You are on page 1of 25

Journal of Operations Management 21 (2003) 19–43

The impact of human resource management practices on operational performance: recognizing country and industry differences
Sohel Ahmad a,∗ , Roger G. Schroeder b,1
b a Department of Management, St. Cloud State University, 720 Fourth Avenue South, St. Cloud, MN 56301-4498, USA Department of Operations and Management Science, Donaldson Chair in Operations Management University of Minnesota, Carlson School of Management, 3-140 CarlSMgmt Building, 321-19th Avenue South, Minneapolis, MN 55455, USA

Received 17 August 2000; accepted 14 January 2002

Abstract The interest in strategic human resource management (HRM) has spawned a number of empirical research studies that investigated the impact of HRM practices on organizational performance. However, very little attention has been paid to address the impact of HRM practices on operations management and to generalize the findings across countries and industries. Success of some business decisions (e.g. globalization and merger and acquisition) necessitates recognition and reconciliation of the differences among HRM practices in different countries and industries. This study attempts to generalize the efficacy of seven HRM practices proposed by Pfeffer in the context of country and industry, focusing primarily on the effects of these practices on operations. The findings provide overall support for Pfeffer’s seven HRM practices and empirically validate an ideal-type HRM system for manufacturing plants. © 2002 Elsevier Science B.V. All rights reserved.
Keywords: Human resource/OM interface; Strategic human resource management; Staffing; Operational performance improvement

1. Introduction Human resources are considered the most important asset of an organization, but very few organizations are able to fully harness its potential. Lado and Wilson (1994, p. 701) define a human resource system “. . . as a set of distinct but interrelated activities, functions, and processes that are directed at attracting, developing, and maintaining (or disposing
∗ Corresponding author. Tel.: +1-320-255-2994; fax: +1-320-255-3986. E-mail addresses: ahmad@stcloudstate.edu (S. Ahmad), rschroeder@csom.umn.edu (R.G. Schroeder). 1 Tel: +1-612-624-9544; fax: +1-612-624-8804.

of) a firm’s human resources.” Traditionally, management of this system has gained more attention from service organizations than from manufacturing organizations. However, to enhance operational performance, effectively managing this system is equally important in both types of organizations. Needless to say, sophisticated technologies and innovative manufacturing practices alone can do very little to enhance operational performance unless the requisite human resource management (HRM) practices are in place to form a consistent socio-technical system. For this reason, manufacturing organizations need to carefully evaluate their existing HRM practices and modify them, if needed, so that employees can effectively contribute to operational performance improvement.

0272-6963/02/$ – see front matter © 2002 Elsevier Science B.V. All rights reserved. PII: S 0 2 7 2 - 6 9 6 3 ( 0 2 ) 0 0 0 5 6 - 6

20

S. Ahmad, R.G. Schroeder / Journal of Operations Management 21 (2003) 19–43

Several studies in the HR literature investigated the impact of HR practices on organizational performance. Although some studies related to HR practices can be found in the operations management literature (Jayaram et al., 1999; Kathuria and Partovi, 1999; Youndt et al., 1996; Kinnie and Staughton, 1991), this discipline has tended to address structural issues and analytical questions, and has paid little attention to human resources issues. A review of empirical articles published between 1986 and 1995 in 13 OM research outlets revealed that less than five percent of these articles fell into the “HRM for operations” category (Scudder and Hill, 1998). This lack of attention is surprising when one considers human resources’ critical role in achieving superior performance in competitive priorities, such as low cost, quality, delivery, flexibility, and innovation. Over the years, researchers have suggested many HRM practices that have the potential to improve and sustain organizational performance. These practices include emphasis on employee selection based on fit with the company’s culture, emphasis on behavior, attitude, and necessary technical skills required by the job, compensation contingent on performance, and employee empowerment to foster team work, among others. Pfeffer (1998) has proposed seven HRM practices that are expected to enhance organizational performance. The practices proposed by Pfeffer (1998, p. 96) are: 1. Employment security. 2. Selective hiring of new personnel. 3. Self-managed teams and decentralization of decision making as the basic principles of organizational design. 4. Comparatively high compensation contingent on organizational performance. 5. Extensive training. 6. Reduced status distinctions and barriers, including dress, language, office arrangements, and wage differences across levels. 7. Extensive sharing of financial and performance information throughout the organization. There are several objectives of the present study based on these practices. First, we investigate whether manufacturing plants’ use of these seven practices differs by country or industry. Next, we assess the impact of each of these practices on organizational

performance which includes (1) operational performance measures: unit cost, quality, delivery, flexibility, and speed of new product introduction and (2) an intangible performance measure: organizational commitment. Lastly, we examine whether these seven practices can form a synergistic HR bundle to represent an ideal HRM system for manufacturing plants and check the efficacy of this ideal system. Since the manufacturing plant is the unit of analysis for this study, we will be testing the HRM theory at the plant or operations level of the organization.

2. Theoretical background and hypotheses Organizations can internalize as well as externalize employment (Lepak and Snell, 1999). Internalization of employment involves building an employee skill base inside the organization, while externalization of employment means outsourcing human resource needs to market-based agents (Rousseau, 1995). Each alternative has its own costs. According to the transaction cost theory (Williamson, 1975), the decision to internalize or externalize a part or all of an operation’s human resource needs should be based on the transactional costs involved. Arriving at a HR outsourcing decision in such a manner is myopic as it overlooks the strategic consequences. For example, outsourcing human resource needs can minimize bureaucratic costs and complexities. However, an operation’s continued dependence on external sources may inhibit its ability to develop core skills and capabilities vital for longterm survival in the marketplace (Lei and Hitt, 1995). The human capital theory recognizes employee skills, experience, and knowledge as assets with the potential to generate economic rent (Coff, 1997). However, this theory evaluates human resources through productivity gains. It falls short of attaching strategic value to causal ambiguity and tacit knowledge embedded within an organization’s human resource system. In recent years, researchers and practitioners have realized that HRM systems can be used as strategic levers to focus on value creation that goes beyond traditionally emphasized cost reduction (Becker and Gerhart, 1996). Drawing on a behavioral psychology perspective, researchers have highlighted the strategic aspect of HRM practices and argued about why these practices can lead to competitive advantage (Schuler

Pfeffer’s seven HRM practices are internally consistent with one another. (c) use of self-managed teams and decentralization. 1993. Huselid. The impact of HRM practices on organizational performance as proposed by Pfeffer (1998) and others can be generalized across manufacturing plants operating in different industries and countries if we find support for the set of hypotheses below. 829) raise concern that the results of their study of HRM practices on organizational performance in the banking industry may not be valid in other industries. Much of the previous research on the relationship between HRM practices and organizational performance has concentrated on a single HR practice. abilities. identifying these qualities should be an integral part of the hiring process. Several researchers (Delery and Doty. 1995.G. 1994) including Pfeffer (1998) have argued why these practices are expected to enhance organizational performance. H1.” Similar cautions can be made for country effects. The literature has emphasized the need for generalizability of the relationship between HRM practices and organizational performance. if they intend to keep employees loyal to them. Organizations need to make extra efforts to reduce status distinctions. etc. an organization promising employment security needs to pay close attention to selective hiring of new personnel. . empirical validation of the findings in operations across countries and/or industries is nearly non-existent and very limited at best. 1995.. After controlling for the industry and country effects. Effectively operating self-managing teams and decentralizing decision making require in-depth understanding of aptitudes. p. the current findings need to be validated in other industries to rule out industry as an important contingency factor. training related expenditures should be strategically evaluated and considered an investment toward human capital rather than merely a cost of doing business. values. idiosyncrasies. (d) use of compensation contingent on organizational performance. Schroeder / Journal of Operations Management 21 (2003) 19–43 21 and Jackson. goals. (Gerhart and Milkovich. Osterman. 1987. The resource-based view of the firm further advanced this argument by stressing that the tacit knowledge embedded in firm-specific human resources is hard to imitate because of its attributes. Ahmad. . . Mutual understanding among employees usually develops when they work together for a long time as occurs in organizations that provide employment security. (b) selective hiring. Barney. “Human resource practices are said to be bundled when they occur in fairly complete. 1998). 1996. 1991). a growing number of researchers have argued for instituting complementary bundles of HRM practices to enhance organizational performance (Ichniowski et al. The impact of HRM practices on organizational performance has been the subject of much attention over the years. Under a longer time horizon. and behavior fit with those of the organization. Recent trends toward globalization and mergers and acquisitions in the business world make the study of HRM practices in the context of country and industry a necessity (Legare. The seven practices suggested by Pfeffer (1998) are expected to foster such inimitable attributes in human resources and. R. Sharing information on organizational strategy. and performance with employees conveys that they are trusted. 1984). and personal traits of fellow employees. Osterman. and causal ambiguity (Russell. Also. These authors urge that “. However. (f) reduced status distinctions and (g) sharing of information. path dependency. 657). organizational performance will be positively related to each of the following seven HRM practices: (a) employment security (alternatively. such as compensation. For example. Information sharing also empowers the employees and fosters organizational transparency which are crucial if the employees are to have long tenures in the organization. However. 1997. selection. temperaments. it makes sense to invest more in training these employees. Schuler and MacMillan. 1995. when an organization institutes performance contingent compensation. mutually reinforcing or synergistic sets” (Dyer and Reeves.S. (e) the ex- tent of training. such as asset specificity. Collis and Montgomery. Delery and Doty (1996. p. For example. 1994. therefore. social complexity. 1990). the employees are motivated to focus on long-term organizational performance rather than short-term gains if the employers provide employment security. thereby help an organization attain competitive advantage. Arthur. employment insecurity is negatively related to organizational performance). Organizations emphasizing employment security intend to keep employees longer. 1994). Employees cannot be retained for a long time unless their attitudes. Therefore.

