A contingent view of quality management--the impact of international competition on quality

INTRODUCTION

The past decade has witnessed the most rapid dissemination of quality management philosophy and practices in American history. Arthur D. Little's study in 1992 indicated that quality management in various forms is being practiced by about 93% of the largest 500 U.S. corporations (Powell, 1995). Similarly, the Boston University manufacturing survey (Kim & Miller, 1992) highlighted the importance of quality as one of the top five competitive priorities of U.S. manufacturers. Nonetheless, quality focus and quality-related investments remain controversial strategic issues in terms of performance deliverables. Managers have expressed reservations on the value of quality management as an implementable and costeffective initiative (Naj, 1993; Schaffer & Thomson, 1992). Benchmarking lowquality-focused companies against strong quality management adopters has not revealed clear distinctions in superiority of outcomes (Fuchsberg, 1993; Mathews, 1992). Although the known complexities of the issue prevent an easy resolution of the quality-performance debate, scholars have been seeking to achieve a clearer understanding of quality practices and performance outcomes (Garvin, 1986; Flynn, Schroeder, & Sakakibara, 1995). These investigations can be classified into two broad investigative streams: studies concerned with the identification and measurement of the elements of quality management, and studies that examine the relationships between quality practices and performance. The findings of this research have been inconsistent, especially in regard to the relationship between quality practices and performance. Quality program disappointments have been attributed to inadequate resources, poor implementation, negligence in making complementary investments in organization structure and human resources, misalignment with marketing and corporate goals, and an inadequate appreciation of system dynamics (Lawler, Mohrman, & Ledford, 1992; Kim, 1993; Powell, 1995; Sterman, Repenning, & Kofman, 1997). Much of this research reflects a resourcebased view of the firm, dealing primarily with internal issues of managerial and technological competence in developing and executing effective quality management strategies.

In contrast, the neoclassical/industrial organization (1/0) perspective on the influence of industry environment on firm practices and performance is conspicuously absent in the empirical quality management literature, despite research calls for such investigations (Sitkin, Sutcliffe, & Schroeder, 1994). Previous 1/0 research on the quality-industry environment relationship is grounded in the economics literature and centers on price and quality competition among symmetric firms in differentiated demand environments (Hotelling, 1929; Moorthy, 1988; Polo, 1991). In a recent study, Banker, Khosla, and Sinha (1998) developed gametheoretic models to examine the impact of competition on quality goals in duopoly, cooperative, and oligopoly situations. In general, quality was found to increase with competitive intensity, provided the cost of quality improvements declined with increased competition in an industry. These virtual models, however, have not been validated by empirical testing. Specifically, the issue of how quality is influenced by the competition has not received attention in the empirical literature. The importance of addressing this gap in the quality literature is evident from the increasing concern over underperforming quality programs and questions about the relevance of quality strategies in highly competitive environments (Taylor, 1992a; Fuchsberg, 1993). This study introduces an important exogenous uncertainty for many companies, international competition, as a contingent influence on quality management work structure, practice, and performance. The research objectives of this study are to: * Develop and validate a quality management framework integrating high involvement work practices, quality practices, customer satisfaction, and firm performance;

* Examine the extent to which international competition affects investments in high involvement work practices and quality practices;

* Examine the extent to which international competition affects customer satisfaction performance;

* Examine the extent to which international competition moderates the interrelationships among high involvement work practices, quality practices, customer satisfaction performance, and firm performance.

A conceptual framework of quality management is developed and tested. The contingent effects of international competition on the constructs and relationships of the conceptual framework are examined. The paper concludes with a discussion of the theoretical and strategic implications of the findings.

CONCEPtual FRAMEWORK Even as calls are made for developing holistic frameworks for quality management research (Ahire, Landeros, & Golhar, 1995), researchers have been developing conceptual frameworks around quality management dimensions. Ross (1993) described quality management as an integrated philosophy, requiring managerial proactiveness in customer orientation, rework reduction, employee involvement, and supplier relationships. Anderson, Rungtusanatham, and Schroeder (1994) perceived quality management as a holistic approach to organization-wide quality, operationalized through leadership, internal/external cooperation, effective process management, product design, learning, customer focus and involvement, employee fulfillment, and continuous improvement. The quality literature has been consistent in describing quality management as an integrated management philosophy, with functional and organizational boundary-spanning attributes. Practitioners have also emphasized the importance of integration of quality actions among different areas in a firm (Pegels, 1995; Talley, 1991; Sashkin & Kiser, 1993). Integration has been identified as a key element of organizational quality. This study integrates relationships among environmental factors, quality strategy, quality practices, work practices, and performance to develop a conceptual framework of quality management (Figure 1).

Given time, technology and market forces generate enough momentum to pressure companies to develop formal strategies for acquiring quality competencies. Such strategic initiatives push changes in organizational work practices (Chandler, 1992). These structural changes, in turn, enable and promote the adoption of appropriate quality practices. Structure and practice conjunctively accomplish successful customer satisfaction performance outcomes. High levels of customer satisfaction performance lead to improved marketing and financial performance. The moderating role of international competition represents an important caveat to these propositions. The current analysis focuses on the "high involvement work practices-quality practices-customer satisfaction performance-firm performance" segment of the conceptual framework.

Three notable features of the conceptual framework distinguish it from its predecessors. The first and perhaps most distinct difference between our framework and others derives from the explicit modeling of international competition as a moderator of work practices, quality practices, and performance relationships. Second, the role of organizational work practices has been maintained as an independent entity, unlike other frameworks, which usually subsume organizational factors into a quality management scale (Flynn, Schroeder, & Sakakibara, 1994; Powell, 1995). There is compelling evidence in the strategy literature to warrant considering organizational work structure as a separate and influential factor in strategy implementation and firm performance (Armour & Teece, 1978; Teece, 1981; Hrebiniak& Joyce, 1984; Dean & Susman, 1989). A third difference lies in the scope of empirical test and validation. Previous empirical research has examined individual quality linkages in isolation, such as the relationship between quality practices and quality performance, or that between quality performance and business performance (Adams, 1994). There is a paucity of empirical studies of sufficiently holistic scope. Grandzol and Gershon (1997) used structural equation modeling to explore selective relationships among quality practices and aspects of quality performance in an organization. Choi and Eboch (1998) used path model analysis to examine the relationship between TQM practices and customer satisfaction. Examinations of larger, more complex models have been constrained by the use of relatively weak analytical techniques, such as correlation analysis (Flynn et al., 1994; Powell, 1995). This study conducts a simultaneous empirical examination of the conceptual quality management framework, using structural equation modeling techniques.

Framework Constructs Quality practices Quality practices are operational interventions introduced by management to achieve quality performance goals. Quality practices encompass a gamut of managerial actions, including workforce training, top management support, customer focus, supplier involvement, an open organizational structure, and the creation of suitable measurement and evaluation systems (Powell, 1995; Flynn et al., 1994, 1995; Evans & Lindsay, 1993).

