60 SITUATING HUMAN RESOURCE MANAGEMENT
well. A whole system or set of HR practices has to be implemented effectively.
This raises two issues. The first is what is meant by good HRM. The second
is where any practitioner should start. Discussion of an HR system is often
linked to the idea of a ‘bundle’ of HR practices. These are typically defined
as key combinations of practices. However, there remains an ambiguity in the
literature about whether bundles should be considered at the level of specific
practices, combinations of practices, or a broadly based approach reflecting a
philosophy of HRM (Delery 1998). Without some clarification, there remains
uncertainty about the level at which to consider application of HRM. More
specifically, if researchers are advocating the need to apply a ‘system’ how
does anyone introduce a system? It is conceivable that this can be done in
the context of greenfield sites; but for a manager who has heeded the message
and wants to get going, where does he/she start? Put another way, to what
extent is it ethical to advocate an approach that many may find impractical to
introduce?
A pragmatic way to address this issue is through some form of statistical
analysis to identify which combination of practices is most likely to be consis-
tently associated with outcomes. This might be achieved through regressions
taking each HR practice rather than some combination of practices as the
independent variables. A variant on this that has been applied to HRM is the
use of sequential tree analysis (Guest, Conway, and Dewe 2004). This builds a
‘tree’ by identifying the key practices and then seeking the best combination
of practices. Studies of this sort, as well as the more qualitative work of Purcell
et al. (2003) point to the importance of job design and of providing scope
for autonomy and discretion as a key practice associated with both superior
performance and job satisfaction. Yet it is not clear that job design typically
falls within the domain of HR managers. For example, it is relatively infre-
quently cited as one of the practices being implemented by them and their
organization. There is a risk, therefore, that recommending that priority be
given to job design as a starting point may serve to disempower HR managers.
Does partnership resolve the ethical issues?
The body of research evidence does show some sort of association, some of
the time, between HRM and some aspects of performance as well as with
job satisfaction and commitment. Put another way, there is evidence that
whatever its shortcomings, many workers prefer to work in an organization
that practices a form of high-commitment HRM to the alternatives. Further-
more, since managers will continue to seek competitive advantage, it may be
better, from a worker’s perspective, if they do so by pursuing HRM rather than
some of the less palatable approaches that were in vogue during the 1990s