The reliability and validity of the constructs were formally tested using data from over 800 respondents in a prior round of data collection in 43 US plants. Japan. The focus of the WCM project is to examine differences in manufacturing practices across plants in different countries and industries (Flynn et al. Whenever possible.22 S. We use a part of this project’s database that addresses HRM issues. 1995). As a result of these tests. then an organization with a HRM system similar to the ideal-type HRM system will explain significantly more variation in organizational performance than any of the individual HRM practices or any combination thereof. The mean age of these plants is about 37 years. . and implement each of these seven practices. The average facility size (production and warehouse) is 160.G. Italy. including both salaried and hourly workers. This ideal-type HRM system is expected to yield the highest organizational performance. 3. Therefore. Asia and Europe. North America. some of the items were clarified or changed to be more representative of the intended constructs. In each of these countries. The four countries were selected to represent the major industrial regions of the world. institute. Schroeder / Journal of Operations Management 21 (2003) 19–43 Table 1 Number of plants by country and industry Electronic Germany Italy Japan USA Total 5 8 13 6 32 Machinery 10 13 12 5 40 Automobile 9 7 14 5 35 Total 24 28 39 17 107 As explained above. Data collected from plants operating in four countries and three industries are used for the empirical analyses. A bundle of internally consistent practices is more effective than the sum of the effects of the individual practices due to their mutually reinforcing support (MacDuffie. Such a HRM system may be termed an ideal-type HRM system. and the USA. We wanted industries that had been threatened by global competition and thus were seeking improvements. these HRM practices are internally consistent with one another and qualify as a synergistic set. if bundling invokes synergy among HRM practices as previously argued. such as total quality management (TQM). plants were randomly selected from three industries: automobile. After the pilot testing. The more similar an organization’s HRM system is to the ideal-type HRM system. and machinery. Face validity of the questionnaires was insured by having three different researchers develop items for the scales. After controlling for the industry and country effects. The countries are Germany. The three industries were selected because the literature suggests that they have been implementing various WCM approaches. scales were selected from the existing literature. Every organization differs in how much effort it puts into harnessing each of the seven HRM practices. The resource-based view also supports this notion by stressing that individual practices have a limited ability to generate competitive advantage in isolation.701 ft2 . in combination. with 32 product lines manufactured on average. R. The response rate for this project was about 60%. and employee involvement (EI). The three researchers then reviewed all of the items for content validity. From the above discussion. Similar arguments can be made about each of the remaining six HRM practices.. just-in-time (JIT). These plants employ 1153 employees on average. typically when an organization exerts the maximum effort possible to develop. it includes 107 manufacturing plants (see Table 1) after eliminating responses with missing data. H2. 1996). 1995). Ahmad. An ideal situation may be one in which each of these HRM practices is explored and exploited to its highest potential. The data collection instrument was pre-tested using 10 industry experts and academics. the degree of dissimilarity (measured as misfit) between an organization’s existing HRM system and the ideal-type HRM system will be negatively related to the organizational performance. these complementary resources can help a firm attain greater competitive advantage (Barney. However. electronics. Data collection We use world class manufacturing (WCM) project data to test the hypotheses. we draw the following hypothesis. the better the organization’s performance. Moreover. “employment security” is internally consistent with other HRM practices.

A value of 2 indicates that the plant does not use any group incentive or profit sharing plans. Measures 4. Each scale was evaluated for its reliability and unidimensionality. Each item of a construct was answered using the following five-point scale: strongly agree (5). Therefore. Organizational performance Past empirical research has mostly investigated the effects of HRM practices on financial performance (cf. The list of scales includes: MFGHRFIT. the most desirable value for this variable is 4. MULTFUN. TEAMS. Schroeder / Journal of Operations Management 21 (2003) 19–43 23 some of the scales were significantly revised. 4. JOBSKILL. The variable CONTCOMP is a composite of two binary measures. We used responses from different people for the dependent (organizational performance) and independent (HRM practices) variables to avoid common respondent bias. BEHAVIOR. Huselid. After purifying a scale. and FEEDBACK. the variable STATDIFF is a composite of four binary measures. while a value of 4 indicates that the plant uses both.2. any variable measured by the scale can range in value from one to five. Ahmad. See Table 3 for details. A value of Cronbach’s alpha of 0. agree (4). The remaining three variables in Table 2 were measured using objective measures. we averaged all of the items in that scale. The US instrument was subsequently translated into German. and ability to provide accurate and unbiased answers to the questions in the survey. the most desirable value of this variable is 0. Similarly. 1995). neutral (3).S. As mentioned earlier. The seven HRM practices Table 2 summarizes the variables used and the methods employed to measure the seven HRM practices. About 60% of the plants contacted participated in the study. Plant managers were contacted by a member of the WCM team and asked for their voluntary participation in exchange for detailed feedback regarding their manufacturing practices in comparison to the industry. as such. Interested plant managers appointed plant research coordinators who maintained contact with the research team. Hence. and its value can range from 4 to 8. engineers. INCENTOB. This variable measures job insecurity rather than job security. The value of this variable can range from 2 to 4. We removed an item if it did not contribute strongly to the alpha value and if its content was not essential for the construct. The higher the value of the variable STATDIFF. some are measured using multiple variables as determined by the scope of the HRM practice and limitations of the WCM database. 4. These scales closely approxi- mate the definition of the seven HR practices being measured. experience. the content validity of a construct was ensured through pre-testing of the questionnaires and structured interviews with the managers and academic experts in the field. For details on the measurement refer to Appendix A and Table 3. However. The research team consulted with the plant research coordinators to identify the right respondents in the plant who had pertinent knowledge. the higher the status differences. The last column of Table 2 lists the most desirable value for each variable. 1978). INTERACT. The variable INSECURE was measured as a percentage of employees laid off during the past 5 years. Italian and Japanese. These plant research coordinators were managers who had at least 3 years of experience in the plants and were knowledgeable about the major responsibilities of the employees working in the plant. Delery and Doty. R. The foreign language version was then translated back into English by another individual and compared for accuracy. very few studies have examined the impact of HRM practices on operational performance measures. Any discrepancies were resolved. A set of Likert scales was used to measure pertinent constructs. STRATCOM. which became the value of the variable representing the construct. and strongly disagree (1). 1996) and some on efficiency and employee turnover (cf.7 or more was used as a criterion for a reliable scale (Nunnally.G. such . where five is the most desirable value. the most desirable value for this variable is 4.1. Most of the variables were measured using perceptual scales with a few exceptions where objective measures were used. disagree (2). Therefore. supervisors and workers responded to these questionnaires. The questionnaires were collected in sealed envelopes to maintain anonymity of responses. While most of these HRM practices are measured using one variable. Managers.

