Human Resource Management and Performance: A Review and Research Agenda

The International Journal of Human Resource
Management
ISSN: 0958-5192 (Print) 1466-4399 (Online) Journal homepage: http://www.tandfonline.com/loi/rijh20
Human resource management and performance: a

review and research agenda
David E. Guest
To cite this article: David E. Guest (1997) Human resource management and performance: a
review and research agenda, The International Journal of Human Resource Management, 8:3,
263-276, DOI: 10.1080/095851997341630
To link to this article: http://dx.doi.org/10.1080/095851997341630
Published online: 18 Feb 2011.
Submit your article to this journal
Article views: 11674
View related articles
Citing articles: 411 View citing articles
Full Terms & Conditions of access and use can be found at

http://www.tandfonline.com/action/journalInformation?journalCode=rijh20
Download by: [Banaras Hindu University BHU] Date: 06 January 2016, At: 21:39
The International Journal of Human Resource Management 8:3 June 1997
Human resource management and

performance: a review and research
agenda
David E. Guest
Downloaded by [Banaras Hindu University BHU] at 21:39 06 January 2016
Abstract There is a growing body of evidence supporting an association between what

are termed high performance or high commitment human resource management (HRM)
practices and various measures of organizational performance. However, it is not clear
why this association exists. This paper argues that to provide a convincing explanation of
this association we need to improve our theoretical and analytic frameworks in three key
areas. These are the nature of HRM, and especially the rationale for the speci c lists of
HR practices; the nature of organizational performance; and the linkage between HRM
and performance. A model is presented within which to explore these linkages. The
existing literature on HRM and performance is reviewed in the light of this analysis to
identify key gaps in knowledge and help to focus further the research priorities.
Keywords Human resource management, performance, outcomes, theory, strategic

integration
The impact of human resource management on performance has become the dominant
research issue in the eld. There has been a rash of studies demonstrating a positive
association between human resource management (HRM) and performance, providing
encouragement to those who have always advocated the case for a distinctive approach
to the management of human resources. While these studies represent encouraging
signs of progress, statistical sophistication appears to have been emphasized at the
expense of theoretical rigour. As a result, the studies are non-additive, except in a very
general way. My aim in this paper is to present a form of research agenda which seeks
to reintroduce theory into the empirical debate and to use this to review some of the
emerging empirical ndings.
If we are to improve our understanding of the impact of HRM on performance, we
need a theory about HRM, a theory about performance and a theory about how they are
linked. The interest in HRM and performance partly re ects a view that the debates
about theory in HRM had become rather introspective and boring. Perhaps it is only
when the empirical data begin to emerge that we realize how important the theory is. I
shall begin by brie y reviewing where we stand on theories of HRM, of performance
and of the link between HRM and performance. I shall then review some of the recent
literature within a simple framework. Finally, I shall outline the issues emerging from
this review and highlight some of the developments that need to occur in theory and
practice if we are to improve our understanding of HRM and performance.
0985–5192 © Chapman & Hall 1997

264 David E. Guest
Theory about human resource management
In any empirical study, we need to have a clear view about our independent variable. At
present we seem to have only confusion. There appear to be three broad categories of
general-level theory about HRM and a host of more speci c and concrete theories about
particular areas of policy and practice, such as quality, commitment and performance-
related pay. The three categories of theory can be labelled strategic, descriptive and
normative.
Strategic theories of HRM

These are primarily concerned with the relationship between a range of possible
external contingencies and HRM policy and practice. Some of the best-known UK
research in this category has been reported by Hendry and Pettigrew (1990) whose main
concern was to identify and classify key environmental in uences on HRM. In their
work, HRM policy and practice becomes, in a sense, the dependent variable, judged in
terms of how well it ts the context. The implicit but untested hypothesis is that a good
t will be associated with superior performance. In their research, Hendry and Pettigrew
concentrated on mapping the context, identifying an inner context (within the
organization) and an outer context (in the wider environment) and exploring how HRM
adapted to changes in context rather than analysing any link to performance.
In the USA, attention has focused more on classifying types of HR strategy, often
drawing on existing models of corporate strategy. One frequently cited example has
been presented by Miles and Snow (1984), building on their earlier work on strategy
and structure. They propose that each of their strategic types of rm will need to adopt
a different set of HRM policies and they are reasonably precise about some of the
variations. Again the hypothesis is that those rms that have a t between business
strategy, structure and HRM policy and practice will have superior performance. A
similar rationale lies behind the work of Schuler and Jackson (1987) which outlines
three HRM strategies linked to Porter’s (1980) three general competitive strategies.
There are several other writers whose work falls within this broad focus on strategy.
They are helpful in identifying in uences on strategic t and sometimes in specifying
types of t. But they are often simplistic in characterizing HRM, usually identifying
about four broad areas of activity such as selection, training and development, rewards
and careers. While the implication is that those rms achieving t between business
strategy and HRM strategy will have superior performance, they are weak in specifying
the process whereby HRM is linked to performance. Finally, they generally adopt a
limited view of performance, de ning it largely in nancial terms.
Descriptive theories of HRM

