Horizontal and Vertical Fit in Human Resource Systems

Human resource (HR) management has traditionally focused on how employment-related
policy decisions, such as selection, development, training, compensation, and work design,
influence organizational effectiveness. Until the last decade or so, effectiveness was almost
exclusively studied at the individual level of analysis (e.g., using individual performance ratings
as the dependent variable). However, the field of HR has evolved such that significant emphasis
is now placed on also understanding linkages between HR practices and effectiveness at the
aggregate level (Becker & Gerhart, 1996), including facility level outcomes such as cost,
productivity, quality, as well as firm level outcomes such as total shareholder return,
profitability, and survival.

The term, HR system, refers to the fact that different combinations of individual HR
practices are possible and a key question is whether the effects are additive or, as hypothesized
by several perspectives, is the effect of any one HR practice dependent on the nature of other HR
practices? A second key question is whether the specific effect of any particular HR practice or
HR system depends on contextual factors, such as organizational strategy. These two questions
have to do with the issues of horizontal (or internal) fit and vertical (or external) fit, respectively

 Although my focus is primarily on the business performance dependent variable, I

recognize that effectiveness can be defined more broadly, for example, using a stakeholder
perspective where, at a minimum, shareholder, customer, and employee outcomes are
considered. Clearly, HR systems can differ in how favorable their consequences are for different
stakeholders. Some HR systems may be of the mutual gains nature (Kochan & Osterman, 1994),
while others may create more of a zero-sum situation. I operate under the assumption that that
some minimally competitive level of business performance is typically good for everyone.
Without it, there will not only be less money to be made by investors and less satisfied
customers, but also fewer jobs to be had by workers. For a discussion of how HR systems may
affect worker outcomes, see Godard and Delaney (2000) and the following comment by Kochan
(2000). Several recent empirical studies are now also available (Godard, 2001; Handel &
Gittleman, 2003; Hunter & Lafkas, 2003; Parker, 2003).

Our goals in this chapter are to (a) review the conceptual basis and measurement of HR
systems in the literature and their evolution, (b) consider evidence on the consequences for
effectiveness of horizontal and vertical fit of HR systems, and (c) provide suggestions for future
research on these issues, as well as a model of HR systems and business performance. I begin
3
with a survey of models and measurement of HR systems and its conceptual basis.

HR Systems

Conceptual Basis

As noted, HR systems have been most often defined in terms of policies (e.g., staffing,
job design, and so forth) and “system” refers to the particular array or combination of HR
practices in an organization. Typically, an HR system is seen as most directly influencing what
might be termed intermediate employment effectiveness outcomes, such as ability, motivation,
attitudes, and performance. (Outcomes such as attraction and retention could also be added.)
These intermediate employment outcomes, in turn, are expected to influence business
performance outcomes. This mediation model, while fundamental to the HR systems literature,
is also largely untested.
AMO Framework. A general framework for conceptualizing and studying HR systems
has emerged in the recent literature on “high performance work systems” (HPWs). Although
Boxall and Purcell (2003) observe that “the definition of components of HPWSs is confusingly
varied” (see also Becker & Gerhart, Table 2), they find an increasingly common “very basic
theory of performance” being used, which they refer to as “AMO theory.” Boxall and Purcell
summarize it as:
P = f(A,M,O)
where P is performance, A is ability, M is motivation, and O is opportunity. In other words, HR
systems will be most effective when they foster abilitiy, motivation, and opportunity to
contribute to effectiveness.
The AMO logic is most clearly spelled out in Appelbaum et al. and Bailey (1993) and
earlier by Katz, Kochan, and Weber (1985) [“many theoretical arguments have suggested that
individual worker ability, motivation, and participation in job-related decision making affect
both organziational effectiveness and individual worker satisfaction”, p. 513]. Recent empirical
studies (e.g., Appelbaum et al; Batt, 2002; Huselid, 1995; MacDuffie, 1995) use this conceptual
framework or something akin to it.
Boxall and Purcell (2003) note that the AMO model is actually an “an old rubric.” This
is certainly correct in that each of the components has been studied extensively (see below),
albeit with different degrees of emphasis depending on the discipline/field. However, the clarity
and parsimony of the AMO model is helpful, and the equal importance it assigns to the A, M,
4
and O components in its holistic approach is fairly novel. (Previous literatures have typically
focused on one or two of the components.) Given the multidisciplinary (and thus somewhat
dispersed) basis for the AMO framework, a brief review of the basis for each component is in
order.