This study views quality practices as comprising two parallel sets of activities. The first set involves decisions and actions internal to the firm, involving design, quality planning and leadership, quality training, quality procedures, quality resource investment and evaluation, benchmarking, and customer orientation. Quality-oriented environments pay attention to product and process design, engage in quality planning with active management initiatives and communication, and invest in statistical process control training programs. Quality agendas include systematic procedures for employee evaluation on quality parameters, and benchmarking and monitoring of customer and competitor behavior. Customer focus is another key emphasis in quality programs, requiring monitoring customer complaints, determining customer expectations, and regular customer satisfaction assessments. The second set of quality practices engages the supply base and is characterized as part of process management in the Baldrige award criteria (Samson & Terziovski, 1999). Strategic sourcing practices include effective information sharing, supplier quality assessment, production schedule synchronization, and technology sourcing (Monczka, Trent, & Handheld, 1997). Research suggests that strategic sourcing of supplier capabilities can be a critical factor in enhancing quality performance (Armistead & Mapes, 1993; Bhote, 1987). This study compiles quality practices under four domains: supply chain management, quality resources and evaluation, quality training, and customer commitment. People management practices were included separately in the high involvement work practices construct.

High involvement work practices High involvement work practices involve an organization's basic design for employee involvement, as well as the management processes required to integrate and control its operations (Hrebiniak, Joyce, & Snow, 1989). Of interest are the structural aspects of work organization (teams, formal interfunctional interfaces) and associated practices (interdepartmental cooperation). Excluded are human resource management practices such as hiring or training that are not so intricately integrated and internalized with the administrative core of the organization (Pil & MacDuffie, 1996). Structural adaptations have been shown to be essential to the success of many organizational programs. For example, high involvement organizational work practices have been identified as a critical element of successful technology implementation. Organizational studies have discussed innovative forms such as group technology cells, cross-functional teams, plants within plants, and networks that go with the grain of new technologies (Sethi & Sethi, 1990). Similarly, as a firm develops a quality strategy, its anatomy should change to accommodate needs for inter- and intrafunctional collaboration. The importance of organizational processes in facilitating quality learning and implementation is documented. Basic design innovations such as cross-functional teaming, together with supporting integration and coordination mechanisms are considered to be primary attributes of a quality-based organization (Powell, 1995; Flynn et al., 1994; Haim, 1993). Powell (1995) observed that a critical success factor in quality programs was an open organization, consisting of interfunctional teams and open horizontal communication. Organizational capabilities in team building and management were found to confer tacit, inimitable benefits for sustainable competitive advantage. Flynn et al. (1994) reported similar findings. Their study of 42 U.S. plants found that cross-functional teaming and other formalized interfacing mechanisms for joint problem analysis and problem solving were integral to the success of quality programs.

Customer satisfaction performance Although quality performance has been conventionally defined in terms of operational performance (cost of quality, re-work costs, WIP inventory levels), recent empirical studies (Adams, 1994; Madu, Kuei, & Jacob, 1996) have introduced strategic customer-based performance considerations. The importance of customer satisfaction in the evaluation of quality performance finds mention in the seminal quality literature (Juran, 1981 a, 198 lb). Similar perspectives positioning customer satisfaction as the final outcome of quality management have been offered in current research (Anderson, Fornell, & Lehmann, 1994; Dean & Bowen, 1994; Reeves & Bednar, 1994). Madu et al. (1996) indicated that responsiveness, availability, and completeness of product/service quality are important considerations in the evaluation of customer satisfaction. Customer satisfaction and retention constitute part of the customer-driven quality focus embedded within the Baldrige Award core values (Evans & Ford, 1997). Roth and Giffi (1995) found that world-class manufacturers value performance improvements in perceived responsiveness, on-time delivery, and customer satisfaction ratings.

Lederer and Rhee (1995) interpreted customer satisfaction as a quality performance measure. This study defines customer satisfaction performance as the achievement of the strategic quality objectives of customer satisfaction, customer retention, and delivery performance. These quality performance aspects involve meeting and exceeding customer expectations (Deming, 1986), and represent a broad-based definition that spans multiple industry criteria.

Firm performance Firm performance is defined as the achievement of financial and market share objectives. Anecdotal evidence has related quality investments to performance gains in sales, ROA, and market share (Sankar, 1995). Return on assets (ROA), profitability, and average sales growth are some dimensions of operational and financial performance employed in the quality literature. Previous studies in quality management have framed firm performance as a combination of operational and financial performance (Sluti, 1992; Adams, 1994). Easton and Jarell (1995) defined firm performance in terms of profit margin, ROA, asset use efficiency, and excess stock returns. Adams (1994) found a significant, albeit weak relationship between these elements of firm performance and a firm's quality improvement approach. A point of interest is the element of time lag between quality investments by a firm and the realization of tangible financial or operational benefits. Such delays are not uncommon with other types of improvement programs, such as advanced manufacturing technology investments. Research indicates that companies implementing quality improvement programs had to wait five to eight years on average before experiencing financial gains (Dusseau, 1996). It becomes important to select a sample frame that includes companies likely to possess adequate quality maturity, in order to be able to discern long-term effects.

HYPOTHESIS DEVELOPMENT

This section develops specific hypotheses for empirical testing, using the conceptual framework of quality management introduced in the earlier section.

High Involvement Work Practices and Quality Practices Deming noted that process management requires cooperation, not competition (cited by Anderson, Rungtusanatham, & Schroeder, 1994). High involvement work practices are an integral feature of the organismic organizational form (Burns & Stalker, 1961). In a comparison of the compatibility of different organizational models with quality practices, Spencer (1994) observed a strong congruence between structural organieity and quality management requirements. She found the organismic model focus on coordination, lateral communication, and learning, synergistic with quality management needs. Similarly, Anderson, Rungtusanatham, and Schroeder (1994) theorized that an organization that fosters cooperation and learning, facilitates the implementation of quality practices. Practitioner literature has also emphasized the importance of interdepartmental communication, crossfunctional coordination, flat organizations, and teamwork support, for effective implementation of quality plans (Murphy, 1996; Tatikonda & Tatikonda, 1996)organizational attributes reflected in high involvement work practices. Although a case conceivably can be made for quality management's affinity with the mechanistic objectives of process efficiency and control (Spencer, 1994), the high involvement form subsumes many mechanistic characteristics such as conformance to standards (mechanistic objective) as a means to achieving customer satisfaction (organismic objective). The classical Burns and Stalker (1961) distinction between different types of organization may in fact represent a continuum of organizational structural development, with different combinations of organismic and mechanistic elements. Organizations move along the continuum in their effort to cope with environmental shifts. The primary contingency of change remains. Change requires a cooperative organismic work environment and forms an essential feature of quality management. Thus, H 1: High involvement work practices are positively related to quality practices.

Quality Practices and Customer Satisfaction Performance Citing Deming (1982), Anderson, Fornell, and Lehmann (1994) emphasized customer satisfaction as the ultimate outcome of quality management process activities. An earlier study reported a positive relationship between quality and customer satisfaction (Anderson et al., 1993). The Baldrige Award recognizes this linkage between quality management practice and quality performance by assigning a 30% weight to customer satisfaction in its evaluation process. Literature ascribes quality performance failures to a lack of customer focus in the firm (Porter, 1996). Clearly, the adoption and effective execution of the appropriate quality practices should result in enhanced customer satisfaction performance. A moot point is the selection of the "right" practices. As the Boston Consulting Group found in their survey of U.S. businesses, 95-99% of internal practices bear little association with consumer satisfaction (Godfrey, 1993). 1993) ------

While contingent factors may influence the extent of implementation of an individual quality practice, there is a general consensus among researchers on the operational constituents of quality practices (Ahire, Golhar, & Waller, 1996; Powell, 1995; Flynn et al., 1994; Anderson, Rungtusanatham, & Schroeder, 1994). General agreement also exists on the performance implications of ignoring or inadequately implementing an operational practice (Tatikonda & Tatikonda, 1996; Masters, 1996; Harari, 1993). Previous empirical studies have documented the influence of individual quality practices such as teamwork, top management commitment, leadership, employee training and empowerment, customer orientation, appropriate measures, and buyer-supplier relationship development, on customer satisfaction performance (Samson & Terziovski, 1999; Flynn et al., 1994; Powell, 1995; Adams, 1994). Such relationships have been explored through correlation or regression analysis. A recent study by Choi and Eboch (1998) found a positive linkage between composite quality practices and customer satisfaction, however, their operationalization of quality practices omitted supply-base practices. The above discussion suggests the following hypothesis:

H2: Quality practices are positively related to customer satisfaction performance.