These set out to describe the eld in a comprehensive way. The two best known are
those presented by Beer and his colleagues from Harvard (Beer et al. 1985) and by
Kochan, Katz and McKersie (1986) from MIT. In both cases, there is an attempt to
capture the broad eld and to address some of the interrelationships. For Beer at al. this
means listing four broad areas of HRM policy and practice and four key outcomes. For
Kochan and colleagues it entails a systems approach, describing the interrelationships
between levels. Both are essentially descriptive, mapping the eld and classifying
inputs and outcomes. Both can be developed into theory, but only at a very general level
of speci cation. A strength and a weakness in this respect is the emphasis on an open
Human resource management and performance 265
systems approach which may accurately capture an element of reality but fails to offer
speci city.
By providing conceptual maps of the eld, these models do provide a broad
classi cation of the content of HRM and of a range of outcomes. Both are useful in
adopting a stakeholder approach and in identifying a range of outcomes of interest to
the various stakeholders. However, they are essentially employee relations models
concerned with the managers’ role in balancing competing interests, in highlighting the
scope for choice and in identifying some of the in uences on the choices. Despite
implicit preferences on the part of their authors, by specifying a range of choices and
not recommending speci c approaches, they are largely non-prescriptive. They
therefore provide no clear focus for any test of the relationship between HRM and
performance.
Normative theories of HRM
Models or theories of this type are more prescriptive in their approach, re ecting the
view either that a suf cient body of knowledge exists to provide a basis for prescribed
best practice or that a set of values indicates best practice. Often these two perspectives
become con ated. One of the best known examples of this approach is Walton’s work
on control and commitment (Walton, 1985). In presenting the contrast between the two
approaches to the management of human resources, he follows McGregor (1960) some
twenty- ve years earlier in saying these are in one sense ideal types but in practice if
you wish to ourish you have no choice. He is prescribing a commitment strategy as the
distinctive basis for HRM. The same general analysis can be found in the work of
Lawler (1986, 1992), although he uses the language of involvement rather than HRM.
More recently, the work of Pfeffer (1994) has attracted a lot of attention. He lists
sixteen HRM practices (subsequently amended to thirteen (Pfeffer, 1995) on the
grounds that the precise number and presumably the precise nature of the practices is
neither clearly known nor particularly important) which he advocates on the grounds
that their positive effects are now well established. This ts with an essentially
atheoretical stream of work about ‘high performance’ work practices.
My own work (Guest, 1987) has attempted to capture some of the spirit of this
approach by seeking to present it within a coherent framework, specifying some of the
links so that the resulting model can at least be tested – and possibly refuted. The
central hypothesis is that if an integrated set of HRM practices is applied with a view
to achieving the normative goals of high commitment to the organization plus high
quality and exibility, then higher worker performance will result. The assumption is
that this will have a positive impact on organizational performance. Unlike other
approaches, this normative perspective argues that speci c practices and speci c HRM
goals will always be superior.
There are a number of problems with this view of HRM. One is that it focuses
predominantly on the internal characteristics of HRM at the expense of broader
strategic issues. In doing so, and in advocating a best set of practices while ignoring the
variety of pressures and consequent business strategies, it is taking a considerable risk
in implying ‘one best way’. A second problem is that, while the goals of HRM can be
reasonably well de ned, the related list of HRM practices is far from clear (for an
outline of the variables included in the various studies, see Dyer and Reeves (1995) and
Becker and Gerhart (1996)) and awaits either a clear theoretical speci cation or a much
stronger empirical base.
266 David E. Guest
Each of the three approaches outlined has some sort of theoretical basis in either
contingency/business strategy, systems theory or OB/motivation theory. Each implies
rather different levels of analysis. Only the second and more particularly the third begin
to specify the dimensions of HRM policy and practice in any way potentially helpful for
measurement and even they are not really suf ciently precise. In summary, we still lack
a coherent theoretical basis for classifying HRM policy and practice, a problem that
becomes more apparent when we start to look at the empirical research.
Theory about performance
There is no general theory about performance per se. However, we have a number of
approaches and models, often built on speci c disciplinary perspectives, such as