Y=Yes=2 and N=No=1 A scale of four items to measure whether the plant’s reward system is consistent with manufacturing objectives and goals A scale of three items to measure if employees’ on the job skills and knowledge are being upgraded in order to maintain a work force with cutting edge skills and abilities A scale of five items to measure the extent to which employees receive cross training so that they can perform multiple tasks or jobs Four questions were asked to judge the use of symbols that indicate status differentials among various employees in terms of the following: the use of assigned parking spots (Y/N). the use of uniforms by workers only (Y/N). R.G. Schroeder / Journal of Operations Management 21 (2003) 19–43 Table 2 Summary of measurements of the seven HRM practices Practice Employment insecurity Variable INSECURE Scales/measurement Employment insecurity Description of measurements (The number of employees who have been laid off during the past 5 years/number of employees in the organization)×100 A scale of six items measuring the degree of cooperation between manufacturing and human resources in designing job descriptions and staffing activities A scale of five items measuring the importance given to a prospective employee’s attitudes and behavior toward teamwork and problem solving during the selection process A scale of five items used to assess the effective use of teams on the shop floor A scale of three items that measures the extent to which supervisors encourage and facilitate workers to work as a team This measure checks whether group incentive plans (Y/N) and profit sharing plans (Y/N) are used in the organization. access restriction to cafeteria for some employees (Y/N). and the use of separate rest-rooms (Y/N) for different employees in the plant A scale of three items to measure the efforts made by management to communicate the plant’s competitive strategy to all employees A scale of five items to measure the extent to which management provides shop floor personnel with information regarding their performance in a timely and useful manner Ideal profile 0 Selective hiring MFGHRFIT Manufacturing and human resources fit 5 BEHAVIOR Behavior and attitude 5 Use of teams and decentralization TEAMS INTERACT Team activities Interaction facilitation 5 5 Compensation/incentive contingent on performance CONTCOMP Contingent compensation 4 INCENTOB Incentives to meet objectives Training on job skills 5 Extensive training JOBSKILL 5 MULTFUN Training in multiple functions Existing status differences 5 Status differences STATDIFF 4 Sharing information STRATCOM Communication of strategy 5 FEEDBACK Feedback on performance 5 . Ahmad.24 S.

G. Schmenner. Hayes and Wheelwright. 1994). the rate of new product introduction has also been included in this list (Vickery et al.S. 791). 1978. such as organizational commitment (cf. 1981. We performed factor analysis to check if these five operational performance measures formed different groups. R.2. the strategic implications of HRM practices make tracking intangible performance measures important. and flexibility. 1996. Because the unit of analysis for this study is a manufacturing plant. Also. 4. 1989) have proposed a wide variety of operational performance measures for manufacturing facilities. but in each case the focus should be on variables that have inherent meaning for a particular context” (Becker and Gerhart. Kalleberg and Moody. We. p. Operational performance measures Researchers (Wheelwright. MacDuffie. 1995) or intangible performance measures. 1997). Ahmad. Lately. The factor analysis revealed that all of these measures loaded on .1. Hill. Schroeder / Journal of Operations Management 21 (2003) 19–43 Table 3 HRM practices measured using objective measures 25 as quality. These include cost. quality. investigate the impact of HRM practices on operational performance measures as well as the intangible performance measure defined below. 1984. cost or delivery (cf. delivery. we argue that HRM practices will impact the operational performance measures at the plant level. therefore. “The appropriate dependent variable will vary with the level of analysis..

Intangible performance measure Researchers have yet to reach a consensus about how best to define strategic HRM. Measure of misfit In context of this paper. The firm has very little control over this human capital as employees may leave the firm or. but the firm does not actually own this human capital. Xk the score of the kth variable of the ideal-type HRM system.e. 4: better than average. Even highly skilled and knowledgeable employees who are uncommitted may not contribute discretionary efforts and will thereby minimize their potential in the organization. they may not be inspired to put forward their best efforts.3. even if they do not leave. the experimental unit is a plant’s existing HRM system as measured by the variables representing the seven HRM practices shown in Table 2. Additionally. n MISFITi = k =1 (Xk − Xik )2 (1) where MISFITi is the distance between the existing HRM system of a particular plant i and the ideal-type HRM system. 2: below average.2. such as organizational commitment are less likely to have strategic impact (Arthur. We identified a theoretical ideal profile by choosing the most desirable values of the variables representing the seven HRM practices shown in the last column of Table 2. we use the following general formula to calculate MISFIT. Snell and Dean (1992) recommend that a firm devise methods to ensure that individuals act in the firm’s best interest over time. This profile represents the ideal-type HRM system that has been theorized to yield the highest organizational performance. . the reliability analysis of these measures yielded a value of Cronbach’s alpha of 0. The number 5: superior or better than average. However. p. See Appendix B for details. Items were dropped to obtain a reliable and unidimensional scale. We. Accordingly. In this study. 1: poor or low end of the industry. 1994). k the number of variables representing the HRM . Ahmad.” Snell and Dean (1992) further stress that a firm invests in employees to strengthen its human capital. HRM practices that fail to elicit specific employee attitudes. Schroeder / Journal of Operations Management 21 (2003) 19–43 Table 4 Operational performance measures (PERFORM) COST Unit cost of manufacturing QUALITY Quality of product conformance DELIVERY On-time delivery performance FLEXBLTY Flexibility to change volume NPDSPEED Speed of new product introduction PERFORM = COST + QUALITY + DELIVERY + FLEXBLTY + NPDSPEED Please circle the number which indicates your opinion about how your plant compares to its competition in your industry. Furthermore. involves designing and implementing a set of internally consistent policies and practices that ensure a firm’s human capital contributes to the achievement of its business objectives. the ideal-type HRM system) and a point representing an experimental unit.G. R. Xik the score of the kth variable of the existing HRM system of a particular plant i.2. an employee with strong organizational commitment will be highly motivated to expend energy on organizational tasks (Anderson et al. identify organizational commitment as an intangible performance measure and measure it using a scale. This composite measure represents a plant’s aggregate achievement in all five areas of performance mentioned above compared to competitors. (1997. Mathematically. to one factor. See Table 4 for details.26 S. 1994).71. Huselid et al. justifying summing up these measures to form a single performance index (PERFORM). 3: average or equal to the competition.. 4. . We conducted reliability and unidimensionality analyses for this scale. misfit represents the dissimilarity between an ideal HRM profile and a plant’s existing HRM profile. therefore. Organizational commitment is an indicator that testifies to whether the HRM practices employed in an organization are able to foster psychological links between organizational and employee goals. The remaining items were then averaged to obtain a score for the scale (COMMIT) corresponding to each plant. misfit is the Euclidean distance between a point defined in a multidimensional space by the ideal profile (i. This is an intangible outcome of a HRM system and is important to retaining employees and exploiting their potential to the fullest extent over time. 172) attest that there is broad agreement in the literature that strategic HRM “. 4.

ITALY (Italy compared to USA). . included the following control variables (indicator variables) in the regression analyses. For example.S. standardized (STD) the squared differences (between the ideal-type HRM system and the existing HRM system of a plant) before adding them together to avoid this problem. Positive correlations among different HRM practices show that when a plant increases its efforts in one of the HRM practices. This analysis is primarily descriptive. it is also more likely to increase efforts in other practices. which allow for some interesting observations. the ranges of the variables in the above equation are not the same. therefore. That is. two industry control variables. although . Analyses and results In this part of the paper. A higher status difference in a plant is associated with lower efforts in other HRM practices. More specifically. a plant with high status difference is expected to have high employment insecurity (i. higher employee layoff rate) since STATDIFF and INSECURE are positively correlated. we first present the descriptive statistics.e. Three country control variables. Next. 4. Employment insecurity (INSECURE) is negatively related to many of the HRM practices which implies that a plant with a high employee layoff rate is less likely to foster growth in other HRM practices listed in Table 5. Ahmad. k varies from 1 to 12 and i varies from 1 to 107. n.G. . analysis to determine if the extent to which plants use the seven HRM practices differs by country and/or industry. standard deviations. and JAPAN (Japan compared to USA). are used to represent the three industries from which the data were collected. Control variables Since we intend to identify impacts of HRM practices on organizational performance that can be generalized across countries and industries. We have. Table 5 shows means. R. MISFIT is calculated as follows: 5. and correlations. for this study. we conduct statistical MISFITi = STD{(0 − INSECUREi )2 } [STD{(5 − MFGHRFITi )2 } + STD{(5 − BEHAVIORi )2 }] + 2 2 [STD{(5 − TEAMSi ) } + STD{(5 − INTERACTi )2 }] + 2 [STD{(4 − CONTCOMPi )2 } + STD{(5 − INCENTOBi )2 }] + 2 [STD{(5 − JOBSKILLi )2 + STD{(5 − MULTIFUNi )2 }] + 2 + STD{(4 − STATDIFFi )2 } [STD{(5 − STRATCOMi )2 } + STD{(5 − FEEDBACK i )2 }] + 2 As mentioned earlier. .4. Canonical correlation analysis is often used to investigate the relationship between two sets of variables. it is unlikely that the HRM practices will flourish in a plant where high status difference exists. This can disproportionately inflate some variables’ contribution to the MISFIT calculation. Schroeder / Journal of Operations Management 21 (2003) 19–43 27 system 1. Similarly. Furthermore. Lastly. the effects of country and industry need to be removed prior to evaluating the relationship between HRM practices and organizational performance. . high variance of the variable INSECURE indicates that plants’ employee layoff rates vary widely. for this study. empirical analyses are performed to test the hypotheses stated earlier. are used to represent the four countries. GERMANY (Germany compared to USA). We. MACHINE (machinery industry compared to electronics industry) and AUTOMOBL (automobile industry compared to electronics industry). The variable that measures status difference (STATDIFF) shows similar results. therefore.

