Human resource management and organizational performance. (Measuring Organizations: New Approaches)

The traditional model of human resource management systems in the United States, based on bureaucratic control mechanisms and designed for mass production, has come under attack in recent years. Pressures deriving from intensified foreign competition, rapid technological change, greater needs for innovation, and workers' demands for empowered jobs have led some American organizations to search for alternatives to this traditional model. Theorists have developed the idea of a transformed or high-performance work system in the United States that represents a composite of several models which are alternatives to mass production (Appelbaum and Batt 1994; Bailey 1992; Kochan and Useem 1992; Lawler 1992; Osterman 1992).

Despite the importance and timeliness of this topic, there are few empirical studies of the effects of high-performance human resource policies and practices on organizational performance(1) that are based on representative samples of diverse work organizations. The few extant national-level surveys of the diffusion of various work reform and employee involvement practices(2) generally do not examine the relationship between human resource management (HRM) practices and organizational performance (see Appelbaum and Batt 1994; Bailey 1992). Some evidence for this comes from case studies and from nonrepresentative surveys (i.e., transformed organizations are more likely to respond) conducted by membership organizations, consulting firms, and other private industry sources. But both these nonrepresentative surveys and case studies vary in quality, leading Appelbaum and Batt (1994) to conclude that "despite the widespread interest in work reorganization, our understanding of what has taken place in American workplaces still is poor. . . . With few exceptions, careful studies . . . have not yet been undertaken".

The National Organizations Study (NOS) data enable us to take a look, albeit a very incomplete one, at the extent to which U.S. work organizations' human resource management practices display features of a high-performance work system (HPO). In this article, we examine the relationships between the few HPO characteristics that are available in the NOS, on one hand, and organizational performance, on the other. We first summarize a portion of the vast and growing literature on HPOs and identify some of the key HRM policies and practices that are commonly associated with this organizational form. We then discuss the measures of these HRM characteristics that are available in the NOS-GSS data set, and assess how they cluster, or produce discrete groups of organizations in the NOS. We next relate these HPO structures to measures of organization performance to test hypotheses about whether organizations that have features commonly associated with HPOs actually perform better than other organizations.

CHARACTERISTICS OF HIGH-PERFORMING ORGANIZATIONS IN THE UNITED STATES

Writers contrasting the ideal-type old and new forms of human resource management systems use different labels to capture this distinction (see Bailey 1992), including mass versus flexible production (Piore and Sabel 1984), industrial versus salaried (Osterman 1988), old competition versus new competition (Best 1990), conflict versus commitment (Walton 1985), cost reduction versus commitment maximizing (Arthur 1992), and high-performance work organization versus mass production-low wage organization (Commission on the Skills on the American Workforce 1990). These labels all point to several basic dimensions that constitute elements of a model of a high-performance work organization. These components of HPOs are summarized in Table 1.

These systems have four main components: (a) management methods (market strategy, organization structure, overall process approaches--e.g., total quality management or just-in-time manufacture); (b) work organization (design of shop floor or frontline jobs and deployment of workers, including such practices as job rotation and teamwork); (c) human resource practices (training, compensation, and strategies to induce worker effort and commitment, such as employee involvement and employment security); and (d) industrial relations (relationships between managers and workers and union-management relations).

The NOS data set was not designed explicitly to measure HPOs, and it contains better …