COLLABORATION AND EVALUATION IN THE SUPPLY CHAIN: THE IMPACT ON PLANT-LEVEL ENVIRONMENTAL...

1. Introduction

During the past decade, an increasing awareness of climate change, the depletion of natural resources, and the generation of waste have moved environmental issues from the periphery of management attention to the forefront. International agencies and national governments met a number of times over the last decade (e.g., Earth Summit, Rio, 1992 and Kyoto, 1997) to establish broad goals for environmental improvement in areas such as greenhouse gas emissions. Moreover, industry associations have adopted voluntary standards and codes, such as Responsible Care in the chemical industry, and established knowledge-sharing networks for environmental issues, such as the Printer National Environmental Assistance Center in the printing industry. Thus, environmental concerns have become a mainstream issue for manufacturing organizations as management is pushed and encouraged to develop appropriate policies and implement related new technologies in their operations.

During this same period, globalization and intensified competition have pressed managers to devote greater attention to supply chain management. Firms increasingly rely on their supply network to design more complicated subcomponents and products, implement more complex process technologies, and meet higher customer expectations. Among these expectations, greater attention is focused on suppliers' social responsibility (Drumwright 1994; Carter 2000) with a particular attention to the responsible use of natural resources and environmental practices (Min and Galle 1997). Options to encourage better practices in suppliers include requirements for audits and certification (e.g., ISO 14001) and obligations to take back used product for recycling (Krut and Karasin 1999; Lippmann 1999), among others. Some organizations have extended this thinking to product stewardship, with an assessment of life-cycle impact of their products (McIntyre, Smith, Henham, and Pretlove 1998b) and assurance of their proper use by customers (Snir 2001).

In response, managers must make plant-level investment in environmental technologies that both comply with regulations and meet customer expectations within the context of other supply chain activities. If more proactive, operations managers can also seek to implement more sustainable and greener practices ahead of these requirements. To that end, greater collaboration among the members of a supply chain can foster the development of improved environmental systems through both technological innovation and better resource management (Handfield, Walton, Seeger, and Melnyk 1997; Geffen and Rothenberg 2000), which can reduce the overall environmental impact across one or more segments of the supply chain.

This paper extends earlier research in supply chain management to further examine its relationship to environmental investment at the plant level. The paper's major contribution is the consideration of the linkage between two types of supply chain activities-namely collaboration and evaluation-and the level and form of plant-level environmental investment. Drawing from the environmental management and supply chain management literatures, the following two sections define the constructs and hypothesized relationships that underlie environmental investment and supply chain management. Next, the survey methodology and construct measurement are detailed. Finally, the results and discussion are presented, along with implications for both future research and managerial practice.

2. Investment in Environmental Management

Since the early 1970s, environmental regulation directed at manufacturing has tended to target pollutants generated by products and manufacturing processes. Such regulations, which have been the predominant approach in much of the world and frequently termed command-and-control (Organisation for Economic Cooperation and Development 1995), specify particular limits for the quantity of pollutants that a manufacturing plant may emit (Bonifant, Arnold, and Long 1995). As a result, significant capital and operating investment has been targeted at reducing the release of particular pollutants into the natural environment.

More recently, various programs and initiatives designed to reduce pollution before its generation have attracted a great deal of interest-motivated by public policy (Freeman et al. 1992), improved technical understanding and broader dissemination (U.S. Environmental Protection Agency 1989), corporate pragmatism, and economic self-interest (Royston 1979; Kemp 1993). Pollution prevention has the potential to eliminate environmental hazards more efficiently and effectively across the entire production system (Porter and van der Linde 1995; Florida 1996).

Yet, despite the appealing rationale behind pollution prevention, manufacturing organizations struggle to move from a pollution control-oriented bias to a prevention-oriented bias. One reason for the apparent reticence is that pollution prevention typically requires more radical technological innovation, with significant changes to product and/or process (Jones and Klassen 2001). In contrast, end-of-pipe pollution controls tend to rely far more heavily on incremental innovation, where existing products and processes change little. Thus, a manufacturer's investment in environmental management is affected both by the level (i.e., extent) of resources invested in environmentally related projects and by the form (i.e., allocation) of that investment across different types of environmental technologies.