The A component can be traced most directly to the industrial/organizational (I/O)
psychology and economics of human capital (HC) literatures. The former provides extensive
evidence that individual abilities strongly predict individual job performance (Heneman & Judge,
2003; Schmidt & Hunter, 1998). The HR system policy implication is that employee selection
practices will be most effective when valid predictors like ability are used and when employers
are selective (Brogden, 1949). Human capital theory (Becker, 1964) and research (e.g., Mincer,
1974) hold that investments in people (e.g., education, training) make them more productive and
that the choice by individuals and employers to invest depends on their expected rate of return.
The I/O psychology literature also gives significant attention to training and development (e.g.,
Noe, 1999).

 The M component is addressed in many literatures (Gerhart & Rynes. 2003). Williamson

et al. refers to the managerial goal of obtaining behavioral “consumate cooperation.” As Simon
(1951; see also Barnard, 1951) pointed out, this depends on workers’ view of their exchange
relationship with the employer: employees provide contributions based on their view of
inducements, monetary and otherwise, provided by the employer. Recent work on different
forms of the psychological contract (Rousseau & Parks, 1992) and of the employment
relationships (Lepak & Snell, 1999; Tsui et al., 1997) uses a similar logic.

 The I/O psychology literature on motivation includes models such as goal-setting (Locke

& Latham, 1990), expectancy (Campbell & Pritchard, 1976, Lawler, 1971, Vroom, 1964), and
equity theories (Adams, 1965), which focus primarily at the individual level and on the
psychological mechanisms that explain motivation (Gerhart & Rynes, 2003). Much of this work
is conducted in laboratory settings. Work that is more grounded in the economics literature such
as efficiency wage (Yellen, 1984), transaction costs (Williamson, 1975), and agency theories
(Fama & Jensen, 1973) typically excludes psychological mechanisms (in empirical work at least)
and focuses more on compensation policies in work organizations.

 One of the more explicit discussions of the A and M components is found in Vroom

(1964, p. 203), who states that “the effects of motivation on performance are dependent on the
5
level of ability of the worker, and the relationship of ability to performance is dependent on the
motivation of the worker.” In other words, “the effects of ability and motivation on performance
are not additive but interactive” (p. 203). Thus, Vroom proposed the following formula:
Performance = f(Ability x Motivation) or P = f(A, M).
Vroom observed that establishing the validity of this formula had “considerable
implications for managerial practice” (p. 203) because “It would suggest that managerial efforts
to obtain and develop persons with skill and ability and to motivate these persons must proceed
concurrently” (p. 203). This is the same logic that characterizes recent HR systems models that
propose the importance of horizontal/internal fit or bundling.
The similarity of the Vroom P = f(A, M) formula to the AMO formula of P = f(A, M, O)
is obvious as is the lone difference: the lack of an O component in Vroom’s formula. Like
Vroom, Campbell’s (1990) I/O Handbook chapter focused on the A and M components as
determinants of job performance (e.g., his Figure 3, p. 707) and used the simplifying assumption
that situational effects are “held constant” (p. 707). This line of work on performance prediction
(e.g., as summarized by Campbell, 1990) led Peters and O’Connor (1980, p. 391) to observe that
situational constraints on performance was “a frequently overlooked construct.” Even here, the
situation was seen more as a constraint on M and A, less as an important contributing factor in its
own right to effectiveness. Although I/O psychology (mostly the ‘O’ part of the I/O field) has
given substantial attention to job design (MacGregor; Turner & Lawrence; Hackman and Lawler,
1971; Hackman & Oldham, 1976), this literature does not seem to have been strongly linked to
that part of the I/O literature (mostly the ‘I’ part) focused on performance as the dependent
variable.

To get any assignment help. Please chat with online assignment adviser