economics, psychology or production management, which help us to understand and
classify aspects of performance. This is not an easy task. In the eld of Organizational
Behaviour, measurement of performance is often described as the criterion problem. We
might be tempted to extend this to the study of HRM.
We can begin to make sense of performance by highlighting a number of distinctions.
First, we can focus on issues concerning the content of performance. Second, we can
consider the types of data. Third, we can consider linkages within a broad view of
performance and thereby begin to explore causal links between HRM and
performance.
The content of performance and outcomes
The issue of the nature of data overlaps with the question of what types of data are of
interest. Who are the stakeholders of performance – and is this the same as outcomes?
Arguably, performance is a company-dominated criterion while outcomes are poten-
tially much broader. They may include environmental issues, job satisfaction,
contribution to community activities and so on. There is a risk, in looking at
organizationally determined performance criteria, that some of these issues are ignored.
It could be argued that, if the concern is narrowly for company performance, this does
not matter. On the other hand, if the implicit theory of HRM is that results are achieved
through effective utilization, as opposed to exploitation, of human resources, they might
be very important. However, even this may not be enough, since we know from a mass
of research that there is only a very weak link between performance and job satisfaction
(Staw, 1986). Also, there is some indication that highly successful companies may not
be those in which workers prefer to work (Levering et al., 1984; for a fuller discussion,
see Guest, 1992).
Building on the stakeholder analysis, one of the terms sometimes used in discussions
about HRM is the concept of ‘the balanced scorecard’. This simply implies that it is not
enough to concentrate on one view of performance at the expense of the others. As in
the old socio-technical systems theory, it is necessary to optimize each dimension rather
than maximizing one at the expense of the others. Typically, the items on the scorecard
are those of importance to the nancial, customer and employee constituencies. If we
accept this, we need multiple criteria of performance, a point highlighted in the
descriptive models of HRM and implicit in the whole idea of taking human resource
management seriously. It means we must be wary of emphasizing one of these interests
at the expense of the other, particularly if we are using one-off rather than time series
measures.
Measures of performance
There are essentially three types of performance data available (Locke and Latham,
1990); these are measures of output of goods and services, which may be quantitative
(units produced, customers served) or qualitative (number of errors, customer
complaints); measures of time, including lateness, absence, lost working time, failure to
meet deadlines); and nancial indicators, which include a large array of possibilities.
These may be interrelated; thus absence reduces unit production levels which reduces
pro ts.
Where ‘hard’ quantitative data are not available, we may wish to resort to measures
of behaviour. Ideally, we observe people at work, noting whether they approach
customers and offer help, or noting whether they observe safety procedures. Failing
that, we can obtain reports either from an ‘observer’, such as the superior or a peer or
subordinate, or from the person under study.
The same perspective can be adopted for performance data of interest to other
stakeholders, for example those concerned with the environment. The measurement
becomes more complex when applied to employees, since we will need to add
subjective evaluations. The important point is that they should be collected system-
atically, perhaps through mechanisms such as attitude surveys.
Of course, an analytic framework for considering the range of performance data
cannot hide biases and preferences – perhaps for nancial measures – nor can it prevent
subjective interpretation and weighing of the information. The simple point is that as
researchers we should perhaps be aware of the range of types of data that might be
collected to indicate performance. There is also an implicit judgement about preferred
priority, starting with the ‘hard’ measures, followed by behaviour, followed by reports
or ratings. We may of course wish to balance the post-modernist view that all
performance measures are social constructions, open to a variety of interpretations,
against a recognition that broadening the de nition of performance and providing
evidence about any link between HRM and performance may be one of the more
effective means of ensuring that policy makers take HR issues seriously.
Linkages in performance data

Data on their own tell us very little. We need some basis for comparison, either cross-
sectional or longitudinal; we also need to understand the relationships between types of
performance data. As already noted, absence may reduce production which may affect
pro ts. Many linkages of this type can be made. If we move towards a behavioural
model, then we might wish to link employee perceptions to their behaviour, to
individual or group-level performance outcomes which affect unit performance and thus
to company pro ts. To understand, as opposed to measuring the performance, we need
to make these linkages. There may be an association between HRM practices and
company pro t, but, without some linkages, we will not know why; we have no theory.
This implies that we need a range of types of performance measure at the individual,
group, unit and company level to begin to establish these linkages. To date, HRM has
been weak in this area.
Theory about the link between HRM and performance