2.1. Environmental Technologies

It is important to emphasize that environmental technologies are broadly defined to include design, equipment, and operating procedures that limit or reduce negative impacts of products or services on the natural environment (Shrivastava 1995; Klassen and Whybark 1999b). Based on theoretical and empirical grounds, researchers have developed a number of frameworks to categorize environmental technologies, i.e., form of technology (a list of seven frameworks is summarized in Jones and Klassen 2001). It is very important to explicitly separate motivation and timing (i.e., reactive versus proactive) from the form of the environmental technologies investment, as there has been growing evidence that proactive environmental strategies can utilize multiple forms of technologies (Aragon-Correa 1998; Klassen and Whybark 1999a). For example, investments in both preventive- and control-oriented technologies might be undertaken ahead of new air emission regulations, making both part of a proactive strategy. Thus, while motivation and timing are critical research issues, they are beyond the scope of this paper. Instead, the focus here is on how supply chain activities are linked to the level of investment in environmental technologies and the form of that investment.

In its own right, the form of investment in environmental technologies is important to track patterns in individual plants, firms, or industries over time, to monitor changes, and to assess implications for performance. Aggregating over time, the collective, observable pattern of investment can be thought of as a portfolio-termed an environmental technology portfolio-with a plant-specific composition of different forms of environmental technologies. Based on the operations strategy and environmental management literatures, three exhaustive and mutually exclusive categories have been proposed: pollution prevention, pollution control, and management systems (Klassen and Whybark 1999a).

Pollution prevention technologies are structural investments in operations that involve process- or product-based changes. Material substitution and source reduction are examples of such technologies. While investments in better housekeeping and integration of environmental considerations in production planning and scheduling have been included in some discussion of pollution prevention (e.g., Hart 1995), these infrastructural investments are captured elsewhere to retain the historical differentiation of manufacturing strategy research between structural and infrastructural investments (Wheelwright 1984). Thus, the emphasis for pollution prevention here is fundamental change to the physical product and/or process.

In contrast to pollution prevention, pollution control technologies are structural investments that capture, treat, or dispose of pollutants or harmful byproducts at the end of a manufacturing process. To accomplish this, additional operations or equipment must be added to the end of an existing manufacturing process, thereby leaving the original product and manufacturing process virtually unaltered. Pollution control technologies include both end-of-pipe equipment and remediation projects to clean up earlier harmful practices.

Finally, management systems are infrastructural investments that improve the way that environmental issues in manufacturing are managed. These systems include introducing formalized procedures for evaluating environmental impacts during capital decision budgeting, routines for assessing the environmental implications of using new raw materials in product design, environmental audits, employee training to prevent spills and reduce waste, and better housekeeping practices. These systems often include elements that both control and prevent environmental degradation.

3. Environmental Management and Supply Chain Management

Manufacturing plants cannot be managed from an isolated perspective: explicit recognition of the interactions with upstream and downstream supply chain members is needed (Davis 1993). Hence, supply chain management has been defined as all activities associated with the flow and transformation of goods from raw material through to the end product, including operations within the focal firm (Handfield and Nichols 1999). This definition captures two important supply chain characteristics: first, it is a holistic and systemic view that goes beyond a simple notion of a value chain; and second, in addition to material flows, information flows reflect the bidirectional nature of supply chains, which comprise upstream and downstream relationships. While research must remain cognizant of the longer supply chain, for pragmatic analytical purposes, this paper adopts a narrower perspective of the supply chain, focusing only on immediate suppliers and immediate customers.