Customer Satisfaction Performance and Firm Performance The quality literature emerges with mixed results in its examination of the impact of customer satisfaction performance on firm performance. The PIMS studies suggested a positive relationship between customer satisfaction performance and a firm's market share (Buzzell & Weirsema, 1981; Craig & Douglas, 1982). More recent studies have provided more evidence of the positive impact of customer satisfaction performance on organizational cost, profit, and sales growth (Dean & Snell, 1996; Madu et al., 1996). Other researchers have reported conflicting findings. These studies found the relationship between customer satisfaction performance and firm performance indeterminate and tenuous at best, and negative at other times (Fisher, 1992; Adams, 1994). On balance, there is enough evidence to suggest that customer satisfaction performance generally relates positively to firm performance. Satisfied customers develop product loyalty, leading to repeat purchases, sales growth, and increased profits. Therefore,

H3: Customer satisfaction performance is positively related to firm performance.

A Contingency Theory of Quality Management Contingency theory prescribes that an organization must be aligned with its environment to achieve optimal performance (Lawrence & Lorsch, 1967; Thompson, 1967). The concept of "fit" underlies the contingency perspective. Venkatraman (1989) defined fit as a theory-based match between two related variables. The organization interacts with externalities to configure and match its internal resources to exogenous requirements. The congruence of the match determines successful outcomes. To illustrate, Chrysler's success in introducing new products is, in part, attributed to the achieved fit between high involvement work practices (cross-functional platform teams) and new product development strategy (Dyer, 1996). A contingency perspective of quality management implies that quality practices and work structures must be consistent with situational factors.

A key situational factor in firm environments is competition, or more fundamentally, competitive scope and intensity. The intensity of competition present in a market environment is likely to be related to its scope and breadth. Companies facing competition from abroad would likely inhabit more challenging environments than companies that encounter solely domestic competition. Common characteristics of international competition are short product life cycles, product design sophistication, consistent high quality, cost reductions, and customization. Rapidfire introductions of new products and continuously rising customer expectations create a volatile and demanding market environment. Customization, product-line proliferation, and intercultural differences in managerial approach and operation, add complexity to the business environment (Ohmae, 1985).

Previous empirical research has not employed a contingency perspective in quality management. Quality program failures have been attributed to a variety of reasons, including the rise of quality bureaucracies, incorrect understanding of quality management content, expectations of instant gratification, and top management orientations toward customers and time (Choi & Behling, 1997). The root causes of quality failures converge on a resource-based perspective of the firm. In contrast, scant attention has been paid to the role of firm externalities. Sitkin et al. (1994) conceptualized a contingency model of quality management effectiveness that distinguishes between the control and learning aspects of quality management. Structure and control are considered appropriate for an environment of low uncertainty, while a learning posture is seen as more effective in high uncertainty situations. Reed, Lemak, and Montgomery (1996) argued that quality management efforts should be targeted toward operational performance goals, selected with due consideration to the degree of uncertainty present in the environment. High uncertainty conditions would require quality management programs directed toward the creation of new product designs and markets. Quality investments in process efficiency and product reliability benefit firm performance under conditions of low uncertainty. Again, the conceptualization does not extend to empirical verification.

This study introduces international competition as an explicit contingent factor in the practice and performance of quality management. Firms facing international competition have to produce and market high-quality products to meet stringent customer and competitive quality standards. Such firms would be required to make continual investments in work practices and quality practices in order to remain competitive. Quality-related investments associated with increased international competition should result in improved products and services, leading to higher levels of customer satisfaction performance. The triad industrial sector comprising Western Europe, Japan, and North America is home to some of the most sophisticated, aggressive, and demanding customers in the world (Ohmae, 1985). Thus, H4: Investments in high involvement work practices and quality practices by North American firms are positively influenced by international competition.

H5: The level of customer satisfaction performance in North American firms is positively related to the level of international competition.

Quality practices alone may not guarantee customer satisfaction in a dynamic, demanding market. Literature suggests that quality performs as a market-entry qualifier in sophisticated markets. Models developed by Lederer and Rhee (1995) show that firms earn zero economic returns on quality investments in highly dynamic competitive environments. Yet, those firms that do not invest in quality are forced from these markets. Conversely, firms can earn economic rent from early implementation of quality programs in less competitive environments. Sterman et al. (1997) studied a plant that implemented quality management with significant success on customer-relevant parameters such as cost and defect reduction, yet failed to translate these gains into firm-level performance. They ascribed the lack of firm-level impact of quality performance to an inadequate understanding of system dynamics, and a misalignment of manufacturing capabilities with actual customer and market requirements. They also demonstrated that nonadoption of quality practices would have likely resulted in far more serious consequences to the firm, chiefly from competition dynamics.

The linkage between quality practices and customer satisfaction performance in highly competitive regimes may be weaker than that found in less competitive regimes. Customer satisfaction performance in intensely competitive, highly discerning markets may be more influenced by product innovation, cost issues, or customization capabilities, than by quality-related factors. Thus, international competition may moderate the influence of quality practices on customer satisfaction performance. H6: The positive influence of quality practices on customer satisfaction performance declines with an increase in international competition.

We did not have good reasons to anticipate moderator effects of international competition on the relationship between high involvement work practices and quality practices, or the relationship between customer satisfaction performance and firm performance. The next section of the paper describes the empirical examination of the above hypotheses.

METHODS Measures Based on the literature, operational measures were developed for the constructs in the conceptual framework (see Figure 1). Following the procedures described in Bollen (1989), each variable was operationalized by: (1) providing a theoretical definition, (2) identifying the dimensions, (3) forming appropriate measures, and (4) specifying the relation between the measures and the construct. The measures were pre-tested through a set of interviews with managers in 14 North American and European manufacturing organizations in order to improve the clarity and content of each measure. (These managers were not included in the sample used in this study.)

High Involvement Work Practices Consistent with the advocacy of an organismic structure in the quality literature, measures of high involvement work practices were based on the attributes of organismic organizations (Burns & Stalker, 1961). Organismic work structures foster employee involvement in decision making and problem solving. High involvement work practices were evaluated in terms of extent of interdepartmental coordination, extent of use of cross-functional teams, extent of formalization of interfunctional interfaces, delegation of quality-related decisions to employees, and employee responsiveness in making suggestions for quality improvements. Similar measures are observed in existing scales of high involvement work practices (Pil & MacDuffie, 1996).