Neither the strategic nor the descriptive model provide much insight into how HRM
policy and practice translates into high performance. They sensibly view HRM as only
268 David E. Guest
part of the explanation for high performance and suggest that, when the various sub-
systems including the HRM sub-system are aligned and supporting each other, superior
performance is likely. The right people will be in the right place doing the right things.
It is a managerial model of organizing, particularly in the strategic version. The
descriptive models, because of their employee relations roots, give more weight to the
active role of employees, but mainly in negotiating order.
In contrast, the normative approach has its roots rmly in organizational psychology
and is built on lower-range, more speci c behavioural theories. The assumption is that
‘appropriate’ HRM practices tap the motivation and commitment of employees. The
factors that constitute ‘appropriate’ practices are derived from the speci c theories of
organizational commitment, job design, goal setting and so on. The key features of this
approach are that HRM provides a coherent integration of these behavioural theories
and that they spell out the linkages between practices and performance.
It is possible to develop theories of alignment, explaining a series of linkages to
performance, both at the broader level of strategy and at the more speci c level of HRM
practices. Starting with the latter, the expectancy theory of motivation provides one
possible basis for developing a more coherent rationale about the link between HRM
practices and performance. An approach close to expectancy theory was adopted by
MacDuf e (1995). As he puts it,
Innovative human resource practices are likely to contribute to improved economic
performance only when three conditions are met: when employees possess knowledge
and skills that managers lack; when employees are motivated to apply this skill and
knowledge through discretionary effort; and when the rm’s business or production
strategy can only be achieved when employees contribute such discretionary effort. I will
argue that all three conditions must be met for HR practices to contribute to
performance.
(MacDuf e, 1995: 199)
Although expectancy theory is concerned primarily with motivation, it is also a theory
about the link between motivation and performance. Speci cally, it proposes that high
performance, at the individual level, depends on high motivation plus possession of the
necessary skills and abilities and an appropriate role and understanding of that role. It
is a short step to specify the HRM practices that encourage high skills and abilities, for
example careful selection and high investment in training; high motivation, for example
employee involvement and possibly performance-related pay; and an appropriate role
structure and role perception, for example job design and extensive communication and
feedback. These are illustrated in Figure 1. We therefore have a theory which links
HRM practices to processes that facilitate high individual performance; it is,
furthermore a theory which ts well with the individualistic orientation of the normative
approaches to HRM.
Before getting too enthusiastic about any theory linking HRM practices and
performance, we should bear in mind that any analysis of in uences on company pro ts
or factory output or even absence levels quickly leads to the conclusion that factors
other than HRM are involved. Therefore, we need to have a theory about when human
resources matter more; for example, the human factor may be more important in the
service sector. We also need a theory about how much of the variance can be explained
by the human factor. Is any statistically signi cant in uence, for example where a
regression reveals that HRM explains 2 per cent of the variance in performance, a good
result? Or should we, as Becker and Gerhart (1996) suggest, be paying much more
attention to size effects than to statistical signi cance? Can we, as Huselid (1995) has
Selection
Socialization Skills and Ability
Training and development (Quality)
Quality improvement programmes
Single status
Job security Effort/Motivation
Internal promotion (Commitment)
Individualized reward systems
Communication
Employee involvement Role Structure and Perception

Team working (Flexibility)
Job design
Flexible job descriptions
Figure 1 Linking HRM practices and HRM outcomes
recently attempted, put a price on the bene ts of good HRM? For this, we need a rather
broader conceptual framework as well as an accepted basis on which to judge what
constitutes an important result.
Moving on to a possible broader framework linking HRM and outcomes, the starting
point should be the recognition, as noted above, that the distinctive feature of HRM is
its assumption that improved performance is achieved through the people in the
organization. Any theory of linkages should explicitly build on this. A model that
illustrates the kind of linkages that might be explored, albeit using arbitrary categories
for HRM strategy and HRM practices, is presented in Figure 2. Its main value in the
present context lies in highlighting the range of related outcomes that need to be
considered in any model that seeks to understand the impact of the ‘human’ factor in
human resource management.
Figure 2 acknowledges a role for external context and strategy but suggests, at its
core, that HRM practices should be designed to lead to HRM outcomes of high
employee commitment, high quality staff and highly exible staff. High commitment
will be assessed using standard measures and based on the de nition popularized by
Mowday, Porter and Steers (1982) of identi cation with the values of the organization,
a desire to belong to the organization and a willingness to do things for the
organization. High quality staff refers to the capabilities and to the knowledge and skills
of staff. Flexibility refers to functional rather than numerical exibility. The speci c
practices that help to achieve these outcomes were illustrated in Figure 1. According to
the model, only when all three HRM outcomes are achieved can we expect behaviour
change and higher performance.
Given the framework in Figure 2, it is essential to measure HRM outcomes to
understand how HRM impacts on performance. As we move on through the remaining
stages in the model, we would expect the impact of HRM to become progressively
weaker as other factors intervene. The model in Figure 2 separates changes in
behaviour, performance and nancial outcomes. It is important to note that under the
broad heading of performance there are further distinctions that can be made, for
example between performance and effectiveness, which is a subjective assessment of
270 David E. Guest
HRM HRM HRM Behaviour Performance Financial

strategy practices outcomes outcomes outcomes outcomes
Selection Effort/ High: Pro ts
Differentiation Motivation Productivity
(Innovation) Training Commitment Quality
Innovation
Focus Appraisal Cooperation
(Quality)
Rewards Quality Low: ROI
Cost Involvement Absence
(Cost- Job design Labour turnover
reduction) Con ict