The interorganizational activities between a plant and its suppliers and customers might embed environmental issues within a broader set of supply chain concerns (e.g., quality appraisal, product development, delivery monitoring). Thus, a range of supply chain activities can potentially influence environmental management within a plant. As a result, several perspectives exploring the linkages between supply chain and environmental management have emerged from the purchasing literature (Min and Galle 1997; Zsidisin and Siferd 2001). A synthesis of this literature, together with research on supplier development (e.g., Krause 1999) points to at least two generally accepted characteristics that capture supply chain activities related to environmental management:

* joint collaborative activities between a plant and its suppliers and/or its customers directed at achieving sustained improvements in environmental performance (Handfield et al. 1997; Geffen and Rothenberg 2000); and

* upstream evaluative activities taking the form of information gathering by manufacturers to assess and to monitor environmental management and performance of their suppliers (Krut and Karasin 1999; Hines, Lamming, Jones, and Rich 2000).

3.1. Collaborative Activities and Environmental Investment

Collaborative activities are characterized by tacit knowledge integration, which occurs through information exchange in a rich communication setting (Purdy and Safayeni 2000). These activities can result in product adaptation, fundamental process modifications, and cooperation in logistics. Suppliers with technical knowledge regarding their products or customers with new product designs can influence a manufacturing plant's environmental investments and technology selection (Ashford 1993). Through knowledge sharing, collaborative activities reduce uncertainty, willingness to change, and other sources of resistance frequently associated with the lack of investment in environmental technologies (Ashford 1993; Kemp 1993).

Collaboration along the supply chain also helps management to identify and evaluate a greater variety of options that might address particular environmental challenges (Bonifant et al. 1995). These options are often linked with improvement in other manufacturing performance dimensions, such as productivity and quality (Porter and van der Linde 1995), which can prompt further investment in projects related to the environment. The incentives to do so are especially strong if shared-savings contracts are in place to encourage investments and innovation from multiple parties along the supply chain (Bierma and Waterstraat 2000; Corbett and DeCroix 2001). For example, Xerox's design-for-remanufacturing initiative required product-based partnerships with its suppliers to increase the quantity of reused and remanufactured components (McIntyre, Smith, Henham, and Pretlove 1998a), coupled with a benefit-sharing mechanism (Krut and Karasin 1999). Thus, collaboration is expected to increase the overall investment related to environmental management.

H1: As collaborative activities between a plant and its suppliers and customers increase, the level of investment in environmental technologies in the plant increases.

While encouraging a higher level of investment in environmental management, collaboration among supply chain members can also alter the means by which any negative environmental impact is reduced. As noted earlier, pollution control only affects the operations of a single plant, and can be implemented in an isolated fashion. In contrast, product-based actions that leverage "design for environment" (DFE) to make such improvements as reduced packaging are likely to be much easier to recognize and implement as one aspect of other collaborative supplier-plant or plant-customer activities. This is likely to be the case for other initiatives too, such as joint recycling of parts and components, and process changes that reduce the use of hazardous materials.

A good example of such collaborative activities can be found in chemical management services. These services, undertaken by a chemical supplier, guide a buying plant to properly use and handle chemicals, thereby potentially reducing spills and consumption. For example, Castrol, a lubricant supplier to the automotive industry, worked jointly with one of its customers' plants. The resulting process modifications reduced the consumption of lubricants, leading to lower cost and environmental impact (Reiskin, White, Kauffman-Johnson, and Votta 2000). Similar case-based evidence has been reported elsewhere in the automotive industry. A paint supplier effectively worked on-site in the paint shop of an automaker to develop a better solution to the ever-increasing pressure for lower emissions of volatile organic compounds (VOCs) (Geffen and Rothenberg 2000). Thus, collaborative activities can shift environmental investment away from pollution control technologies toward more efficient and effective preventive technologies.

H2: As collaborative activities between a plant and its suppliers and customers increase, the form of investment in environmental technologies in the plant is increasingly allocated toward pollution prevention.