G. R.28 S. Ahmad. Schroeder / Journal of Operations Management 21 (2003) 19–43 .

he found that cultural values varied significantly by country and region of the world. On the other hand. Huselid. 1999). Table 6 shows the canonical cross-loadings for the first canonical pair. and (3) measure of redundancy for the percentage of variance accounted for by the two sets of variables. each of the dependent variables is significantly related to the independent canonical variate (canonical variate representing HRM practices). Although there are no guidelines about the minimum acceptable value for the redundancy index. except for flexibility to change volume (FLEXBLTY). governmental regulations/policies (Morishima. Ahmad. MacDuffie.3896 TEAMS 0.31 is considered significantly different from zero at a level of significance of 0. 1980).4047 JOBSKILL 0. which is quite low. 1995).1128. Researchers have argued that HRM practices can differ across countries and/or industries for several reasons including: cultural idiosyncrasy (Salk and Brannen. particularly where the variables are highly correlated. R. Ichniowski and Shaw. 1998). 2000). (2) magnitude of the canonical correlation. In a single company study. the central foci of these studies were not to compare systematic differences that may have existed in HRM practices in the different countries and industries in which the organizations operated. However.G. Schroeder / Journal of Operations Management 21 (2003) 19–43 29 it can be used for predictive purposes. Some studies used data collected from multiple industries in one country (cf. Canonical cross-loadings have been suggested Table 6 Results of canonical correlation analysis Canonical correlation 0. these weights are subject to considerable instability due to slight changes in sample size. We. it is important to understand the differences that may exist .S. The canonical cross-loadings show the correlations of each of the dependent variables with the independent canonical variate.1128 Correlations between the operational performance measures and the first canonical variable of the HRM practices COST 0. Most of the empirical studies related to HRM practices have been conducted using data collected in a single industry within one country (cf. 1994). Hofstede argues that national cultures impact the attitudes and behaviors of employees (Hofstede. use this method to identify the relationship between HRM practices variables and operational performance measures.56) was moderate. canonical pairs have been interpreted by examining the sign and the magnitude of the canonical weights. 1999).0911 NPDSPEED 0. Arthur.4334 DELIVERY 0..3227 FLEXBLTY 0. A loading of at least 0. Traditionally.4436 FEEDBACK 0. and vice versa.3785 as a preferable alternative to the canonical weights (Hair et al.4625 STATDIFF −0.3328 Correlations between the HRM practices variables and the first canonical variable of the operational performance measures INSECURE −0. such as JIT and quality management (Snell and Dean. all independent variables (HRM practices) except for employment insecurity (INSECURE) and status differences (STATDIFF) are significantly related to the dependent canonical variate (canonical variate representing operational performance measures).4228 CONTCOMP 0.4121 INTERACT 0. (1998). 1995.05 (Graybill.3101 BEHAVIOR 0. Since we intend to identify generalizable impacts of HRM practices on organizational performance across countries and industries (H1).4553 MULTFUN 0.0064 Redundancy index 0. and adoption of managerial practices. 1992). 1961). 1995). As suggested by Hair et al. According to this criterion.3153 INCENTOB 0. The redundancy index was found to be 0. competitive priorities (Boxall and Steeneveld.3912 QUALITY 0. generally the higher the value of the index the better. three criteria were considered when determining the number of important canonical pairs: (1) level of statistical significance of the function.1183 MFGHRFIT 0. Only the first canonical pair was statistically significant (see Table 6).2224 STRATCOM 0. Empirical examination of broad-based HRM practices across industries and/or countries is very limited in the literature (MacDuffie and Kochan. However. and some studies were conducted on data collected from a single industry in multiple countries (cf. 1995). therefore.5603 Level of significance 0. The canonical correlation (0.

We use one-way ANOVA to identify differences in HRM practices among plants operating in four countries. in HRM practices in various countries and industries.40 16. Plants in Japan emphasized some HRM practices significantly more than plants in other countries.88 2. These practices are: behavior and attitude (BEHAVIOR). while plants in Germany lagged behind other countries in this HRM practice.26 3.00 0. 4)∗∗ (3.41 14.00 0. 2)∗ (3. 2)a.86 0. 2)∗∗ (3.36 (1. F-statistics for all of the HRM practices are found to be highly significant except for the scale representing manufacturing and human resources fit (MFGHRFIT).76 3.76 3.00 0.10 3.41 3.83 2. R.00 Pairwise differences F-value Significance NS: not significant.57 71.55 3.00 0. 4)∗∗ (1. The last two columns of Table 7 show the values of the F-statistics and their levels of significance. ITL: Italian plants.19 0. USA: American plants.25 21.40 3. 2)∗∗ (4. + P ≤ 0.65 USA (4) 20.56 2.5 2.70 JPN (3) 0. and Appendix A for definition and measurement of these and other HRM practices.38 3.00 0. 1)∗∗ (3. Ahmad. 2)∗∗ (1.∗∗ (1. Next.13 41.82 4. 1)∗∗ (3. 4)∗∗ (3. communication of strategy (STRATCOM).05. We were surprised by this observation since Germany. Plants in Japan put significantly more effort into training on the job skills.25 3. 1)∗∗ (3. 2)∗∗ (4. 2)∗∗ (1. 2)∗∗ (3.16 11. interaction facilitation (INTERACT). 2)∗∗ 7.09 3.78 2. 4)∗ (3.44 3. and incentives to meet objectives (INCENTOB).74 4. These HRM practices include team activities (TEAMS). contingent compensation (CONTCOMP).17 3. ∗ P ≤ 0. Schroeder / Journal of Operations Management 21 (2003) 19–43 Table 7 HRM practices across countries Practice Countries GER (1) INSECURE MFGHRFIT BEHAVIOR TEAMS INTERACT CONTCOMP INCENTOB JOBSKILL MULTFUN STATDIFF STRATCOM FEEDBACK 42.51 3.97 3. Employment insecurity is the highest in Germany and the lowest in Japan.63 3. Compared to other countries in this sample. The training on job skills scale (JOBSKILL) measures if employees’ on-the-job skills and knowledge are considered important and whether these are upgraded on a regular basis to maintain a work force with cutting edge skills and abilities. 2)∗∗ (3.71 3.87 13.6 3.38 3. 3)∗∗ NS (3.G.13 3. offers apprenticeship training to secondary school students to facilitate the school-to-work transition (MacDuffie and Kochan. 1)+ (3. 2)∗∗ (1. 2)∗ (1.30 5.23 3.54 7.22 ITL (2) 1.77 0.00 0. mean efforts expended by plants differed in all but one of the HRM practices in at least two countries. 3)∗∗ (2.00 0.11 2.17 3.62 3.00 0. ∗∗ P ≤ 0. we conducted the Scheffe pairwise comparison tests of mean differences to better understand how HRM practices differed between each pair of countries.62 3. Our expectation was that German plants would . That is. 2)∗ (3. 2)∗∗ (1. 1995).78 3. a The average percentages of employees laid off in the past 5 years from the plants in Germany and Italy differ at a level of statistical significance of P ≤ 0. We investigate these differences below. 2)∗∗ (4. 2)∗ (3.79 3.29 3.00 0.20 5. under a national industrial and educational policy. The well known lifelong employment policy in Japan seems to be evident in this finding.64 3. 1)∗∗ (4. 2)∗∗ (4. 1)+ (3. GER: German plants.30 S. plants in Italy seem to be significantly lacking in their efforts in several HRM practices.98 10. 4)∗∗ (2. Refer to Tables 2 and 3. 2)∗ (3.32 2. 2)∗∗ (3.00 0.1. 2)∗∗ (4.70 3. This comparison revealed several important aspects of HRM practices as they are used in different countries. 2)∗∗ (1. Statistical insignificance of the F-statistic for MFGHRFIT suggests that the level of cooperation between manufacturing and human resources in designing job descriptions and staffing activities did not differ significantly by country. JPN: Japanese plants.01. and feedback on performance (FEEDBACK).20 3.33 3. training in multiple functions (MULTFUN).55 3.52 2.01. 3)∗∗ (1.59 6.67 3.