Involvement Flexibility Organizational Customer complaints
citizenship Labour turnover
Status and
Security
Figure 2 Linking HRM and performance
performance and perhaps productivity; and between performance, which is concerned

with output, and productivity, which is concerned with input-output ratios. Since
effectiveness may be considered quite differently by the various stakeholders, it will be
desirable where possible to include additional measures of effectiveness alongside the
more objective outcome measures.
The empirical evidence about HRM and performance
A conceptual framework
There are now a growing number of studies which make a serious attempt to link HRM
and performance. (There are, of course, hundreds which examine particular aspects of
HRM and performance.) They are not all looking at quite the same thing so they are
cumulative in only a very general sense. One interesting attempt to identify the various
possible approaches has been derived from the work by Venkatraman (1989) whose aim
was to classify the various types of ‘ t’ in research on strategy. More recently, his
approach has been adapted slightly by Sivasubramaniam and Kroeck (1995) to classify
the various perspectives on HRM. The concept of ‘ t’ or ‘integration’ is central to many
attempts to theorize about HRM so this seems an interesting way of trying to proceed.
Essentially, they suggest that the various types of HRM t can be ordered along two
dimensions of Internal–External t and Criterion speci c or Criterion free. This
provides four main possibilities, although one is further sub-divided. I have made
considerable amendments to the descriptions used in both previous papers:
Fit as strategic interaction seeks to link HR practices to the external context and
re ects the standard strategic approach. A key point is that there is a choice about how
to respond to and interact with the environment. Once that choice has been made, then
HR strategy and practice must match it. The hypothesis is that those organizations with
the appropriate response and the right match will report superior performance. The
typical test of this is to examine the link between the Miles and Snow strategic types
and the HRM practices associated with each and relate this to some measure of
Criterion specific Criterion free
Internal Fit to an ideal set of practices Fit as gestalt

Fit as ‘bundles’
External Fit as strategic interaction Fit as contingency
Figure 3 Forms of HRM t

performance. The level of theory speci cation implies that performance will be
measured mainly through nancial criteria.
Fit as contingency re ects the traditional contingency approach, suggesting that those
organizations whose HR policy and practice is more responsive to external factors will
report superior performance. These external factors may include the nature of the
market, legislative changes or features of the speci c sector. This approach tends to
assume that a particular response will always be superior, but it does not specify the
type of response or any class of responses, although there is no reason why this could
not be done.
Fit as an ideal set of practices implies that there exists a set of ‘best HRM practices’
– Pfeffer’s (1994) list might be one example – and the concern is with how close
organizations get to the ideal set of HR practices. The hypothesis is that those closer to
the ideal type will report higher performance.
Fit as gestalt implies that the key to effective HRM lies in nding an appropriate
combination of practices. In addition, it is assumed that the sum is greater than the
parts. It may be the synergistic combination of all the practices or it may be the speci c
‘architecture’ or culture which binds them together. This is sometimes seen as the non-
replicable element which can provide organizations with an HR-based source of
competitive advantage. One implication of this model is that the study of speci c
aspects of HRM such as pay or training may provide a distorted picture. Unless very
careful controls are used to take account of all other HRM practices, it is possible that
the results may overstate the in uence of the speci c practice by picking up the
combined impact of the set of practices. In principle, this approach differs from the
ideal type concept of t in not specifying the dimensions of HRM practice; and from
both the ideal type and the ‘bundles’ approach described below in being multiplicative
rather than additive. By implication, if one key aspect is missing, the gestalt may not
exist.
Fit as ‘bundles’ implies the existence of distinctive patterns or con gurations – what
are sometimes called ‘bundles’ – of practices and the key is to determine which are the
most effective. In principle there may be a number of possible combinations or
con gurations of practices which will lead to high performance; for example, some
organizations may emphasize job security as the building block; others prefer training
and development. The other practices t around these. To test this, the key is to look not
so much at the total number of HRM practices but to take those who adopt above a
certain number, perhaps the median, as long as a distinctive core exists. Within each
bundle some substitutability may exist; for example, statistically, there may be little to
be gained by reporting both use of realistic job previews and psychometric tests in
selection since both are equally indicative of careful selection. Although in principle
272 David E. Guest
this approach allows for equi nality, or some substitution, in practice certain clusters
may consistently emerge, indicating the limited range of effective combinations.
Most reported studies of HRM and performance can be classi ed within one or other
of these categories. However, we are some way from seeing all the possibilities tested.
Indeed, Delery and Doty (1996) argue that there are really three main approaches – the
strategic contingency approach, the best practice approach and the con gurational
approaches, representing external t, internal t and systems theory – and that these
provide rich competing theoretical perspectives. Since there has been very little
research on the pure contingency approach and, apart from some speci c case studies,
little exploration of the ‘gestalt’ perspective, the next section concentrates on these three
main approaches.
The empirical studies of HRM and performance