3.3. Evaluative Activities and Environmental Investment

Supply chain activities also involve monitoring and assessment, which are collectively termed evaluative activities. These activities emphasize gathering and processing information in order to assess operating performance including legal compliance, and to mitigate any associated risks. Many evaluative activities are based on preestablished performance standards for the quality of the materials, delivery reliability, speed, and customer service that are driven by priorities of downstream supply chain members (Leenders and Fearon 1997). Good supply chain practices also recognize the importance of supplier awards and feedback (Krause, Scannell, and Calantone 2000).

Because large numbers of suppliers might be covered by the same evaluation process, these activities are often conducted in a transactional context and focus on specific, easily verifiable measures. Evaluative activities take the form of questionnaires, nonregulatory standards, and third-party audits that also extend to environmental criteria (Min and Galle 1997; Walton, Handfield, and Melnyk 1998). For example, AMD, a manufacturer of integrated circuits for computers, made extensive use of questionnaires and audits to evaluate environmental practices and performance of its suppliers (Krut and Karasin 1999). In addition to identifying high-risk suppliers, management reduced the likelihood of supply interruption problems and associated liabilities. This evaluation is expected to drive greater investment in environmental technologies across multiple members of the supply chain, driven both by greater external scrutiny (i.e., evaluation of suppliers by a plant) and the obligation not to appear hypocritical (within the plant).

H3: As evaluative activities between a plant and its suppliers and customers increase, the level of investment in environmental technologies in the plant increases.

Alternatively, management might find evaluative activities a convenient means by which to shift environmental responsibility upstream, with little meaningful change in their own operations (Florida 1996). In addition, because of the significant cost of external evaluative activities (Krut and Karasin 1999), plant managers might be tempted to reduce their own internal investment.

Given that evaluative activities tend to focus on easily verified measures, these activities also can be expected to affect the form of any investment in environmental technologies. First, rather than encourage systemic changes, which often require long-term planning and joint implementation, suppliers might instead favor "quick fixes" such as those offered by end-of-pipe pollution control equipment (Noci 1997). Second, as the objective of evaluative activities is primarily to monitor and control, thereby forcing improvement, these activities are in many ways similar to environmental regulations that prescribe compliance for particular emissions and specific paperwork. Focusing on compliance can implicitly limit the set of technological options considered, which in turn favors the adoption of end-of-pipe technologies (Bonifant et al. 1995).

H4: As evaluative activities between a plant and its suppliers and customers increase, the form of investment in environmental technologies in the plant is increasingly allocated toward pollution control.

4. Research Methods

4.1. Sample

The relationship between supply chain management and environmental investment was explored using a plant-level survey. This research was an extension of an international research project undertaken in about a dozen countries to survey the motivations for, implementation process of, and benefits associated with ISO 9000 and 14001 certification for quality and environmental management, respectively. To a short international survey instrument, additional sections were appended for Canadian plants to measure constructs related to environmental investment and supply chain activities. Thus, only the Canadian data could be used to assess the hypotheses proposed in this paper.

Rather than focus on a single industry, a cross-section of manufacturing plants in Canada was desirable to ensure variation in supply chain relationships and contextual variables, and to improve generalizability. To fulfill obligations to the international research project, all surveyed plants had either ISO 9000 certification (e.g., ISO 9001), ISO 14001, or both. Certification also offered one signal that some attention was being paid to managing external supply chain relationships, as ISO 14001 certification has been linked with supplier and customer benefits (Darnall, Rigling Gallagher, and Andrews 2001). While certification did not ensure strong performance-that has been cited as a weakness of the standard-it did ensure that a basic system aimed at improvement was present that captured both suppliers (e.g., raw materials) and customers (e.g., feedback on satisfaction).

Based on these criteria, a sample frame was constructed from all ISO 9000 and 14001 certified firms in Canada (World Preferred Inc. kindly provided access to their comprehensive database). Of the approximately 9,000 certified manufacturing plants in Canada, only about 5% had ISO 14001 certification. Thirteen SIC codes were chosen based on having large numbers of ISO certifications (SIC 24-30, 33-38). All ISO 14001 certified plants were targeted in these SIC codes (i.e., 261 plants) along with a similar number of randomly selected ISO 9000-only certified plants from the same industries (i.e., 300 plants).