36 0.90 3.46 5.64 0.32 3. Ahmad.05.20 2. ∗∗ P ≤ 0.43 3. A closer look reveals that these HRM practices are often emphasized in plants that implement manufacturing practices.29 3. The Scheffe pairwise comparison tests of mean differences revealed that plants operating in the machinery industry seem to put significantly less effort into team activities (TEAMS).76 3.58 0. such as quality management and/or lean production. The automobile industry was at the forefront of the quality management and JIT manufacturing revolutions in past decades (Soderquist and Motwani.35 3. These institutions exhibit an important and pervasive influence on the HR practices employed. 5.35 2.93 2.67 5. Only the two-way interaction effect for the variable INSECURE was found to be statistically significant. Table 8 shows that the F-statistics corresponding to most of the HRM practices are insignificant.70 4. 2)a. and feedback on performance (FEEDBACK) than plants operating in the automobile industry (see Table 8). AUTO: automobile industry.00 Pairwise differences F-value 31 Significance NS: not significant.17 3.61 2.23 0. Italian plants are second.88 0.57 3. Schroeder / Journal of Operations Management 21 (2003) 19–43 Table 8 HRM practices across industries Practice Industries ELEC (1) INSECURE MFGHRFIT BEHAVIOR TEAMS INTERACT CONTCOMP INCENTOB JOBSKILL MULTFUN STATDIFF STRATCOM FEEDBACK 22. National culture. 1999.59 NS NS NS (3.39 3.90 2.1. ∗ P ≤ 0.48 3. we used one-way ANOVA to identify plants’ differences in HRM practices in the three industries.01.85 3.89 5.11 0. 2)+ (2.00 0.26 0.48 3. 1990). interaction facilitation (INTERACT).59 0.1.+ (3. This well-known fact probably explains the difference in HRM practices.99 2. Also.95 0. the plants in the machinery industry exhibited significantly higher status differences (STATDIFF) than those in the electronics industry.13 2. Womach et al. MACH: machinery industry. Again. the two-way interaction effects were also tested using general linear models.01 0. the country . First.G.37 3. we have found that HR practices vary widely by country and to some extent by industry.24 8. We also note that German plants exhibit the highest status differences (STATDIFF) among all of the countries.43 3. 2)∗∗ 1.29 2.05 0. 2)∗ (3. This is consistent with institutional theory when the institutions are taken to be country or industry.96 AUTO (3) 5.02 0. ELEC: electronics industry. a The average levels of effort put in team activities (TEAMS) by the automobile and machinery industries differ at a level of statistical significance of P ≤ 0.1.45 0.36 3. 2)∗∗ NS NS NS (1.50 3. show a similar proclivity toward developing job skills in plants.S.21 0. training in multiple functions (MULTFUN).74 3. 1)∗ NS (3.57 3.53 4. R.29 3.73 3.. This is not surprising given our earlier findings which revealed that this variable showed fairly high variance across countries and industries.71 3.27 MACH (2) 11. Hypothesis 1 This hypothesis is tested by hierarchical regression analyses using PERFORM (Table 9) and COMMIT (Table 10) as dependent variables.47 3. + P ≤ 0.78 3. industry competition and other factors may account for the differences we observed among the HR practices adopted in different countries and industries.72 3. In addition to conducting one-way ANOVA as discussed above. The general perception of work environments in the machinery and electronic industries supports these findings.61 3.06 1.06 0. In summary.

Schroeder / Journal of Operations Management 21 (2003) 19–43 . R.32 S.G. Ahmad.

R. Schroeder / Journal of Operations Management 21 (2003) 19–43 33 .G. Ahmad.S.

5. 1976). Regression analyses show that employment insecurity (INSECURE) and status differences (STATDIFF) were not significant for either of the two dependent variables. (3) linking misfit with organizational performance. the more controlling the employees perceive the compensation system to be. when Z is regressed on X. According to the criteria mentioned above. the less organizational commitment it will engender (Deci. Additionally. Henderson and Lee. both INCENTOB and CONTCOMP had to be significant. This probably explains why we failed to observe a significant relationship between contingent compensation (CONTCOMP) and the intangible performance measure. We. JOBSKILL. These HRM practices have positive associations with the dependent variables.G. (c). hypotheses (b). organizational commitment (COMMIT) acts as a mediating variable for MFGHRFIT. According to the behavioral approach to strategic HRM. the correlation matrix (Table 5) shows that employment insecurity is negatively related to several HRM practices. and AUTOMOBL) were entered. we were surprised that employment insecurity (INSECURE) was not significant. (2) calculating misfit. contingent compensation (CONTCOMP) was not significant for the intangible performance measure. We completed the first . (2) X significantly affects Y. 1986): (1) X significantly affects Z. INCENTOB. and FEEDBACK. 1996). JAPAN. ask the reader to exercise caution while interpreting results related to Hypothesis 1 due to the possibility of omitted variable bias since the correlations between some pairs of HRM practices are quite high. Table 11 shows the results related to mediating effects of the intangible performance measure (organizational commitment). Specifically. Thus. if the following are true (Baron and Kenny. Status difference (STATDIFF) shows a similar relationship with other HRM practices. 1985) is used to test this hypothesis. when Y is regressed on X. (3) Z significantly affects Y. This investigation is conducted as follows. INSECURE and STATDIFF seem to have no impact on operational performance. such as organizational commitment. In order for hypothesis (d) to be fully supported. it can sometimes de-motivate them (cf. when Y is regressed on both X and Z. Here. Therefore. There are three steps in this method: (1) identifying the ideal profile. all hypotheses are supported except for hypotheses (a) and (f). organizational commitment) as a mediating variable in HRM system’s influence on operational performance is worth being investigated. the mechanism through which a HRM system contributes to operational performance is by eliciting behaviors required to accomplish operational goals. CONTCOMP influences operational performance directly. Ahmad. STRATCOM. TEAMS. the analyses conducted to test the direct impact of HRM practices on operational performance and the subsequent analyses for mediating effects reveal the following.34 S.e. Next. Ryan. INTERACT. Therefore. For the dependent variable COMMIT. 1982). MULTFUN. BEHAVIOR. From that standpoint. for the dependent variable PERFORM. Therefore. the results show that most of the HRM practices explain a significant incremental level of the variance. 1972. the HRM practices were independently entered into the equation. providing overall support for this hypothesis. Schroeder / Journal of Operations Management 21 (2003) 19–43 and industry control variables (GERMANY. and (g) are supported and hypothesis (d) is partially supported since the variable INCENTOB is found to be significant but the variable CONTCOMP is not. The literature finds mixed impact of contingent compensation on intangible performance measure. Hypothesis 2 The profile deviation method (Drazin and Van de Ven. However. Kohn. thereby minimizing the potential of the HRM practices as a whole. For each of the dependent variables. In the literature. employment insecurity seems to hinder the development of other HRM practices. (e). however. the term “control” implies management’s attempt to ensure desired outcomes by trying to influence employee behavior (Lawler and Rhode. the role of an intangible performance measure (i. R. and the rest of the HRM practices influence operational performance indirectly through the mediating variable COMMIT. While contingent compensation can sometimes motivate workers to put forward their best efforts (cf. A variable Z is said to be a mediator of a relationship between two variables X (independent variable) and Y (criterion variable). MACHINE. 1992). 1993a) because contingent compensation can be perceived by the employees as a management control mechanism.2. empirical evidence shows that employment insecurity is associated with lower performance (Delery and Doty. ITALY.