A number of studies linking HRM and performance have been published; others have
been presented in conference papers or exist in the form of working papers. They are
mostly, but not exclusively, cross-sectional. The studies vary in quality, particularly
with respect to their measurement of HRM policy and practice, although they share a
tendency to display some statistical sophistication. While some range widely across
industry, others concentrate on particular sectors.
External t: HRM as strategic integration There is some support for this approach.
Huselid (1995) has found that those organizations that link HRM practices to strategy
report higher nancial performance outcomes. Delery and Doty (1996) found modest
support for a t with the Miles and Snow typology. Youndt et al. (1996) found support
for this type of t in a sample of ninety-seven manufacturing plants, more particularly
with respect to the t between high performance HRM practices and a quality strategy.
MacDuf e (1995), in contrast, explicitly rejects this hypothesis, claiming that in his
international study of car-manufacturing plants he found no evidence that a ‘ t’ of
appropriate HRM practices to mass production was able to compete with exible
production.
Internal t: HRM as an ideal set of practices This is the most widely tested and the
most strongly supported type of t. Almost every study, including those already cited
but also those reported by Huselid and his colleagues (Huselid, 1995; Huselid, Jackson
and Schuler, 1995; Delaney and Huselid, 1996), by Arthur (1994), Ichniowski, Shaw
and Prennushi (1994) and MacDuf e (1995) support this type of t. Essentially, what
they all show, either across industries or within a speci c sector, is that the more of the
high performance HRM practices that are used, the better the performance as indicated
by productivity, labour turnover or nancial indicators. Where different types of t
were compared, this was invariably the one that received strongest support. The main
exception to this general pattern, to date, is the study by Youndt at al. (1996) which
nds that a ‘ t’ between high performance HRM practices and a quality strategy
provides the best results.
Con gurational t: HRM as bundles Huselid and Becker (1995), in a panel study,
have provided a partial test of this by examining the impact of three separate factors
which emerged from their factor analysis of a list of HRM practices. These they label
selection and development, motivation and HR strategy, though the labels may not be
very appropriate. Although support was strongest for the cumulative measure of ‘ t as
ideal type’, there was signi cant independent support for each of these factors. In a
study using subjective measures of organizational performance, Delaney and Huselid
(1996) failed to nd any positive impact for speci c combinations of practices as
opposed to the total number of HRM practices. However Delery and Doty (1996) found
some support for the con gurational approach in their study in the banking sector.
Guest and Hoque (1994) examined the impact of HRM comparing those above and
below the median number of HRM practices in a sample of green eld sites. When this
was combined with a measure of the presence of an HRM strategy, it was found that
establishments in this group reported superior HRM outcomes (commitment, employee
quality, aspects of exibility) and employee relations outcomes but not superior
productivity or product quality. This highlights the need to take into account the range
of outcomes and performance indicators. The study also hinted at a ‘cluster’ or ‘bundle’
re ecting deliberate low use of the proffered set of HRM practices. Those who did this
as a deliberate strategy had poorer HRM and employment relations outcomes but
reported no differences in productivity and product quality. This in turn raises a
somewhat neglected issue of the costs of HRM practices.
Bringing these results together, there is empirical support for each of the three main
approaches to HRM and performance but consistently stronger support for the internal
t model with its view that those organizations that use more high performance HRM
practices report higher performance. The general approach represented by this stream of
research is suf ciently encouraging to suggest that it is continued and improved. There
is no doubt that it is attracting a lot of research interest in North America. It is to be
hoped that it can do the same in Europe. However, in taking it forward, we need to be
aware of the shortcomings and in so doing return to the issue of theory.
Emerging issues for future research on HRM and performance
The rst key issue is the lack of theory about the nature of HRM practices. As I have
argued elsewhere (Guest, 1987), it is not the presence of selection or training but a
distinctive approach to selection or training that matters. It is the use of high
performance or high commitment HRM practices. There is little consensus on what
these are and little interest to date in developing theory about what they might be. I
would propose that we build on something like expectancy theory, in a way outlined
above, to provide a sensible rationale for these practices. As MacDuf e (1995)
indicates, expectancy theory, or some variation of it, can also point to a range of
con gurations depending on preferred rewards and perhaps indicating when perform-
ance-related pay can be considered a high performance practice. The alternative is to
adopt a statistical approach and to see what emerges from factor analysis or some
variation (for an example of this, see Wood and Albanese, 1995). In building a set of
best practices, we should also take care to account for cultural differences, for example
in practices associated with training, job security or trade unions. Finally, in collecting
information about high performance HRM policy and practice, we need to be cautious
about the validity assumptions underlying use of data collected at head of ce as a basis
for company-wide statements about HR practices in what might be multi-plant sites.
A second concern is to improve our measures of performance. Indeed, performance
may be the wrong term. It might be more sensible to use the term ‘outcomes’. One
argument might be that the unitarist philosophy underlining HRM implies that the
employees share the concern of shareholders for pro t. However, it would be unwise to
274 David E. Guest
accept this assumption. What we need are outcomes that re ect the concept of the
balanced score card. The idea of balance, like the concept of optimizing socio-technical
systems, implies that compromises may sometimes be necessary. The plant-level
studies reported by Arthur (1994), MacDuf e (1995) and Guest and Hoque (1994)
measure productivity and quality, but those by Huselid (1995) and others using
company-level measures overwhelmingly emphasize nancial criteria. Despite the
attractions of nancial indicators for any attempt to convince senior managers of the
impact of HRM, we need to use a greater range of outcome measures if only to
understand how and why HRM has an impact on nancial results.
The study by Guest and Hoque raises the question of contradictory criteria. Their
establishments categorized as ‘ugly’, with a deliberate strategy of low uptake of high
performance HRM practices, reported performance outcomes as good as any other

