improvements in output at the margin. These practices may have direct and
indirect impacts on productivity or output—direct by, for example, improving
the selection and thus the quality of labor, indirect by changing worker morale
and thereby eVort levels. A standard economic framework (the Cobb–Douglas
production function, see above), whereby additional units of an HR practice are
adopted provided the marginal revenue exceeds the marginal cost, is then used to
explain why not all organizations are interested in developing high-commitment
systems and to move beyond the assertion of the existence of synergies between
the diVerent elements of the HR package in the HPWS literature into an empir-
ically testable hypothesis. This approach serves to introduce a healthy note of
scepticism as to the eYcacy of universal HR best practice bundles, but the
framework assumes that the costs and beneWts of HR practices are known and
calculable. Nor is it able to deal with the issues of long-term and strategic
Xexibility (Wright and Snell 1998 ; Purcell 1999 ) which may require the building
in of a capability to respond to future needs.
A second example of an integrated approach to choice of HR strategy is the HR
architecture model provided by two HR theorists, Lepak and Snell ( 1999 ), who
‘draw on the resource-based view of theWrm, human capital theory, and transac-
tion cost economics to develop a HR architecture of four diVerent employment
modes: internal development, acquisition, contracting and alliance’ ( 1999 : 31 ). Two
variables explain the choice of HR practice—the value of skill and the speciWcity
(or uniqueness) of skill. The market versus hierarchy interface between HRM and
economics is expanded into a richer, more multilayered approach that distin-
guishes usefully between the value and the speciWcity of skill and between relational
and transactional contracting.
Following the personnel economics and HRM traditions, the focus is on describ-
ing practices within the organization and not on the interactions between HR
policy and the operation of the labor market. For example, in deciding between
making or buying skilled labor (internal development or acquisition), the institu-
tional arrangements that produce a supply of ready skilled labor are not consid-
ered. As economists have demonstrated (Marsden 1986 ), an eVectively functioning
occupational labor market (where a ready supply of skilled labor can transfer
between organizations) requires that there is an institutionalization of systems of
training, skill-based job titles, and occupational structures. DiVerences in make/
buy decisions between organizations, sectors, and countries may therefore depend
more on the availability of ready trained labor than on the importance of the
uniqueness of skill. Another problem with the HR architecture approach is its focus
on the value of skill and not on the interactions between diVerent job categories.
According to Boxall ( 1998 ), the strategic HRM literature has focused on the
contribution HR makes to strategic goals rather than operational eYciency. In
Lepak and Snell ( 1999 ), external contracting is proposed where human capital ‘is
generic and of limited strategic value’ and can therefore be ‘treated essentially as
economics and hrm 77