Quality Practices Powell (1995), after reviewing the seminal quality literature and the Baldrige award criteria, identified 12 encompassing constructs of quality management. These constructs include executive commitment, adopting the philosophy, closeness to customers, closeness to suppliers, benchmarking, training, open organization, employee empowerment, zero defect mentality, flexible manufacturing, process improvement, and measurement. Harris (1995) presented a similar thematic focus on process improvement, customer concern, teamwork, top management commitment, and training. Several researchers have operationalized these quality management constructs through quality management scales (Saraph, Benson, & Schroeder, 1989; Flynn et al.,1994; Ahire et al., 1996). Saraph et al. identified eight critical factors of quality management: top management leadership, quality data and reporting, process management, product/service design, training, supplier quality management, role of the quality department, and employee relations. The customer element is not explicated in their analysis. Flynn et al. (1994) extended the SBS framework to include customer orientation, but omitted training and benchmarking items in their instrument. Ahire et al. (1996) addressed these issues in their scale of quality management constructs; however, their inclusion of outcome measures such as product quality in the same scale blurs the demarcation between quality management practice and performance. A similar problem exists in Black and Porter's (1996) factor analysis of quality management constructs, wherein performance (customer satisfaction) and practice are combined into a joint measure.

This study groups quality practices into four major dimensions: supply chain management, quality resources and evaluation, quality training, and customer commitment. The measures for these domains include items related to quality planning, leadership, process management, design, and benchmarking, elements of the Malcolm Baldrige award framework (Samson & Terziovski, 1999). Human resource management practices were included separately in the high involvement work practices construct domain. This classification is consistent with the literature (Harris, 1995; Powell, 1995). Each of these dimensions was measured through multiple reflective indicators and subsequently coalesced into a single secondorder quality practices factor.

Our operationalization of the quality practices latent variable, while including most major quality practices, was not exhaustive in scope. We make no claim that these domains collectively represent a full census of quality practices, if indeed such an enumeration is possible. Omitted, for instance, are environmental factors that may now be a part of some quality management programs.

Customer Satisfaction Performance This study defines customer satisfaction performance as the achievement of the strategic quality objectives of customer satisfaction, customer retention, and delivery performance. Instead of using conventional quality conformance measures such as defect rates and rework costs, our study assesses quality performance through customer satisfaction performance. This approach has two advantages: first, it reflects the growing focus on customer satisfaction as a principle criterion in evaluating quality program effectiveness (Zeithaml, Parasuraman, & Berry, 1990; Lederer & Rhee, 1995); second, it elevates quality performance from the operational to the strategic level, enabling a more coherent link between quality performance and firm performance. Customer satisfaction performance is measured through a combination of respondent perceptions of customer satisfaction, and internal measures of customer retention and on-time delivery. Studies have shown that it is easier and more profitable to retain customers than to create new ones (Reichheld, 1996; Achrol, 1997).

Firm Performance Firm performance is defined as the achievement of financial and market share objectives and is measured through market growth and financial performance. Respondent perceptions of such measures have been widely accepted in organizational research (Lawrence & Lorsch, 1967; Powell, 1992). The convergent validity of these subjective measures was verified by comparison with objective performance data (Compustat), for a subsample of the firms in this study.

Measures for the constructs were compiled into a survey instrument. In determining the sample for the study, an important criterion to consider was the unit of analysis. Benson, Saraph, and Schroeder (1991) found that company type, size, and manager type had no effect on the variation in their "ideal quality variables," which justifies to a certain extent the use of a cross-industry sample with both small and large firms. Furthermore, these authors provided evidence to indicate that quality performance can vary significantly within the same organization across business units. For this reason, it was deemed appropriate that the unit of analysis be the division level. The instrument was finalized after extensive iterations involving pre-tests in industry and academia. Face and content validity issues were identified and addressed at this stage. All of the questions in the survey were further pre-tested at meetings with a top-level quality manager at 14 firms. During these interviews, major segments of the questionnaire were revised for clarity and understanding. The 14 responses were not used in the ensuing analysis.

International Competition International competition was modeled as a function of competition from Japanese companies. The primary reason for adopting Japan to represent international competition was the documented dominance of Japanese industry in many sectors of U.S. enterprise in recent times. Japanese competitors constitute the primary competition in a majority of the industries represented in the study data, including automobiles, electronics, and consumer goods. To illustrate, the North American trade volume with the EC countries in 1970 was almost twice as large as its trade with five major Asian nations (including Japan). During the subsequent years (1970 to 1995), North America's trade with the EC increased threefold, while its trade with Japan surged thirteenfold, from $13 billion to $186 billion (Kotabe, 1998). The flow of products into the U.S. has been much greater than the outflow, as evidenced by the ongoing debates about the almost $60 billion U.S. trade deficit with Japan (Naughton, 1999). Competition from Japanese companies has been intensifying. Japanese manufacturers accounted for nearly one out of every three cars sold in the U.S. in 1992, up from one in five in 1982 (Taylor, 1992b). Steel imports from Japan in the first nine months of 1998 were up by 148% over the same period in the past year (Costa, 1999). No other country in the world has posed such a formidable and sustained challenge to U.S. business across a broad range of products. A related factor in choosing Japanese competition to represent international competition was the overwhelming interest in Japanese industrial philosophies and operational methods shown by U.S. industry and academia over the past two decades. Focusing on Japan, we hoped, would make our investigation relevant and important to a wider cross-section of business and academia than if we had selected another lesser competitor such as Germany or Taiwan. For similar reasons, we did not aggregate competition from multiple countries into a composite variable. Doing so may have masked the impact of leading-edge competitors such as Japan and confounded its effects with that of less demanding competition. Sampling

Cross-sectional industry data was collected through a mail survey from senior quality executives. A list of 3,000 quality directors and vice presidents was obtained from the American Society of Quality. From this list, a subsample of 1,469 manufacturing firms in the automotive, chemical, computer, construction, consumer products, defense, electronics, industrial products, medical device, packaging, pharmaceutical, paperboard, semiconductor, and telecommunications industries was identified. The firms were located in all 50 states of the United States. Two mailings with one follow-up reminder returned 307 completed surveys (a 21% response rate). Further screening reduced the effective sample size to a total of 290 companies. On average, respondent firms had 7.02 years of formal quality planning history. A distribution of the sample by industry is shown in Table 1.

RESULTS Structural equation modeling was employed to analyze the data. It improves on other multivariate analysis techniques such as multiple regression, in several ways. SEM explicitly incorporates errors of measurement in its analysis. Another unique feature of SEM is its ability to evaluate models holistically, conducting formal tests of both global and individual parameter fit. SEM also allows the simultaneous modeling and analysis of relationships between latent variables.

The univariate distributions of the variables were examined for excessive skewness and/or kurtosis using D'Agostino's test statistic, which showed that the skewness and kurtosis of the variables were not statistically different from that of normal distributions, a prerequisite for using structural models. All of the variables were examined for outliers and other departures from non-normality. No significant outliers were detected. The obtained sample size (n = 290) appeared adequate to test a full simultaneous measurement-and-path structural mode, based on previous SEM research (Hair, Anderson, Tatham, & Black, 1995; Rigdon, 1998). Thus, there was a choice between employing a full structural equation model and conducting concurrent tests for measurement and model fit, or conducting a two-stage measurement and path analysis.