22 2.08∗∗ 0.88+ 2.06∗ 0. Next. P ≤ 0.60 2.41∗ 0. Hypothesis 2 will be supported by the regression model below if a significant negative value of β 6 is observed.01.52 0.25∗∗ 0.06 0.21 0.71 1. . Table 12 shows the results of the hierarchical regression analyses.69 2.14∗∗ 0.22∗∗ 2.49∗∗ 1. ORG PERFi = β0 + β1 GERMANYi + β2 ITALYi +β3 JAPANi + β4 MACHINEi +β5 AUTOMOBLi + β6 MISFITi + εi (2) P ≤ 0. where ORG PERFi is the organizational performance of plant i.04 −0.29 2. Two sets of equations correspond to each dependent variable.32∗∗ 35 INSECURE COMMIT MFGHRFIT COMMIT BEHAVIOR COMMIT TEAMS COMMIT INTERACT COMMIT CONTCOMP COMMIT INCENTOB COMMIT JOBSKILL COMMIT MULTFUN COMMIT STATDIFF COMMIT STRATCOM COMMIT FEEDBACK COMMIT + ∗ and second steps earlier. (1)).06∗∗ 0. The country and industry control variables are entered in the first step (Eq. (2) (coefficient) 17.53∗∗ 0. (2) (coefficient) 3.74∗∗ 0.29∗∗ 1.90∗∗ 1. ∗∗ P ≤ 0.07∗∗ −0.06 0.43∗∗ PERFORM (coefficient) 0.43∗∗ 0.62∗∗ 0.41∗∗ 0.36∗ 2.73∗∗ 0.05∗∗ 0. (1) (coefficient) 3. (2).90∗∗ −0.05.38 −0. GERMANYi .30∗ 0.49∗∗ 0.49 16.17+ 0.00 2.46 17.13+ −0.56 0. one at a time.21 4.51∗∗ 0.01 1.08∗ 2.1.40 3.16∗∗ 2.78∗∗ 0.G.00 0.13 3.02 0. Table 12 shows that Eq.23 2.53∗∗ 0. which represents PERFORMi or COMMITi as a dependent variable. (1) (coefficient) Constant GERMANY ITALY JAPAN MACHINE AUTOMOBL MISFIT R2 F Adjusted R2 + ∗ COMMIT Eq.31 2.01 2.76∗∗ 0.06 0.14 0. and JAPANi are three indicator variables representing four countries. MISFITi is the value of the variable MISFIT for plant i.05.01 0.24∗∗ PERFORM (coefficient) −0.S.1.19 1.88∗∗ 0. ITALYi .48∗∗ 0.34 0.86 2. MISFIT is entered into the Eq. Schroeder / Journal of Operations Management 21 (2003) 19–43 Table 11 The mediating effect of organizational commitment (COMMIT) Independent variables Dependent variables COMMIT (coefficient) 0.97 2. The variable MISFIT is found to be negatively Table 12 Results of hierarchical regression analysis of MISFIT on PERFORM and COMMIT Variables PERFORM Eq.04∗∗ 1.68∗∗ 0.47 −0. (2) for each of the two equations is significant and the coefficient of MISFIT (β 6 ) is negative and significant.47∗∗ 0.07 −0.67∗∗ 0. MACHINEi and AUTOMOBLi are two indicator variables representing three industries.09 1. R.17 −0.65∗∗ −0. and the third step is completed here by linking MISFIT with organizational performance according to the following regression model. ∗∗ P ≤ 0.85+ 1.36∗∗ −0.58∗∗ −0. P ≤ 0.14+ 0.55∗∗ 0. Ahmad.04 P ≤ 0.01.97+ 0.17 Eq.09 Eq.33∗∗ 0. thus providing support for Hypothesis 2.

this relationship between MISFIT and organizational performance was observed after controlling for country and industry effects. cannot be empirically validated. we were able to empirically validate those HRM practices that are expected to yield higher performance regardless of the country and industry in which the plant operates. one choice may be to institute these HRM practices for the combined (new) organization. Now. employment insecurity and status difference seem to hinder development of other HRM practices.G. Therefore. practitioners and researchers agree that human resources can be a source of competitive advantage and should be managed strategically. the focus of a HRM system has been short-term. Particularly. this finding suggests that a manager intending to enhance operational performance should create a conducive organizational climate that fosters employees’ commitment to the organization. Several of these mergers and acquisitions are occurring between organizations operating in different countries (e. Another noticeable trend has been mergers and acquisitions among companies. This is particularly important if organizations involved in M&A are following different HRM practices. Hypotheses (a) and (f). Results of the present study show that differences in HRM practices exist in plants operating in different countries. and critical resources. as mentioned earlier. The findings of the present study also offer important implications for several distinct trends observed in the business world today.g. the proposed direct relationship between employment insecurity and organizational performance. Although this was previously implied in the literature.g. and the system has been used as a bureaucratic control mechanism to enhance efficiency (Kalleberg and Moody.. such as human resources. Organizations involved in mergers and acquisitions should take this opportunity to evaluate their existing set of HRM practices and make necessary changes to facilitate post-merger integration. Also. We obtained mixed results when the HRM practices were compared across three industries. however. customers. organizations are discovering this is easier said than done. The mediating effect analysis revealed that most of HRM practices impact operational performance indirectly through organizational commitment. 6. Discussion Traditionally. therefore. 1998. Daimler-Benz and Chrysler Corporation) and industries (e. 2000). These trends pose a unique challenge for HRM (Legare. Many organizations are going through globalization to take advantage of proximity to suppliers. and between status difference and organizational performance. R. While the majority of HRM practices used by plants did not differ by industry. Schroeder / Journal of Operations Management 21 (2003) 19–43 related to the performance measures which implies that as a plant’s HRM system deviates from the ideal-type HRM system its performance suffers. Researchers and practitioners have strongly emphasized that M&A provide a window of opportunity for restructuring HRM practices in the combined (new) organization (Galpin and Herndon. Therefore. comparison of a comprehensive list of HRM practices among countries was lacking. we did find several HRM practices that differed significantly among the three industries. fine-tuning them according to the strategic intent of the . conclude that this ideal-type HRM system is valid for a plant regardless of the country or industry in which it operates. We find overall support for Hypothesis 1 as most of the relationships specified in Hypothesis 1 are found to be significant. However. 1999). which HRM practices should a combined (new) organization choose when M&A is taking place between organizations operating in different industries and/or countries? By controlling for country and industry in our analyses. Additionally. Ahmad. 1994). were not supported for any of the two dependent variables. Our analyses show that plants operating in different industries and/or countries use and emphasize HRM practices differently. This finding indicates that management choices concerning HR practices do indeed make a difference even after accounting for country and industry factors. We can. Therefore. and thereby influence the work environment and minimize the potential of HRM practices as a whole. However. the extent to which some HRM practices are used in plants operating in the machinery industry consistently laged behind that found in plants operating in the automobile industry.36 S. Time Warner and America Online). Lubatkin et al. This finding is important as it refines our understanding of the nature of relationship between HRM practices and operational performance.

a future longitudinal study could focus on the dynamic nature of the HRM practices and uncover the challenges of the implementation process at the plant level. yet only one may successfully attain higher organizational performance because of differences in the implementation process.S. it failed to show the expected level of variation explained. 1995). Based on the literature.G. and this relationship is statistically significant (the coefficient of MISFIT (β 6 ) is negative and significant). R. the variation in organizational performance explained by a HR bundle should be significantly greater than that explained by an individual HR practice in that bundle. 1996). we have useful data from more than twice as many plants operating in Japan (39) than in the USA (17). since we used cross-sectional data. it was not significant for the intangible performance measure. For example. Although the number of plants did not vary greatly among the three industries. Future studies can use objective performance measures at the plant level to check the robustness of our findings. our results empirically validate the proposed ideal-type HRM system because as a plant’s HRM system deviates from the ideal-type HRM system. 1989). Future research can investigate when and why employees perceive contingent compensation as controlling rather than motivating and how this ill effect can be minimized. the findings of our study provide general directions for managers to achieve better operational performance through HRM systems integration in cross-country and/or cross-industry mergers or acquisitions. Taylor. Therefore. the use of objective mea- sures is generally preferred. Ideally. the operational performance measure (PERFORM) could be measured using objective data. Although the use of perceptual measures is quite prevalent in the literature. Although support for Hypothesis 2 in our study empirically validates an ideal-type HRM system. the number of plants varied quite a bit among the four countries (see Table 1). Schroeder / Journal of Operations Management 21 (2003) 19–43 37 new organization. we could suggest little regarding the process of implementation of these practices or the causal relationship between use of these HRM practices and organizational performance. However. Contingent compensation (CONTCOMP) was found to be insignificant for the intangible performance measure. However. Due to the limitations of our data. Existing literature suggests that the level of trust and type of relationship between superior and subordinate may determine whether or not an incentive will be perceived as controlling by the subordinate (Kohn. whether these findings are generalizable across country and/or industry needs . A well-designed research study using longitudinal or panel data can also better address the issue of causality. the plant’s performance decreases. 7. Nonetheless. results of our study failed to show significant increments in variation explained (R2 ) for the HR bundle. Further research is needed to understand how an organization’s strategic context influences the choice of HRM practices and its impact on performance. As a result. However. Thus. and conclusions An important threat to the validity of our findings is the distribution of the number of plants in our sample. Implementing these HRM practices is not an easy task (Pfeffer. Another noteworthy concern is that we used perceptual measures to gauge organizational performance. our results may be more representative of Japanese plants than American plants. While the intangible performance measure (COMMIT) is inherently perceptual. we speculated that employees might have perceived that they were being controlled by this HRM practice. this was not possible due to missing observations. Ahmad. we would have liked to use data from the same number of plants for each country-industry combination. 1993b. hence. future research. Limitations. Therefore. we did not investigate the impact of organizational strategy on these HRM practices. Earlier attempts to empirically validate ideal-type HRM systems have received mixed confirmation (Delery and Doty. while contingent compensation (CONTCOMP) was found to be significant for the operational performance measure. 1994). We empirically showed which HRM practices are expected to enhance performance. Also. Two organizations may correctly identify which HRM practices to implement. This is explained as follows: by definition an HR bundle is a set of interrelated and internally consistent HR practices that are expected to create mutually reinforcing and synergistic impacts on performance (MacDuffie.