After an initial telephone call to the plant manager to confirm contact information, a three-wave survey process similar to that prescribed by Dillman (1978) was followed. Conducted in 2001, the survey was offered to targeted plant managers in both official languages, English and French, to encourage national participation. A total of 202 plant managers returned the survey, yielding a response rate of 36%. Two tests of independence were performed to determine if the distribution of the respondents varied from the targeted sample. The chi-square statistic was not significant for each test: SIC code; and ISO 14001 certification. No evidence was found that the respondents were not representative of the target sample.

4.2. Environmental Investment

The level of investment in environmental management was assessed by using two metrics (Appendix). First, respondents reported the percentage of the capital budget allocated to environmental projects. Second, a three-item, five-point scale captured the extent to which resources have been invested in three programs related to environmental management over the previous 2 years: recycling of materials, pollution prevention, and waste reduction (Cronbach [alpha] = 0.88).

Managers were also asked to report the allocation of resources between different forms of environmental technologies using a previously validated scale (Klassen and Whybark 1999a). Respondents were asked to divide 100 points across the categories of environmental technologies based on the use of resources over the previous 2 years. For consistency with plant-level manufacturing practices, pollution prevention was specified as two subcategories: product adaptation and process adaptation. Pollution control was further specified as two categories: end-of-pipe controls and remediation. The final category was management systems. For each category, a short definition was provided to ensure consistency between respondents (Appendix).

4.3. Supply Chain Management

Based on earlier research, collaborative and evaluative activities were assessed by drawing on a variety of items reported in the literature (Krause 1999; Krause et al. 2000). Using these items, two scales were constructed for plant-initiated collaboration and evaluation of primary suppliers. Another two scales were constructed for collaboration and evaluation initiated by primary customers toward the plant, yielding a total of four scales (Appendix). It should be stressed that all the scales were reported from the perspective of the responding plant manager to ensure that management did not have to speculate about the operations of another firm (i.e., my plant assesses supplier performance. . . ; my customers assess my plant's performance. . . ; etc.). Thus, these four scales only capture supply chain activities that were oriented upstream (Figure 1).

To assess validity, first, the scales were tested for internal reliability ([alpha]). All exceeded 0.67 and were corroborated by composite reliability measures all exceeding 0.69 (Table 1), which is acceptable (Nunnally 1978). Second, a confirmatory factor analysis (CFA) was conducted by using maximum likelihood estimation for the two factors (i.e., scales) pertaining to plant-initiated activities with their suppliers, then for the two factors related to the customerinitiated activities with the plant. All parameter estimates were statistically significant (Table 1). While the chi-square statistics were significant, other fit statistics were generally within acceptable ranges (i.e., [chi]^sup 2^/df < 3, adjusted goodness of fit index > 0.9, comparative fit index > 0.9, normed fit index > 0.9) (Hair, Anderson, Tatham, and Black 1995).

Third, convergence and discriminant validity were also assessed. Convergence validity is supported if the correlation between the latent constructs is significantly greater than zero, as was the case here for both pair of constructs (Table 2, p < 0.01). Discriminant validity is supported if a two-factor measurement model fits significantly better than a constrained single-factor model, which was the case here ([Delta][chi]^sup 2^ = 9.6 and [Delta][chi]^sup 2^ = 9.1, df = 1; p < 0.01). Based on this CFA analysis, the average of the scale items was used in the subsequent regression models.

4.4. ISO 14001 and Other Plant Characteristics

ISO 9000 and 14000 are families of management systems standards established by the International Organization for Standardization (ISO) to formalize strategic approaches to quality and environmental management. Plant managers were asked to report the year in which their plant was first certified for each standard, namely ISO 9001 or 9002, and ISO 14001, if applicable. The status of the plant with respect to ISO 14001 certification was coded by using a dichotomous variable. Of the 202 surveys, 110 plants (54.5%) have ISO 14001 certification.