Because the quality practices index represented the most complex (second order) construct in the model, it was decided to add rigor to the measurement of this concept through a separate confirmatory factor analysis. This approach also reduced the possibility of interpretational confounding, encountered at times in complex models (Burt, 1971). Measurement models are traditionally evaluated through either exploratory or confirmatory factor analysis. Literature has identified several inherent limitations of exploratory factor analysis (Ahire et al., 1996). Following precedent (Anderson & Gerbing, 1982; Fornell & Larcker, 1981), confirmatory factor analysis was employed to build and verify the validity, unidimensionality, and reliability of the measurement model of the quality practices index construct (Figure 2). Table 2 lists the initial and final set of item measures for the quality practices construct.

The overall fit indices for the CFA indicated good model fit (CFI: .977; NFI: .937; NNFI: .970). Common method variances were found in five cases, each case involving associations between items measuring the same factor. Unobserved firm effects could account for the relatedness of these items. Following Bagozzi and Yi (1988), we accounted for these variances through correlation of errors, especially since the correlations appeared logical and were not excessive in number. Convergent Validity The confirmatory factor analysis realized substantive convergent validity with individual item loadings in the range of .60 to .81 for all items, except for two items (0.52 and .59) loading on supply chain management practices, one item loading on quality resources and evaluation (0.57), and one item loading on quality training (0.55). These items were retained in view of their centrality to their respective first-order constructs, and considering the strong overall loadings. All loadings were significant (p < .01).

Reliability (composite reliability) for the first-order factors ranged between .60 (supply chain management practices) and .87 (quality resources and evaluation). Composite reliability estimates were preferred over Chronbach's alpha, since the former is a closer approximation of reliability. Chronbach's alpha, unlike composite reliability, assumes tau equivalency among measures, and tends to be a lower bound of reliability (Werts, Linn, & Joreskog, 1974). The average variance explained in each factor was above .50, except for supply chain management practices (0.33). However, careful content validation of the included items, the developmental nature of the scale, and the overall fit of the CFA model provided justification for accepting the measures for the supply chain management practices factor.

Discriminant Validity All the interfactor correlations were positive and significantly different from zero. Applying Fornell and Larcker's (1981) stringent rules for establishing discriminant validity, the average variance extracted for each factor was compared and found to be greater than the squared correlations between that factor and all other factors (except for supply chain management practices). A single exception was a squared interfactor correlation of .43 found for the supply chain management practices factor, which exceeded the variance extracted for that factor (.33). Nonetheless, the difference was small enough to warrant further tests on the factor for other evidence of discriminant validity. Confidence intervals (.05 significance level) were drawn around each interfactor correlation to verify that the correlations were significantly different from 1.00. None of the confidence intervals contained 1.00 in their range. Viewed collectively, the results of the analysis indicate substantive discriminant validity among the constructs of the quality practices index.

The CFA analysis provided support for the multidimensionality of the quality practices construct. Scores on the four dimensions were combined and averaged to form a composite quality practices construct.

Measurement models for the remaining constructs of the conceptual framework were developed simultaneously with latent variable path analysis, using a full structural equation model. Table 3 describes the measurement properties for all the constructs used in data analysis. The next section describes the structural equation modeling analysis of the conceptual framework.

Structural Model Figure 3 shows the structural model used for data analysis. Table 4 presents the summary results of the structural model analysis and the correlation matrix used for the analysis. Quality practices was considered an observed variable for the purposes of this analysis.

The overall fit indices evidenced good support for the model (CFI=.959; NNF1=.942; NFI= .917; Hoyle & Panter, 1995; Hu & Bentler, 1995). The statistically significant chi-square (p = .002) was not unexpected, given the large sample size. The sensitivity of the chi-square statistic to sample size is well known (Bentler & Bonnet, 1980; Bentler, 1990), and trivial specification errors have been found to lead to model rejection with large samples. The largest standardized residual (.18) was much below .30. Collectively, the overall fit indices substantiate the internal validity of the structural model.

The strong (>.49) and significant (p < .05) individual item-construct loadings indicated convergent validity. Construct reliabilities (composite reliability) ranged between .68 - .89, well exceeding the minimum limit (.60) suggested for new scales (Nunnally, 1978). The average variances extracted for each construct were reasonably close (.42 - .51) to Bagozzi and Yi's (1988) suggested minimum of .50.

Two approaches were used to assess the discriminant validity of the model constructs. First, a pairwise comparison of the latent factors indicated that all the interfactor correlations were significantly different from 1.00, providing evidence of discriminant validity (Challagalla & Shervani, 1996). The average variance extracted for each factor was also found greater than the squared correlations between that factor and the other latent factors in the model (Fornell & Larcker, 1981). The results of the structural model analysis supported the following hypotheses:

Hl: A significant path from High Involvement Work Practices to Quality Practices-.876 (p < .05), supporting the hypothesized relationship between an organismic organizational structure and the effective deployment of quality practices (variance explained = .77).

H2: A significant path from Quality Practices to Customer Satisfaction Performance-.360 (p <.05), supporting the hypothesized relationship between the use of quality practices and the attainment of quality performance goals (variance explained = . 13).

H3 A significant path from Customer Satisfaction Performance to Firm Performance-.152 (t-value = 1.939), supporting the hypothesized relationship between quality performance and overall firm performance.

A significant posthoc path from High Involvement Work Practices to Firm Performance-.460 (p <.05), was indicative of a positive effect of an organismic organizational structure on overall firm performance. Although this path was not modeled in the conceptual framework, it is not entirely unexpected. As Snell and Dean (1992) reported, aspects of the organismic structural form are compatible with other performance improvement initiatives such as JIT and the implementation of advanced manufacturing technologies. Customer Satisfaction Performance and High Involvement Work Practices together explained 28% of the total variance in Firm Performance. Contingency Effects The sample was segregated into three groups: firms facing strong competition from Japan, firms facing moderate competition from Japan, and firms facing weak competition from Japan. Over 95% of all respondents reported high levels of domestic competition. Cases with low domestic competition were eliminated from the groups, to control for domestic competition. The cut-off points for the low, medium, and high competition groups were 1-2, >2 to <5, and 5-7, respectively, on a 1 to 7 Likert scale.

A latent means structural analysis using EQS was employed to evaluate the effect of international competition on high involvement work practices, quality practices, and performance means (H4 and 115). Significant differences were found among the group means for high involvement work practices and quality management practices. There was no significant difference in customer satisfaction performance between the groups. Table 5 shows the results of the analysis.

The results provide partial support for H4 and H5. Firms facing strong competition from Japan were found to focus more on high involvement work practices and quality practices, relative to firms facing low Japanese competition. Firms in the medium competition group also were seen to place greater emphasis on quality practices relative to firms in the low competition group. There were no significant differences between the mean scores of the medium competition group and those of the high competition group for any of the practices. Interestingly, no significant difference was found in customer satisfaction performance between high competition, medium, or low competition environments.

Next examined were the relationships between the quality framework constructs in conditions of high, medium, and low international competition. Two basic techniques are used in SEM to examine interaction effects: a multigroup approach in which one or both of the interaction variables is discrete, or an indicant product analysis procedure with continuous interaction variables. In the multigroup approach, groups are formed for different levels of one or both of the interacting variables, and an interaction is indicated when parameter values change significantly across groups (Bagozzi & Yi, 1989). In contrast, the indicant product approach involves modeling an interaction term that is usually a multiplicative function of the interaction variables (Kenny & Judd, 1984). In a comparative evaluation of the two techniques, Rigdon, Schumacker, and Wothke (1998) concluded that while the product indicant approach may offer more parsimony, the multigroup approach provides more flexibility and avoids multicollinearity and distributional problems that can occur with the indicant product approach. Another advantage with the multigroup approach is ease of implementation on any SEM software. An interaction effect SEM using the multigroup approach was employed to compare the relationships across the three competition categories. A fundamental assumption of the interaction effect SEM multisample approach is that the interaction variable (international competition) does not covary with the variables in the model. A MANOVA was used, only to test this assumption; the MANOVA showed that the vector of means for each group did not vary systematically with changes in the level of international competition. Table 6 shows the results of the multigroup SEM.

