This study established propositions and hypotheses based on two competing perspectives—the universalistic perspective and the contingency (fit) perspective—regarding the configurational characteristics of human resource management (HRM) systems that affect organizational performance, and empirically tested them. According to the universalistic perspective, the effect of an HRM system on organizational performance is proportional to the extent to which high-performance HRM practices are utilized within that system. In contrast, according to the contingency perspective, which considers interactions among individual HRM practices, the effect of an HRM system on organizational performance varies depending on the degree of fit among the individual HRM practices comprising the system. The empirical test results showed that hypotheses based on the universalistic perspective were not supported, while hypotheses based on the contingency perspective received empirical support. The distinguishing features of this study compared to prior research are that, in classifying HRM systems, a broad range of HRM practices were included in the analysis rather than only a small number of high-performance HRM practices, and that data were collected at two time points to attempt a longitudinal analysis.