 0.080 0.047 0.189 0.142 0.039
Endowment with qualified employees is high 0.769*** 1.472*** 1.837*** 0.703** 1.069*** 0.366 0.192
Employees have up-to-date knowledge 0.709** 1.198*** 1.191*** 0.489* 0.482 -0.007 0.204
Employees show high performance level 0.847*** 1.441*** 0.996*** 0.595** 0.149 -0.446 0.170
Employees are highly-cost oriented 0.460* 1.207*** 0.824*** 0.747*** 0.364 -0.382 0.082
Employees are highly quality oriented 0.811*** 1.293*** 1.109*** 0.482* 0.299 -0.184 0.129
Employees are highly innovation oriented -0.014 0.822*** 0.645*** 0.836*** 0.659** -0.177 0.148
Employees are highly motivated 0.702** 1.986*** 0.958*** 1.284*** 0.257 -1.027*** 0.185
Employees show high job satisfaction 0.676** 1.878*** 1.341*** 1.203*** 0.665* -0.537* 0.183
Employees show high commitment 0.510* 1.483*** 0.823*** 0.973*** 0.313 -0.660** 0.119
Number of employees can be adjusted easily 0.260 0.386* 0.626*** 0.127 0.367 0.240 0.068
Allocation of personnel is very flexible in
0.330 0.966*** 0.798*** 0.636** 0.468 -0.168 0.083
terms of working time
Allocation of personnel is very flexible in
0.369 0.745*** 1.078*** 0.376 0.709*** 0.333 0.125
terms of task allocation

Notes: Estimates of ordinal logit regression. Control variables included. Levels of significance: * 5%; ** 1%; *** 0.1%.

28
Table 5: Effect of Clusters on Firm Outcomes

In reference to In reference to In reference to

Cluster 1: High-control system Cluster 2: Long-term-oriented Cluster 3: High-
R
control system commitment
system
Cluster 2: Cluster 3: Cluster 4: Cluster 3: Cluster 4: Cluster 4:
Long-term- High- Regulated High- Regulated Regulated
oriented control commitment commitment commitment commitment commitment
system system system system system system
Compared to our competitors, we realize…
…successful opening up of new markets. 0.037 0.624*** 0.918*** 0.587** 0.881*** 0.294 0.126
…high sales growth. 0.475* 1.058*** 0.877*** 0.584** 0.402 -0.182 0.126
…low costs of units of products. 0.150 0.334 0.325 0.184 0.175 -0.009 0.052
…high productivity. 0.575* 1.207*** 0.785*** 0.633** 0.210 -0.423 0.093
…low prices. 0.051 -0.368 0.065 -.419* 0.014 0.433 0.055
…high quality level. 0.852*** 1.534*** 1.575*** 0.682** 0.724** 0.042 0.178
…high innovation level. 0.032 1.014*** 0.693*** 0.982*** 0.661** -0.321 0.177
…strong differentiation. 0.245 0.715*** 0.464** 0.470* 0.219 -0.251 0.068
…building strong brands. 0.418* 0.958*** 0.551*** 0.540** 0.133 -0.406 0.115
…high profit. 0.232 0.576** 0.206 0.345 -0.025 -0.370 0.065
…high return on equity. 0.621** 0.716*** 0.366* 0.095 -0.255 -0.350 0.057
…positive business development. 0.624** 1.346*** 1.273*** 0.722*** 0.648* -0.074 0.134

Notes: Estimates of ordinal logit regression. Control variables included. Levels of significance: * 5%; ** 1%; *** 0.1%.

29
 Figure 1: Profile Data for HRM Practices

Note: The profile data contains the share of total agreement answers within the respective cluster.

30

View publication stats