The path model analysis showed significant differences in relationship patterns across different levels of international competition. Specifically, the relationship between quality practices and customer satisfaction performance, while not significant in low competition regimes, became significant for medium and high international competition environments. The path coefficients for the relationship were significantly different among the three groups and was observed to be strongest in medium competition conditions. The results represented a reversal of the hypothesized negative moderator effect of competition on the linkage between quality practices and customer satisfaction performance (H6). The change in relationship magnitude between quality practices and customer satisfaction performance followed the hypothesized direction in going from medium to high competitive regimes. However, the relationship became nonsignificant in low competition conditions-conditions that were expected to be extremely conducive for the relationship. The path model analysis also indicated significant and similar relationships between customer satisfaction and firm performance across all three competitive environments. Finally, the analysis revealed that the post hoc relationship between high involvement work practices and firm performance remains significant (and equal) for low and high competition conditions. Surprisingly, the relationship was not significant for the medium competition group of firms. DISCUSSION This section discusses the implications of the results in the context of the conceptual framework. Also discussed are implications for theory and practice. Conceptual Framework The results of the full-sample structural model analysis of conceptual framework support the integration paradigm in quality management. Although earlier frameworks have been examined by studying individual linkages in isolation, or related through simple correlation analysis, this study provides a methodologically rigorous demonstration of the holistic nature of quality management. Validation of the practice-performance linkage adds credence to Deming's (1982, 1986) chain reaction theory of the performance implications of quality investments. The findings also support Anderson, Rungtusanatham, and Schroeder's (1994) quality management framework of relationships between organizational practices, quality practices, and process outcomes. Researchers have observed that piecemeal approaches to quality management are likely to lead to unsuccessful outcomes (Choi & Behling, 1997). The current study supports the notion of integrative, organization-wide implementation of quality precepts and programs. This has considerable implications for top management support, resource availability, and personnel "buy-in" into quality initiatives. Smaller firms without the capital or expertise to implement organization-wide programs may be at a disadvantage, relative to organizations of larger size and resources. Not surprisingly, Powell (1995) noticed that while quality management philosophies had permeated widely among Fortune 1000 firms, smaller firms had not adopted the quality philosophy to a great extent. The unavailability of complementary resources in smaller companies was suggested as one reason for this disparity. Conversely, one could argue that small organizations could capitalize on their limited size and nimbleness to propagate, inform, educate, and involve members in concerted quality strategies. More research is required to obtain a better understanding of these issues. A detailed discussion of the individual results of the full structural model follows:

H1: High involvement work practices have a positive correlation with quality practices.

The strong (path coefficient = .876) and significant relationship between high involvement work practices and quality practices points to the necessity of social and structural reengineering in the organization for successful investments in quality practices. High involvement work practices accounted for 77% of the variance in quality practices. This suggests that managerial attention to structural issues might be a prerequisite for quality practice initiatives. From the familiar strategy-structure perspective, a firm's strategic posture on quality would influence its choice of organizational work practices. If quality practice interventions are considered critical to survival or performance, management should prepare the ground by inculcating organismic elements in the organization. Lateral communications, coordinating structures such as cross-functional teams, and open information access would be some positive steps in this direction. Such structural and social reengineering is not easy to accomplish. The tasks involved are complex and difficult to learn. However, for these very reasons of social complexity, ambiguity, and imperfect irritability, firms that can restructure can generate sustainable competitive advantages (Barney, 1991).

112: Quality practices have a positive correlation with customer satisfaction performance. The positive path between quality practices and customer satisfaction performance offers a degree of resolution to the ongoing quality management benefits debate. While quality practices alone may not be enough to ensure customer satisfaction (path coefficient .36), the variance explained is not insubstantial (13%). Investments in the four domains of quality practices: supply chain management practices, quality resources and evaluation, quality training, and customer commitment should lead to improved delivery performance and increased customer satisfaction and retention. The item loadings offer a more detailed picture of critical success factors for each of the four quality practice domains (Figure 2).

Actionable areas in the supply chain management domain include technology sourcing, sensitive information sharing, and establishment of clear supplier performance goals. Supply chain management practices are becoming a priority area for many companies. The increased dependency on the supply chain for cost, quality, and other performance objectives requires a sharper emphasis on supply chain management practices. Garvin (1983) discovered that plants with explicit criteria and goals for supplier quality produced the highest quality products. Other research has found that open exchanges of technical, economic, and managerial information with the supply base facilitates quality certification and promotes relationship building for quality goals (Cole, 1981; Manoocheri, 1985).

The item loadings suggest specific steps that can be taken to improve the quality resources and evaluation dimension. Management should be explicit and unambiguous in its communication of quality goals. Clear articulation of quality objectives should be reinforced by visible resource commitments. A quality performance evaluation system should be formally developed for management and employees. For psychological reasons, it may be desirable for management to be conspicuous in its self-administration of quality-based evaluation systems, prior to imposing such metrics on firm employees. In the quality training domain, training in quality awareness and statistical techniques should extend to hourly employees. Advanced SPC tools may also be taught at all levels, as employees and managers gain in quality maturity. Customer commitment constituted the remaining domain of quality practices. The findings suggest that providing opportunities for customer-employee interactions leads to customer-relevant reliability and responsiveness goals for an organization. Most firms do not consistently afford such opportunities to shopfloor employees. Investments in customer complaints evaluation and expectation monitoring systems are also suggested in order to design products and processes to meet and exceed customer expectations.

H3: Customer satisfaction performance has a positive correlation with firm performance.

Consistent with previous research (Lederer & Rhee, 1995), the data show a (marginally) significant relationship between customer satisfaction performance and firm performance. The relatively low (. 152) path coefficient can have different explanations. Companies might focus their customer satisfaction and retention efforts on existing markets, ignoring the development of new markets. Such a strategy may have adverse effects on long-term sales growth. The presence of omitted variable effects may provide another explanation. Firm performance is affected by multiple factors, including product price and cost, strategic positioning, effective supply chain management, manufacturing efficiencies and responsiveness, logistics, and marketing performance. From this perspective, the relationship found between quality-driven customer satisfaction and firm performance cannot be considered trivial. The findings provide grounds for management to justify resource commitments for quality programs.