Given the broad range of manufacturing plants represented in the sample, several additional variables were included to account for plant-specific characteristics. Plant size (Grant, Jones, and Bergensen 2002) and equipment age (Klassen 2001) have been associated with environmental management. In addition, differences in process configurations were expected to be related to investment in environmental management. Based on two-digit SIC code, plants were classified into one of four production process categories: continuous flow (e.g., petroleum refining); discrete non-engineering (e.g., textile products); discrete engineering (e.g., machinery); and mixed production (e.g., printing and publishing) (per Statistics Canada; Arundel and Sonntag 1999).

5. Results

Bivariate correlations for the variables are reported in Table 2. Not surprisingly, the two plant-initiated supply chain scales were highly correlated, as well as the two customer-oriented scales. However, discriminant validity was maintained. Given potential concerns about multicollinearity, all variables were standardized with the exception of the dichotomous indicators. The condition indices, variance inflation factors, and incremental variance explained were assessed as one step in the regression modeling (Hair et al. 1995): no major concerns related to multicollinearity were detected.

Hierarchical linear regression was used to test the relationship between supply chain activities and the level and form of investment in environmental technologies. The regression models were structured to report the incremental variance explained by three different groups of variables: first, plant characteristics including ISO certification; second, plant-initiated supply chain activities (i.e., evaluation and collaboration); and third, customer-initiated supply chain activities. The first set of two models report the final parameter estimates and incremental variance explained for level of investment in environmental management (Table 3, related to Hl and H3), and the second set of three models summarize the results for form of environmental technologies (Table 4, related to H2 and H4).

5.1. Level of Environmental Investment

It is noteworthy that several plant characteristics were strongly related to investment in environmental management. Not surprisingly, Canadian plants with continuous flow processes invested significantly more in environmental management than discrete engineering production processes, based on capital expenditures (Table 3). This is consistent with data from the U.S. Bureau of the Census (1985-1991). Similar results were evident for the five-point scale assessing recent investment in environmental programs. Larger plants also invested relatively more in environmental management (while not quite statistically significant for environmental programs, it was directionally consistent). Finally, ISO 14001 certification was significantly related to both measures of the level of investment.

In general, only limited support was found to link collaborative activities (H1) and evaluative activities (H3) with the level of investment in environmental management. First, plant-initiated collaboration with suppliers was significantly related to investment in environmental programs (p < 0.05). Second, customer-initiated evaluation was positively related to investment in environmental programs (p < 0.01).

5.2. Form of Environmental Investment

Based on the findings related to plant characteristics noted in the previous section, it was very interesting to explore if there was a consistent pattern to how this investment was allocated. The results reported in Table 4 indicate that, for plants with continuous flow processes, relative to those with discrete engineering production, this investment was clearly allocated toward pollution control technologies and away from management systems. There was also a significant indication that management systems received a proportionately smaller allocation as the plant size increased (p < 0.05). Moreover, while possibly expected, the relationship between the average age of plant equipment and the relative allocation of investment in pollution control was marginally significant, with pollution control receiving higher allocation as the age increased (p < 0.1).

Finally, as might be expected, ISO 14001 certification was significantly related to an increased allocation of environmental investment toward management systems (p < 0.01). This shift toward management systems came at the combined expense of pollution prevention and pollution control (recall that the three are linearly dependent).

Some substantiation was found for a linkage between supply chain activities and the form of investment in environmental technologies. First, greater customer-initiated collaboration was related to a shift in investment toward pollution prevention (p < 0.05), offering support for H2. However, collaboration with suppliers was not related to the allocation of investment across the three forms of environmental technologies. Finally, no evidence was found that supply chain evaluative activities influenced the allocation of environmental investment between the three forms of technologies in the plant, counter to H4.