category but had poorer HRM outcomes and employee relations outcomes. The rms
got good results but were not attractive places in which to work. The other studies took
no account of contradictory outcome criteria.
Only when we have made progress in measuring the independent and dependent
variables can we begin to give full attention to the way in which they are linked. At
present the studies report a promising association between HRM and outcomes, but we
are not yet in a position to assert cause and effect. We need to develop theory which
integrates aspects of strategy and strategic integration with something like expectancy
theory to create speci c hypotheses about linkages. Case study research can help to
generate some insights which can be more extensively tested. So too can studies of
speci c practices or issues such as quality or commitment where theory is sometimes
more speci c. To establish linkages, we also need longitudinal research designs, ideally
with some sort of interventions to alter HRM practices.
In summary, we need to improve our theory and empirical testing on all three fronts
– the nature of HRM, the nature of the outcomes and the nature of the linkages. As
usual, we need to nd the right combination of survey-based and case-study-based
research. And we need to ensure that the research is not con ned to the USA. We can
now say with increasing con dence that HRM works. But this is a skeletal nding and
we need to put a lot of esh on the bones.
Birkbeck College
University of London
Note
An earlier version of this paper was presented to the seminar. ‘Human Resource Management and
Performance’ held at the Tinbergen Institute, Holland, 22 September 1995.
References
Arthur, J.B. (1994) ‘Effects of Human Resource Systems on Manufacturing Performance and
Turnover’, Academy of Management Journal, 37: 670–87.
Becker, B. and Gerhart, B. (1996) ‘The Impact of Human Resource Management on
Organizational Performance. Progress and Prospects’, Academy of Management Journal, 39(4):
779–801.
Beer, M., Spector, B., Lawrence, P., Quinn Mills, D. and Walton, R. (1985) Human Resource
Management: A General Manager’s Perspective, New York: Free Press.
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):
949–69.
Delery, J. and Doty, D.H. (1996) ‘Modes of Theorizing in Strategic Human Resource
Management: Tests of Universalistic, Contingency and Con gurational Performance Predic-
tions’, Academy of Management Journal, 39(4): 802–35.
Dyer, L. and Reeves, T. (1995) ‘Human Resource Strategies and Firm Performance: What Do We
Know and Where Do We Need to Go?’, paper presented to the 10th IIRA World Congress,
Washington, 31 May – 4 June.
Guest, D.E. (1987) ‘Human Resource Management and Industrial Relations’, Journal of
Management Studies, 24: 503–21.
Guest, D.E. (1992) ‘Right Enough to Be Dangerously Wrong: An Analysis of the In Search of
Excellence Phenomenon’. In Salaman, G. (ed.) Human Resource Strategies, London: Sage.