The post hoc fitted relationship between high involvement work practices and firm performance is supported by the human resource management literature. Specifically, formal information sharing and coordination, employee participation, and job design are among the work practices that have been widely linked with firm-level outcomes (Huselid, 1995). Firms implementing innovative work practices that emphasize teams, cooperation, and problem resolution had higher productivity and lower costs relative to firms using traditional work systems (Katz, Kochan, & Gobeille, 1983; Cutcher-Gershenfeld, 1991). The direct influence of high involvement work practices on firm performance is thus entirely consistent with the literature. The value added by this study lies in the identification of the significant quality practices and customer-satisfaction-mediated indirect effects of high involvement work practices on firm performance. Contingency Effects-Hypotheses 4, 5, and 6 Differences in multigroup construct means Highly dynamic competitive environments appear to stimulate the development of high involvement work practices. The results of the latent means structural analysis (Table 5) indicate that mean investment in high involvement work practices was significantly higher in the high international competition group, compared to the low competition group. International competitive pressures have engendered social, cultural, and structural moves towards more agile, open organizations in U.S. industries (Kanter, 1989). Companies protected from such sophisticated, intense competition seem to be averse to transition to nontraditional structures. The complexity and costs of organizational restructuring may deter initiatives to rock the boat, until the threat of international competition looms large on the horizon. A responsive posture to competitive threats could lead a firm to simply run out of time in its efforts to meet competitive demands. A proactive approach to business change would scan and anticipate changes in the environment for taking timely action.

A difference in means was observed for quality practices between firms facing low and medium/high degrees of international competition. International competition appears to drive improvements in the quality practice domains of supply chain management, quality training, quality resources and evaluation, and customer commitment. Testimony to this observation is provided by the 1980's surge in quality practices in the U.S. The loss of market share to better quality overseas competition galvanized firms such as Ford, Xerox, and Motorola to invest in quality improvement practices. Although not explicitly investigated in this research, it is a fact that many large firms facing international competition have international operations and international supply chains. The configuration, transference, and coordination of quality practices across multinational, multiple ownership supply chains is a daunting task for the best managed firms (Young, 1992). Difficult goals raise the level of required effort, necessitating an increased emphasis on quality practices.

The analysis did not show any significant means difference in customer satisfaction performance (or firm performance) among the different groups. This could be because different competitive environments encounter different customer expectations of quality. For example, the automobile industry in India was protected from international competition until a few years ago. Buyer quality expectations were established by the quality levels of indigenously produced vehicles, with designs and technologies from the 1950s. Consequently, customer expectations were conditioned and fulfilled in terms of the limited range of product choices available in the market. However, with economic liberalization and exposure to international products and services, the same customers now demand world-class quality in automobiles (and other products). Customer satisfaction is a relative concept. Firms in different competitive environments benchmark their quality performance using available competition and remain content with their relative customer satisfaction performance perceptions. Where firms may falter is in the process of transitioning to the quality demands of a more competitive, exacting market environment. Differences in multigroup construct relationships The stability of interconstruct relationships across low, medium, and high international competition conditions was examined using multigroup path analysis. The results raise some intriguing issues. The relationships between high involvement work practices and quality practices, and between customer satisfaction performance and firm performance, proved to be significant and not statistically different among the three groups, The relationships between quality practices and customer satisfaction, and between high involvement work practices and firm performance, were found to be different across the groups. Each of these competition-moderated relationships is discussed in turn.

Companies in the low international competition category did not show a significant relationship between quality practices and customer satisfaction performance. However, the relationship becomes significant in both medium and high competition environments, and is strongest in medium competition conditions. Why should the relationship between quality practices and customer satisfaction performance peak in medium competition regimes, in sharp contrast to the (hypothesized) dominant effect of quality initiatives in low competition markets?

An examination of the data revealed that the length of time a company has engaged in formal quality plans and programs is affected by the intensity of its international competition. Companies in high competition regimes were found to have invested in quality plans for the longest period of time, followed by companies in medium and low competition environments. The means for the high, medium, and low competition groups were 8.85, 6.21, and 5.53 years, respectively. ANOVA confirmed that these means were significantly different from one another (p < .01). There is evidence to indicate that quality investments require a gestation period to impact performance (Dusseau, 1996; Handfield & Ghosh, 1994). Firms in benign environments with the shortest history of investments in quality could still be in the gestation period for performance gains. As time goes by (and competition increases), the investments made earlier begin to mature-witness the strong effect of quality practices on customer satisfaction in medium competition environments. Until these positive developments are experienced, however, management may not be inclined to incur further quality expenditures. Bolstered by these performance gains, increasingly pressured by competitive moves, and now past the initial learning curve, management increases investments in quality practices, as reflected in the higher quality practice means of the medium and high competition groups. The performance benefits accrue, albeit at a decreasing rate. The law of diminishing returns operates, competition catches up, quality becomes a routinized customer expectation, and the foci of customer satisfaction now probably shifts to price, customization, speed, or product innovation. This is suggested by the relatively weaker (vs. medium competition) though still significant relationship between quality practices and customer satisfaction performance in the high competition group. Figure 4 shows the envisaged relationship between quality practices and customer satisfaction performance across different levels of international competition. A confirmatory quadratic regression run with customer performance as the dependent variable and quality practices as the independent variable revealed a significant effect (p < .05) of the quality practices squared term on customer satisfaction performance.

These results have clear strategic implications for marketing and business strategy. As with any other technology, quality begins to lose its competitive edge with rapid dissemination. It can offer economic rents to early adopters (after an initial gestation period), when competitors are slow to adopt similar technologies. A strategy designed to exploit first-mover quality excellence can fetch positive, albeit temporary returns in a low/moderate competition market. Toyota and Suzuki are exemplars of such strategic successes in Southeast Asia. Suzuki entered a technologically obsolescent auto market in India and swiftly wrested substantial market share from domestic manufacturers by exploiting its considerable quality advantage. Late entrants such as Ford, GM, and Hyundai face escalated consumer quality thresholds and have not been able to parlay quality performance into a competitive advantage. Higher quality can be exploited through increased customer satisfaction, market segmentation, and premium pricing. Under less competitive conditions, firms may be able to create rare, imperfectly irritable, and causally ambiguous quality attributes for firm-specific competitive advantage. The laws of economics suggest that such asymmetric positions are unlikely to last in any market, unless specifically protected by law or tariff. Economic returns attract new entrants and investments, unless strong and persistent entry barriers exist-an unlikely proposition in the long term, even in controlled economies such as China and India. While low to moderate competition conditions last, however, management can customize strategies for individual markets to maximize firm performance. Although it is doubtful if poor quality will sell anywhere in the world today, investments in work and quality structures and practices can be stabilized in low/moderately competitive markets and still yield excellent perceived customer satisfaction. As long as those customer quality expectations remain static, market and financial performance can be attained without increasing quality-related investments. However, companies may choose to be proactive in introducing quality improvements to preempt the inevitable threat of new entrants in such attractive markets. Such quality-generated advantages may not accrue to firms that operate in quick response, agile, high competition conditions. In such environments, quality innovations could be quickly imitated and standardized into routine customer expectations. Companies have to invest more in quality-related practices just to be able to maintain customer satisfaction and market position status quo. Companies that cannot do so are forced from these markets. A parallel can be found in the technology literature, where competitive intensity was reported to actually adversely moderate the relationship between technology investments and firm performance (Dean & Snell, 1996).

With quality becoming more of an order-qualifier in highly competitive environments, companies also must pursue differentiation strategies through other means. New product introduction and customization are some alternatives. An example can be found in the varying strategic emphasis placed on quality by the U.S. auto industry at different points in time. Faced with the Japanese onslaught during the early 1980s, U.S. auto makers rapidly adopted and advertised quality concepts to meet rising market expectations. However, quality gains were slow in coming, possibly due to gestation/learning factors. Quality persisted as a selling criterion for quite some time, with imports successfully employing quality-based differentiation strategies for market share growth. The late 1980s and early 1990s saw U.S. auto manufacturers narrow the quality gap and begin to recoup customer confidence in their products. Although competition has since intensified further, quality no longer remains a primary strategic differentiator in the U.S. auto industry. Customers expect a minimum quality/cost standard in all product offerings. Marketing strategies in the auto industry now pursue differentiation through new product introductions, innovative financing programs, and special after-sales service packages. However, even low-cost product positioning strategies in high competition environments cannot neglect quality, for customer dissatisfaction and forced market-exit concerns. Kia Motors is a recent exemplar, that while adopting a bottom-line pricing approach, simultaneously seeks to reassure customer quality concerns by providing extended product, parts, and service warranties, and reliability-centered advertisements.