6. Discussion

6.1. Supply Chain Activities

The empirical evidence relating evaluative supply chain activities with the level of investment in environmental management indicates that tighter control and monitoring by a plant of its suppliers can have significant, positive benefits. More specifically, customers who implement supplier audits, establish formal evaluation processes, and offer feedback to suppliers are likely to see positive changes in how suppliers regard environmental issues. The greater direct scrutiny by customers likely captured the attention of suppliers' plant managers and encouraged greater environmental investment. This presence of this relationship effectively demonstrates the important effect that customers can have on their supply base, and this effect extends beyond traditional manufacturing performance dimensions, such as cost and quality, to environmental management. These findings are quite consistent with at least one aspect of Xerox's experience, where the customer played a critical role in advancing greener designs and processes (McIntyre et al. 1998a).

To a lesser degree, there was limited evidence that collaboration with suppliers (e.g., site visits, exchanges of personnel, and technical assistance) also favored greater investment in environmental programs in the initiating plant. Together, these results may suggest that collaboration with suppliers reflects a change in the contract structure that redirects the responsibility for environmental programs from a plant to its suppliers. Thus, greater supply chain collaboration has implications for multiple echelons along the supply chain.

While customer-initiated collaboration did not affect the overall level of environmental investment, it did affect the allocation of that investment. For pollution prevention, customer-initiated collaboration was clearly more critical to implementation in a particular manufacturing plant than was that plant's collaboration with its suppliers. While any collaboration undoubtedly contributes to a better comprehension by both parties of each other's operations, structural change (i.e., implementation of pollution prevention) tended to occur in the supplier plant rather than the customer (i.e., upstream rather than downstream). Again, one impediment for a supplier to move forward with structural change is the resistance it might face from its customers: structural changes carry an element of risk, where the outcomes for complex processes and products may not be well understood. Possibly, this collaboration may offer a signal to the plant that customers want to see more radical, preventative structural changes, and are willing to accept some inherent risk. Alternatively, the customer may see collaboration as a way of fostering pollution prevention that it may be unwilling or unable to implement itself.

Collectively, these findings indicate what operations managers can expect from their investments targeted at improving environmental management. First, if a plant manager develops collaborative activities with suppliers, it is not likely to generate structural changes that prevent pollution directly in that manager's products or manufacturing processes. Instead, the pollution prevention is likely to occur in supplier plants. Carter and Carter (1998) reported similar findings for purchasing behavior, with the demands and pressures from downstream members of the supply chain having a significant influence on the focal plant's environmental purchasing behavior, while suppliers' actions did not affect behavior.

Second, because only customer-initiated collaboration was found to favor pollution prevention, plant managers must take other actions within their control to emphasize prevention in their own internal manufacturing processes and products. Such actions might include the development of more proactive environmental policies and commitment by senior management. It is also possible that internal efforts to shift away from pollution control toward pollution prevention may signal suppliers that similar demands will be made of them in the near future. In fact, although not explicitly assessed in this study, this signal may be a necessary condition to foster customer-initiated collaboration, and ultimately environmental improvement by suppliers.

6.2. ISO 14001 and Other Plant Characteristics

ISO 14001 certification was significantly, and consistently, related to more extensive investment in environmental management. Given that the reference group of plants had ISO 9001 certification, and thus arguably should have generally sophisticated management systems, this finding was quite encouraging. In particular, this result runs counter to the contention of some critics that ISO 14001 is simply a marketing tactic, and instead provides evidence that ISO 14001 is significantly related to plant-level investment. However, further research is needed to better understand if ISO 14001 certification directly results in more investment, or if another background variable, such as environmental orientation, drives both certification and investment.