The findings of this study are expected to help operations and human resource managers recognize the potential of these seven HRM practices and assist them in designing HRM systems at the plant level to gain superior performance. Their combined and synchronized efforts are needed. Appendix A Scales used to measure HRM practices Variable MFGHRFIT. Item questions The human resources department communicates closely with manufacturing when writing job descriptions Job design at this plant is closely coordinated with manufacturing The human resources department has a close and positive working relationship with manufacturing Staffing. These issues cannot be resolved by isolated efforts made by operations managers or human resource managers. α = 0. we empirically validate an ideal-type HRM system for a manufacturing plant. training and development of employees is closely coordinated with manufacturing Manufacturing works well with human resources staff when changes take place in the manufacturing process Human resources staff knows what manufacturing considers important in the training of employees for new skills We use attitude/desire to work in a team as a criterion in employee selection We use problem-solving aptitude as a criterion in employee selection We use work values and behavioral attitudes as a criterion in employee selection We select employees who can provide ideas to improve the manufacturing process BEHAVIOR.89 Behavior and attitude . Although this was the focal research issue. Cloud State University. Lastly. Delery and Doty (1996) have also reported similar results. The present study brings some of these issues into focus in the context of manufacturing plants operating in different countries and industries.80 Scales Manufacturing and human resources fit Specifically. Future study may shed some light on this matter by theoretically deriving and empirically testing several context specific ideal-type HRM systems. Traditionally.38 S. Despite the compelling theoretical argument. Schroeder / Journal of Operations Management 21 (2003) 19–43 to be investigated. Our study provides empirical validation for the efficacy of the seven HRM practices proposed by Pfeffer (1998).G. α = 0. R. our study failed to show that HR practices are synergistic. This study also evaluates HRM practices taking into account country and industry contexts. thus making the findings generalizable across countries and industries. Ahmad. the present study investigates the mediating effect of organizational commitment which helps us better understand the nature of the relationship between HRM practices and organizational performance. the findings and implications of our study go beyond just testing the potency of Pfeffer’s seven HRM practices. Acknowledgements The first author appreciates the faculty research grant provided by the St. the operations management literature has paid little attention to human resources issues.

α = 0. R.78 Training on job skills . α = 0. we make an effort to get all team members’ opinions and ideas before making a decision Our plant forms teams to solve problems In the past 3 years. α = 0. some employees lack important skillsb Plant employees receive training and development in work-place skills on a regular basis The management at this plant believes that continual training and upgrading of employees’ skills is important Employees at this plant have skills that are above average in this industry INTERACTa . Schroeder / Journal of Operations Management 21 (2003) 19–43 39 Appendix A (Continued ) Variable Scales Item questions We select employees who are able to work well in small groups TEAMS.G.91 Team activities During problem solving sessions. α = 0. Ahmad. many problems have been solved through small group sessions Problem solving teams have helped improve manufacturing processes at this plant Employee teams are encouraged to try to solve their problems as much as possible Supervisors encourage the persons who work for them to work as a team Supervisors encourage people who work for them to exchange opinions and ideas Supervisors frequently hold group meetings where the people who work for them can really discuss things together Our incentive system encourages us to vigorously pursue plant objectives The incentive system at this plant is fair at rewarding people who accomplish plant objectives Our reward system really recognizes the people who contribute the most to our plant Our incentive system at this plant encourages us to reach plant goals Our incentive system is at odds with our plant goalsb Persons (and/or teams) who achieve plant goals are rewarded the same as those who do not achieve plant goalsb Our plant has a low skill level compared with our industryb At this plant.S.92 Incentives to meet objectives JOBSKILL.89 Interaction facilitation INCENTOB.

R. employees are encouraged to learn skills in depth. where 1: I strongly disagree and 5: I strongly agree.88 Feedback on performance α = Cronbach’s alpha.G.92 Communication of strategy FEEDBACK.85 Scales Training in multiple functions Item questions Employees receive training to perform multiple tasks Employees at this plant learn how to perform a variety of tasks/jobs The longer an employee has been at this plant. All scale questions use a five-point Likert response scale. employees only learn how to do one job/taskb At this plant. α = 0. the more tasks or jobs that employee learns to perform Employees are cross trained at this plant so that they can fill in for others if necessary At this plant. . b Indicates a reversed scale question.40 S. a Taylor and Bowers (1972). rather than develop a broad skill baseb In our plant. Schroeder / Journal of Operations Management 21 (2003) 19–43 Appendix A (Continued ) Variable MULTFUN. objectives and strategies are communicated to me Strategies and goals are communicated primarily to managersb I know how we are planning to be competitive at this plant I understand the long-run competitive strategy of this plant Charts showing defect rates are posted on the shop floor Charts showing schedule compliance are posted on the shop floor Charts plotting the frequency of machine breakdowns are posted on the shop floor I am never told whether I am doing a good jobb Information on quality performance is readily available to employees Information on productivity is readily available to employees My manager never comments about the quality of my workb STRATCOM. goals. Ahmad. α = 0. α = 0.

.. Human assets and management dilemmas: coping with hazards on the road to resource-based theory. D.B. 1985. M. Human resource strategy and competitive advantage: a longitudinal study of engineering consultancies.). 514– 539.G. Becker. 1972. 1991. α = 0. Academy of Management Journal 39 (4). Schroeder / Journal of Operations Management 21 (2003) 19–43 41 Appendix B Intangible performance measure Variable COMMITa .. Academy of Management Journal 37 (3). 1995.89 Scales Organizational commitment Item questions I am willing to put in a great deal of effort beyond that normally expected in order to help this organization be successful I talk up this organization to my friends as a great organization to work for I would accept almost any type of job assignment in order to keep working for this organization I find that my values and the organization’s values are very similar I am proud to tell others that I am part of this organization This organization really inspires the best in me in the way of job performance I am extremely glad that I chose this organization to work for over others I was considering at the time I joined I really care about the fate of this organization For me. Dyer. Gerhart. The relationship between quality management practices and performance: synthesis of findings from the world class manufacturing project. 802–835. Journal of Personality and Social Psychology 51 (6).E. Barney.H. Drazin. Schroeder.. A theory of quality management underlying the Deming management method. a Mowday and Steers (1979).. 1996. The moderator-mediator variable distinction in social psychological research: conceptual. 1996. Coff. 1995. A..W.. J.. J. strategic. Harvard Business Review 73 (4).B. D. 1994. R. Academy of Management Journal 39 (4). J. J.. 443–444.H.B. E.A.. Baron. R. M. 49–61.. Administrative Science Quarterly 30. 374–402. Journal of Management 17. J... Kenny. S. References Anderson.M.G. B. contingency.. 1995.. Deci. Schroeder.. Reeves. Collis.. Organizational Behavior and Human Performance 3. D. Boxall. Sakakibara. (Eds. this is the best of all organizations for which to work α = Cronbach’s alpha. Van de Ven. 472– 509. and statistical considerations. Flynn. Human resource strategies and firm performance: what do we know and where do we need to go? The International Journal of Human Resource Management 6 (3). Modes of theorizing in strategic human resource management: tests of universalistic. 1173–1182. Journal of Management Studies 36 (4). and configurational performance predictions. Barney. 1999. 118– 128.C. Advances in the .. Ahmad. In: Fedor. Academy of Management Review 19 (3). Doty.. The impact of human resource management on organizational performance: progress and prospect.. C. Competing on resources: strategy for the 1990s. Steeneveld. Firm resources and sustained competitive advantage. Effects of human resource systems on manufacturing performance and turnover.. Rungtusanatham. 99–120. R. 1986.A. 1997. 217–229.. 779–801. Alternative forms of fit in contingency theory. 1996. D.. 1994.S. T. Looking inside for competitive advantage... Arthur. R.L.G.. R. B. L.J. R. The effects of contingent and non-contingent rewards and controls on intrinsic motivation. Academy of Management Executive 9 (4). Montgomery. 656–670. P. Ghosh. 670–687. Academy of Management Review 22 (2). S. B. Delery.