The differential impact of competition on the relationship between quality practices and customer satisfaction performance has equally significant, albeit subtler, implications for operations managers. Although the rate of return on quality practice investments declines with an increase in international competition, investments in quality structures and practices have to be increased for market survival purposes. Manufacturing, design, and sourcing will be required to improve quality standards on a continual basis, even though such performance improvements may not be proportionately reflected in gains in customer satisfaction performance. One notable issue that arises is the possibility of loss of management interest in quality initiatives. Obtaining resources for manufacturing and supply chain qualityrelated programs will become more difficult in the absence of a visible link between such investments and customer satisfaction. Symptoms of customer dissatisfaction become evident only when competition moves ahead in the quality arena and market share begins to be adversely affected. A resurgence in management interest in quality at that time may come too late in the day. Sterman et al. (1997) noted a similar chain of events in a firm that witnessed management interest in quality decline, as the relationship between quality investments and performance became less perceivable.

Our tri-group path analysis also revealed interesting differences in the relationship between high involvement work practices and firm performance among different competitive environments. High involvement work practices affected firm performance similarly in both low and high competitive conditions, and yet had negligible impact in the medium competition group. The differences in high involvement work practice means followed an identical pattern. The mean in the high competition group was significantly higher than the mean in the low competition group. The mean in the medium competition group was not statistically different from those in either the low or the high competition group. Because the work practice-firm performance path was a post hoc model modification, there are no a priori rationales to offer for the findings.

A possible explanation could be that companies in low competitive regimes that begin investing in high involvement work practices such as cross-functional teaming and interfunctional coordination, obtain early gains in work performance. Teaming and coordination involve enhanced access to cross-departmental information and delegation of decision-making responsibility. Employees feel enthused with the new open work environment, and performance improves. There is evidence that the financial returns associated with progressive work practices can be substantial (Huselid, 1995). With the passage of time, increasing competitive pressures demand increased performance. However, having already plucked available low-hanging fruit, companies discover that further work practice initiatives do not yield continued performance returns. Consider the well-known dysfunctionalities of teaming, including loss of identity, stymied career paths, and disconnects in evaluation and compensation systems. Thus, companies facing moderate forms of international competition may not have enough incentive to increase their investments in high involvement work practices. Interfunctional coordination and teaming involve learning. Research reports that inertial organizations characterized by benign competitive environments display truncated learning patterns (Doz, 1996). Competition is the primary stimulus for organizational experimentation and learning (Barnett & Hansen, 1996). Companies in highly competitive markets are compelled to be more aggressive and persistent in their improvement efforts, and may continue to experiment with organizational interventions, hoping for performance gains (that eventually materialize as they mature in their integration expertise). Osterman (1994) reported that work practice innovations are typically found where the product market has international competition.

Another factor to consider is that companies under competitive pressure may prefer to re-engineer work structures to the fullest extent, before turning to more capital-intensive, fixed-nature investments in technology. Researchers have also noted that as work practices grow in complexity, complementary assets in the form of budgetary support, employee release time, appropriate performance evaluation and reward procedures, and other human resource and technology practices are necessary to enhance implementation and performance (Pils & MacDuffie, 1996; Ichniowski, Shaw, & Prennushi, 1997). Companies with strong competition have necessarily developed competencies in many such areas to survive. The work practices-performance relationship can be exploited with some thought by operations managers. Managers may wish to emphasize the considerable direct effects of organizational restructuring on firm performance, in addition to its indirect impact through quality practices and customer satisfaction performance. The lesson here for operations managers is to persist while adopting high involvement work practices, and prepare management for anticipated plateaus in performance raW me

Viewed collectively, our study provides several insights of theoretical and managerial interest. The findings support the viability of the conceptual quality management framework. The synergistic associations between work practices, quality practices, customer satisfaction, and firm performance are entirely consistent with the essence of Deming's (1982, 1986) chain reaction theory of quality. Introducing international competition in a moderator role extends the theory by providing an understanding of how competition can affect quality-performance relationships. From a practical perspective, manufacturing can adopt an aggressive role in shaping business strategy by informing and educating management not only of its own quality capabilities, but also about significant quality initiatives observed in the competition. In order to be able to integrate its strengths and knowledge with business strategies, manufacturing will require enhanced strategic vision and market-scanning systems. It should develop or participate in systems to monitor and interpret changes in competitive intensity, competitor product/service quality strategies, and shifts in customer expectations. Cross-functional teams with active input from marketing and sourcing may be a useful mechanism for this purpose. Regular participation by manufacturing in business strategy meetings, manufacturing inputs into marketing plans, and key processes ownership may also be useful. It may not be easy for manufacturing to perform these tasks. Manufacturing and supply-base potential may not be readily apparent to marketing or finance-oriented management. Such latent capabilities should be highlighted by manufacturing at every available forum, including quality-based business case presentations to senior management, clearly outlining the strategic impact of quality investments in different competitive conditions.

CONCLUSION This study contributes to the operations management literature in several ways. The primary contribution consists of the development and test of a contingency framework of quality management. Quality program failures to improve performance have been attributed to a variety of causes, ranging from cultural misfit to poor management support and misaligned objectives. This research proposed and validated competitive intensity as an alternative explanation for the absence of returns from investments in quality capital. Obtaining customer satisfaction performance from quality practices was shown to be contingent on the degree of international competition present in the business environment. The findings indicated that consistently high quality investments are a necessary condition for customer satisfaction in conditions of moderate to high competition. The implications of these findings for strategic planning and practice were discussed.

This research also developed and tested a nomological framework of quality management. Within the framework, the role of work structure in quality management was examined separately, with interesting results. Earlier studies have collapsed organizational processes and quality practices in the same construct (Flynn et al., 1994; Ahire et al., 1996).

The conceptual framework remains to be examined in full. Future research can extend this study to include quality strategy, environment (product complexity, technology volatility, etc.), and other factors. The framework can also be applied to international settings. Insights on differential relationships among work practices, and quality practices and performance across different cultures and technology levels should be useful to research and industry.

There are limitations to the study. Our operationalization of international competition was limited to a single, albeit important region, Japan. Future studies could include competition from Europe and Asia, and also possible competition within affiliated units (e.g., Honda America vs. Honda Japan). A research issue that merits attention is the impact of U.S. competition on the practices and performance of foreign industry. Another limitation, imposed by the low sample size of the individual groups, was our use of path model analysis for the multigroup analysis, instead of continuing with full structural equation model analysis. Future research could try to obtain larger sample sizes in different competitive environments to allow the use of full structural model treatment for multigroup analysis. Finally, our operationalization of quality outcomes was confined to customer satisfaction performance. Although customer satisfaction is admittedly a primary objective of quality management, other outcomes such as process and product quality have been recognized in the literature. [Received: September 29, 1998. Accepted: August 23, 2000.]