The general type of manufacturing process had a significant impact on the level and form of environmental investment. While it was not surprising that continuous flow processes had higher overall investment, it was worthy of note that this investment still tended to more strongly favor pollution control technologies than in other manufacturing processes. By way of possible explanation, continuous flow processes have been recognized as being relatively more complex and capital intensive than other process configurations because of high automation and tightly coupled, interrelated steps. Thus, managers of such processes were likely reluctant to introduce technologies such as pollution prevention that can disrupt existing, presumably predictable processes (Rajaram and Corbett 2002). Management systems may also be less critical with capital intensive process. In contrast, pollution control can be more loosely coupled to an existing, complex process, and involves only incremental innovation (Jones and Klassen 2001). If this industry-level generalization can be applied at the microlevel within an industry, plants with simpler, less connected processes may have fewer barriers to implement a combination of pollution prevention and management systems, rather than pollution control.

Finally, the linkage between the age of plant equipment and the allocation of investment toward pollution control underscores one of the challenges that managers of older plants face. Instead of building capabilities for broad-scale environmental improvement, as suggested by pollution prevention, management tended to adopt the less-demanding and more familiar investment pattern of pollution control. In addition, smaller firms may invest relatively less in environmental management because of a lack of resources, less scrutiny by different stakeholder groups, or the inability to generate economies of scale.

Several limitations of this study are important to note. First, while the sample size was quite large given earlier research in operations management, statistical power may still be insufficient to detect smaller effects related to those hypotheses for which no support was found. As such, the primary contribution here is the identification of supply chain activities that were related, rather than establishing those that were not. Second, the potential for respondent bias is always a concern. However, validation was found for some variables in earlier research, such as the findings based on type of manufacturing process. Third, because the sampling frame only assessed plants with either quality or environmental certification, there may be concerns about generalizing to manufacturers that have chosen not to obtain (or could not obtain) either certification. Fortunately, particularly in the area of ISO 9001, this is a diminishing group. Financial performance was not controlled for, which might influence the degree of organizational slack available for environmental investments. Finally, only two types of supply chain activities were considered, both with an upstream orientation. Additional research considering downstream oriented activities, or other factors such as commitment (Krause 1999), might also offer new insights. A dyadic approach to complement the current research in chemical management services (Snir 2001) would also be an interesting avenue for further empirical research.

7. Conclusions and Future Research

Research in two streams, supply chain management and environmental management, provided the theoretical foundation for this investigation into how supply chain activities affect plant-level investment in environmental management. Earlier literature in supplier development and supply chain management indicated that important constructs include, at a minimum, collaborative and evaluative activities. Collaborative activities encompass such actions as site visits, exchanges of personnel, and technical assistance, while evaluative activities include assessing supplier performance, relying on supplier certification to reduce the need for inspection, recognizing supplier achievements, and offering feedback. While not exhaustive, these activities provide the means by which multiple echelons along a supply chain can encourage and enable environmental improvement in at least a semicoordinated fashion. Thus, the focus here was to explore the relationship between these supply chain activities and, first, the level of environmental investment, and second, the allocation of that investment between forms of environmental technologies.

Based on a sample of Canadian manufacturing plants, supply chain collaboration was found to significantly impact both the level and form of investment in environmental technologies. Of greatest importance, as customer-initiated collaborative activities increased, managers of the upstream plant increasingly allocated a greater proportion of resources toward pollution prevention, and away from management systems. In addition, increased collaboration also was significantly related to greater overall investment in environmental programs in the initiating plant. In contrast, only very limited evidence was identified that evaluative activities influenced investment, with some indication that greater evaluation of suppliers increased investment in their plants.

This study took a broad perspective of supply chain activities, and additional work is needed to examine in greater detail the implications of individual interorganizational supply chain activities. It is very likely that specific activities will prove more beneficial to environmental management and performance under particular circumstances. For example, environmentally focused collaborative activities (e.g., technical interchanges to develop pollution prevention) might prove critical for durable goods. In contrast, evaluative activities such as environmental audits and life-cycle assessment might be more important for consumer products. Finally, the relative position of a manufacturing plant along a supply chain, such as the physical or organizational distance from commodity material suppliers and end-consumers, might create additional challenges for managers seeking to develop a greener supply chain.1