Weber. The impact of human resource management practices on manufacturing performance. Praise.. An Introduction to Linear Statistical Models. A’s. T. Huselid. Hofstede. Henderson. Management of Organizational Quality. 1990. R. 1994..C... 1993. Tatham. 141–184. Salk. NJ. Harvard Business School Press. Human resource management and organizational performance. Managing I/S design teams: a control theories perspective. J. C.Y. Upper Saddle River. Hair.. Herndon.A. productivity. How common is workplace transformation and who adopts it? Industrial and Labor Relations Review 47 (2). Vol....T. 1995. 2000. 1979. Nunnally. Schuler. R. Vickery. Gaining competitive advantage through human resource management practices. R. International Journal of Operations and Production Management 11 (9). T. A. Technical and strategic human resource management effectiveness as determinants of firm performance. M. New York. Greenwich.L. 55–73.C. Black. K. 757–777. The effects of human resource management systems on economic performance: an international comparison of US and Japanese plants. MacMillan. A. Strategic restructuring and outsourcing: the effect of mergers and acquisitions and LBOs on building firm skills and capabilities.. R. 1995. 1998. J. M. Rhode. Lado. 635–672. 21–39. 1995.. R. Ichniowski. S. Ichniowski.. San Francisco. Boston. Multivariate Data Analysis.. Journal of Operations Management 18. Schuler. 1. Lei. 450–461.H. W. Lawler. Sage.. The Complete Guide to Mergers and Acquisitions. J.. Why incentive plans cannot work. Work force management practices for manufacturing flexibility. Jackson. G. . S.P. Journal of Operations Management 16. Industrial and Labor Relations Review 48 (2).W. Houghton Mifflin. Hill. D. Staughton.. G. MacGraw-Hill. 835– 859.. G. Journal of Vocational Behavior 14. Osterman. British Journal of Industrial Relations 33 (4). 197–221. S..J. 1993a. A. J. 1997. 1995.. Vol. Wiley. Steers.A. Top management turnover in related M&A’s: an additional test of the theory of relative standing. W. Wheelwright. J. MA... Psychometric Theory. The measurement of organizational commitment. Galpin. JAI Press. Harvard Business Review 71. Concepts and Situations. Do US firms invest less in human resources? Training in the world auto industry. A review and classification of empirical research in operations management. C. Kochan.. Shaw. Scudder.. 1995. 24–40. N. S. 1999. Human resource systems and sustained competitive advantage: a competency-based perspective.. networks and individual influence in a multinational management team. 1998. R. Y.. IL. Hitt.. Moody. 699–727. A.E.C.. R. R.M. Journal of Management 21. J. 1978. MacDuffie.. 1981. Snell.. Kinnie. 1994. Beverly Hills. J. J. Incentive Plans. 1989. B.A..P. Punished by Rewards: The Trouble with Gold Stars.C. Kohn. F. National culture.W. 1984. M.C. Homewood. Jackson. 54–60.A..E. Russell. M. Gerhart. D. Kohn... J. Human Resource Management 23.A. 191–202. CA. D. S.C.. Academy of Management Executive 1 (3). Journal of Management 25 (1). 2000.. Shaw. Hayes. Human Resource Planning 21 (1). Droge. 663–691. Academy of Management Journal 40 (1).J. 1976..S. 1st Edition. Human resource bundles and manufacturing performance: organizational logic and flexible production systems in the World Auto Industry. 1984. McGraw-Hill..L. Irwin. Jr. M. 207–219. 1982.. Ryan. The human resource architecture: toward a theory of human capital allocation and development.. I. 704–721. M. CA. Journal of Operations Management 18.G. Schmenner.E. Lepak... 1991.E. Production/Operations Management.. The impact of human resource management practices on turnover. Prennushi. 1997.. Restoring Our Competitive Edge: Competing through Manufacturing.S. 1995. E. Seven practices of successful organizations.. R.T.C. Academy of Management Journal 43 (2).. 948–962. R. M. The effect of human resource management practices on productivity. 1999. Information and Control in Organizations. Organizational differences in managerial compensation and firm performance. Lee.... pp. Brannen..P. Prentice Hall. CA.A. P. Academy of Management Journal 38 (3). 1994.A. New York. Schroeder / Journal of Operations Management 21 (2003) 19–43 Legare.. CT. 1999. Competitive Advantage Through People. 171–188. 1–20. Embedding HRM in a social context. Goodyear Publishing Company. Implementing manufacturing strategy: the human resource management contribution. C. 1999. 1999.W. 31–48.. 1961. 32–41. Milkovich. 173–188. 241–255. J.Y. California Management Review 40 (2). R. American Behavioral Scientist 37 (7).A. 91–101. 96–124.K. Anderson. unpublished paper. Psychological Contracts in Organizations: Understanding Written and Unwritten Agreements. 1992. Jossey-Bass Publishers.. R. Huselid. Human assets and management dilemmas: coping with hazards on the road to resource-based theory.L. MacDuffie. G. S.. 1987.S. Management Science 38.. IL. Jayaram. F. Wilson. MA. 1998. Academy of Management Journal 19 (4). Rousseau..42 S. Pfeffer. Academy of Management Review 22 (2). Hill.. C. 1998. 1980. Mowday.F.. Science Research Associates. Academy of Management Journal 33.M. T. 374–402. Academy of Management Review 24 (1). Schuler... R... Kalleberg. Control and information in the intrapersonal sphere: an extension of cognitive evaluation theory. Thousand Oaks.. Manufacturing Strategy Text and Cases.V.. 1. Morishima.. 1994.G. D. and corporate financial performance. Boston.. Culture’s Consequences. Management Science 45 (5). Industrial Relations 34 (2). Sage. Chicago. Journal of Personality and Social Psychology 43. and Other Bribes. 224– 247... Columbia University. T... J. Linking competitive strategies with human resource management practices. 617–640. Pacific Palisades. Graybill. The human side of mergers and acquisitions.E. Kathuria. Schweiger.D. K. CA. 147–168. M. Pfeffer. New York. 1993b. Lubatkin. Ahmad. Partovi.

C. 1997. O. Markets and Hierarchies: Analysis and Antitrust Implications. C. Schroeder / Journal of Operations Management 21 (2003) 19–43 Snell. 1999. Journal of Operations Management 15. Institute for Social Research. Michigan. Dean. Total Quality Management 10 (8). Dean Jr. Markland. Ahmad. Integrated manufacturing and human resource management: a human capital perspective.. Human resource management: manufacturing strategy and firm performance.. 1996. 467–504. The role of trust in labor-management relations.W. J. S. Motwani. 836–866. J. M..P. . Taylor. Reflecting corporate strategy in manufacturing decisions. Soderquist. Survey of Organizations: A Machine Scored Standardized Questionnaire Instrument. R. Business Horizons 21 (1). R... 1975. Organization Development Journal. Womach.. R.K. Summer. J.. D. D.A.W. Roos.A..G.A. The Machine That Changed the World.. Academy of Management Journal 39 (4). 43 Vickery.G. Droge... Taylor. 85–89. K. Free Press. D. New York. S. Lepak. 317–330..S.E.E. Williamson. 1972. New York. Dimensions of manufacturing strength in the furniture industry... Academy of Management Journal 35 (3).G. 1990. 1978. 1992.. 1989. Wheelwright. D..P. Youndt. S. J.. Snell.. S. Quality issues in lean production implementation: a case study of a French automotive supplier. 57–66..C. Bowers. 1107–1122. Jones. University of Michigan. Rawson-Macmillan.. J.