Guest, D.E. and Hoque, K. (1994) ‘The Good, the Bad and the Ugly: Human Resource
Management in New Non-union Establishments’, Human Resource Management Journal, 5(1):
1–14.
Hendry, C. and Pettigrew, A. (1990) ‘Human Resource Management: an agenda for the 1990s’,
International Journal of Human Resource Management, 1(1): 17–43.
Huselid, M.A. (1995) ‘The Impact of Human Resource Management Practices on Turnover,
Productivity and Corporate Financial Performance’, Academy of Management Journal, 38:
635–70.
Huselid, M.A. and Becker, B. (1995) ‘High Performance Work Practices and the Performance of
the Firm: The Mediating Effects of Capital Structure and Competitive Strategy’, paper
presented to the Academy of Management Conference, Vancouver, 6–9 August.
Huselid, M.A., Jackson, S.E. and Schuler, R.S. (1995) ‘The Signi cance of Human Resource
Management Implementation Effectiveness for Corporate Financial Performance’, paper
presented to the Academy of Management Conference, Vancouver, 6–9 August. In press,
Academy of Management Journal.
Ichniowski, C., Shaw, K. and Prennushi, G. (1994) ‘The Effects of Human Resource Management
Practices on Productivity ’, working paper, Columbia University.
Kochan, T., Katz, H. and McKersie, R. (1986) The Transformation of American Industrial
Relations, New York: Basic Books.
Lawler, E.E. (1986) High Involvement Management, San Francisco: Jossey-Bass.
Lawler, E.E. (1992) The Ultimate Advantage. San Francisco: Jossey-Bass.
Levering, R., Moskowitz, M. and Katz, M. (1984) The 100 Best Companies to Work for in
America. Reading, Mass.: Addison-Wesley .
Locke, E.A. and Latham, G.P. (1990) ‘Work Motivation: The High Performance Cycle’. In
Kleinbeck, U., Quast, H-H., Thierry, H. and Hacker, H. (eds) Work Motivation, Hillsdale, NJ:
Lawrence Erlbaum.
MacDuf e, J.P. (1995) ‘Human Resource Bundles and Manufacturing Performance: Flexible
Production Systems in the World Auto Industry ’, Industrial Relations and Labor Review, 48:
197–221.
McGregor, D. (1960) The Human Side of Enterprise, New York: McGraw-Hill.
Miles, R. and Snow, C. (1984) ‘Designing Strategic Human Resource Systems’, Organizationa l
Dynamics, Summer: 36–52.
Mowday, R.T., Porter, L.W. and Steers, R.M. (1982) Employee-Organization Linkages: The
Psychology of Commitment, Absenteeism and Turnover, New York: Academic Press.
Pfeffer, J. (1994) Competitive Advantage Through People, Boston, Mass.: Harvard Business
School Press.
Pfeffer, J. (1995) ‘Producing Sustainable Competitive Advantage through Effective Management
of People’, Academy of Management Executive, 9(1): 55–69.
Porter, M.E. (1990) Competitive Strategy, New York: Free Press.
276 David E. Guest
Schuler, R.S. and Jackson, S.E. (1987) ‘Linking Competitive Strategies with Human Resource
Management Practices’, Academy of Management Executive, 1(3): 209–13.
Sivasubramaniam, N. and Kroeck, K.G. (1995) ‘The Concept of ‘‘Fit’’ in Strategic Human
Resource Management ’, paper presented to the Academy of Management Conference ,
Vancouver, 6–9 August.
Staw, B. (1986) ‘Organizational Psychology and the Pursuit of the Happy/Productive Worker’,
California Management Review, 18(4): 40–53.
Venkatraman, N. (1989) ‘The Concept of Fit in Strategy Research: Toward Verbal and Statistical
Correspondence ’, Academy of Management Review, 14(3): 423–44.
Walton, R.E. (1985) ‘From Control to Commitment in the Workplace ’, Harvard Business Review,
63: 77–84.
Wood, S. and Albanese, M.T. (1995) ‘Can We Speak of Human Resource Management on the
Shop Floor?’, Journal of Management Studies, 3(2): 215–47.

Youndt, M., Snell, S.A., Dean, J.W. and Lepak, D.P. (1996) ‘Human Resource Management ,
Manufacturing Strategy, and Firm Performance’, Academy of Management Journal, 39(4),
836–66.

Human Resource Management and Performance: A Review and Research Agenda

Uploaded by

Copyright:

Available Formats

Human Resource Management and Performance: A Review and Research Agenda

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Human Resource Management and Performance: A Review and Research Agenda

Uploaded by

Copyright:

Available Formats

The International Journal of Human Resource

ISSN: 0958-5192 (Print) 1466-4399 (Online) Journal homepage: http://www.tandfonline.com/loi/rijh20

Human resource management and performance: a

To link to this article: http://dx.doi.org/10.1080/095851997341630

Published online: 18 Feb 2011.

Submit your article to this journal

Article views: 11674

View related articles

Citing articles: 411 View citing articles

Full Terms & Conditions of access and use can be found at

Human resource management and

Abstract There is a growing body of evidence supporting an association between what

Keywords Human resource management, performance, outcomes, theory, strategic

0985–5192 © Chapman & Hall 1997

Strategic theories of HRM

Descriptive theories of HRM

Normative theories of HRM

Theory about performance

approaches and models, often built on speci c disciplinary perspectives, such as

The content of performance and outcomes

Linkages in performance data

Theory about the link between HRM and performance

Employee involvement Role Structure and Perception

Figure 1 Linking HRM practices and HRM outcomes

HRM HRM HRM Behaviour Performance Financial

reduction) Con ict

Figure 2 Linking HRM and performance

performance and perhaps productivity; and between performance, which is concerned

The empirical evidence about HRM and performance

Criterion specific Criterion free

Internal Fit to an ideal set of practices Fit as gestalt

External Fit as strategic interaction Fit as contingency

Figure 3 Forms of HRM  t

The empirical studies of HRM and performance

Emerging issues for future research on HRM and performance

performance HRM practices, reported performance outcomes as good as any other

Excellence Phenomenon’. In Salaman, G. (ed.) Human Resource Strategies, London: Sage.

Shop Floor?’, Journal of Management Studies, 3(2): 215–47.

You might also like

approaches and models, often built on speci c disciplinary perspectives, such as

reduction) Con ict

Figure 3 Forms of HRM t