SUSTAINABLE METAL RECYCLING SUPPLY CHAINS: PRIORITIZING SUCCESS FACTORS APPLYING COMBINED AHP & PCA TECHNIQUES

International Journal of Managing Value and Supply Chains (IJMVSC) Vol. 6, No. 3, September 2015 DOI: 10.5121/ijmvsc.2015.6303 31 S USTAINABLE M ETAL R ECYCLING S UPPLY C HAINS : P RIORITIZING S UCCESS F ACTORS A PPLYING C OMBINED AHP & PCA T ECHNIQUES Jitendar Khatri 1 and Ashutosh Dash 2 1 Department of Operations Management, Management Development Institute (MDI), Gurgaon, India 2 Department of Finance Management, Management Development Institute (MDI), Gurgaon, India A BSTRACT Triple Bottom Line (3 BL) Sustainability has been expressed as a tool in the hands of policy makers to achieve competitive advantage. Closed-Loop Supply Chain (CLSC) conceptualizes the design, control, and operation of the supply chain systems to recover the value from the product even after their useful life through the processes of reuse and recycle. Efficient management of CLSC operations in metal recycling industry will enhance 3BL sustainability. The present study investigates and shortlists critical sustainability factors in metal recycling CLSC operations. The study applies Analytic Hierarchy Process (AHP) to identify the most important factors and Principal Component Analysis (PCA) techniques to check the reliability and validity of these factors through statistical analysis. The study identifies five most important factors to achieve sustainability in metal recycling CLSC. Although the article is focused on the Indian metal recycling industry but factors can be relevant for researchers in other developing countries, where similar economic, social and environment al conditions exist. K EYWORD Analytic Hierarchy Process, Closed Loop Supply Chain, Principal Component Analysis, Triple Bottom Line, Sustainability 1. I NTRODUCTION During the first industrial revolution period (176-1820) and subsequent industrial revolutions traditional unit production methods were replaced by the mass production methods and manual operations were substituted with machines. These changes in production methods have lead to increasing growth in resource consumption all over the World. The large scale industrialization in the expanding global markets, coupled with increasing population, has created imbalance between the availability and consumption of natural resources. The simultaneous increase in demand and reduction in metallic resource reserves, the degradation of existing reserves, the uprooting of inhabitants associated with mining operations, the deteriorating natural environment, and rising awareness among public has compelled business, policy makers and governments to begin thinking in terms of sustainability. Sustainability is considered to consist of three components: the conservation of natural environment, benefits of the local communities and society at large and economic growth. Researchers argue assimilation of these three concepts into business policies and practices are of paramount importance for sustainable development (SD) [1], [2], [3]. Closed-Loop Supply Chain (CLSC) is one such practice adopted by businesses and it targets recovering the left over values from the product after it has served the intended International Journal of Managing Value and Supply Chains (IJMVSC) Vol. 6, No. 3, September 2015 32 purpose [4]. Researchers have highlighted the importance of identifying important factors which can contribute to 3-BL sustainability in metal recycling industry in order to achieve sustainable metal supply for future generations [5]. This current study investigates the most critical factors responsible for 3-BL sustainability in Indian metal recycling industry through application of Analytic Hierarchy Process (AHP) and Principal Component Analysis (PCA) methodology. 2. LITERATURE REVIEW 2.1. Triple Bottom Line (3-BL) Sustainability In order to address the new challenges on natural resources; both perishable as well as nonperishable, by growing industrialization, businesses needs to ensure that their activities eliminate waste and become sustainable [6], [7]. Sustainability or Sustainable Development (SD) has been proposed as a tool for decision-makers to repay their due towards society by taking into account the balanced economic, environmental, and societal growth. After a thorough study of the consumption pattern of natural resources over a period of 30 years, in their famous book, Limits to Growth”, researchers [8] warn that the world might collapse one day as it greatly relies on diminishing global resources and produces excessive emissions. The European Union (EU), considering the importance of sustainability, has urged member countries that current and future legislation must integrate sustainability into implementation orders. Many other countries are also introducing regulations that address sustainability issues [9]. In general terms, sustainability is expressed as a concept that enables the current generation meeting their requirements without harming the ability of future generations through thoughtful exploitation of resources [1] (WCED, 1987). Researchers suggest that by adopting simultaneous economic, ecological and societal factors of sustainability (3 BL) into business objectives, we can ensure future of next generations [3] [5] [10] [11]. While the economic aspect of the 3BL is widely understood and used in business and industry and its measurement criteria are also well defined, the environmental and social concepts are far less understood and practiced. However, most corporate are preparing to include societal and environmental objectives into their policies due to increased internal and external pressures [12]. 2.2. Closed-Loop Supply Chain and Triple Bottom Line Sustainability. Since the sub-processes of supply chain process involve the initial processing of raw material through final production and up to delivery of the product or services to potential customers, supply chain sustainability is considered as the single most important contributor towards business sustainability. Growing number of researchers and policy makers have started realizing that supply chain sustainability has the latent potential to realize 3-BL business objectives [13], [14]. Hence Sustainable Supply Chain Management (SSCM) has been defined as a strategic management function that incorporates societal and environmental objectives along with economic objectives for the successful long term successful performances of the organizations [15]. Successful SSCM demands the effective synchronization of available resources, business processes and stakeholders’ needs into organization’s vision for generating healthy returns on assets. Researchers and policy makers [16], [17] now agree that closing the loops are a prerequisite for supply chain sustainability whether it is measured in terms of economic, environment or societal context; and hence the concept of CLSC was introduced. CLSC operations are designed to recover efficiently the remaining values of the used products in the forward supply chains through reuse or recycling thus providing additional sustainability to supply chains [17] [18]. International Journal of Managing Value and Supply Chains (IJMVSC) Vol. 6, No. 3, September 2015 33 2.2.1. Achieving Triple Bottom Line Sustainability through Closing Metal Supply Loop Researchers have identified major wastes and concerns that emerge in the metal forward supply chains, including supply chain disruption and discontinuity, inadequate or inconsistent product quality, unpredictable delivery times and substantial, unanticipated additional costs, including premium freights [19], [20], and [21]. On the other hand, closing the loop activities are mainly concerned with waste reduction through efficient collection, recycling and integration of the recyclable material/ products back into the manufacturing stream. Metal recycling advances the necessary conditions for promoting triple bottom line sustainability by conserving perishable virgin metal resources, reducing mineral processing energy consumption, reducing landfill requirements, and protecting natural environment and creating job opportunities [22], [23] [24]. The same researchers highlight the need for identifying important CLSC factors and addressing them to exploit the maximum latent potential for achieving sustainability. 2.3. Review of Previous Research for Sustainability Studies in Supply chain. Research methodology is the research approach followed by the researcher from theoretical considerations to data collection and analysis [25], [26]. The intent of this section is twofold: First to review the literature on the sustainable supply chain operation and important factors identified and investigated previously by researchers. The second objective is to review research methodologies commonly used and also the less common methodologies with high potential research opportunities to investigate the triple bottom line sustainability in supply chains. Conceptual /theory and case studies have been the most common methodological approaches to sustainability and supply chain relations study till date [27] [28]. Some authors specifically developed concepts and proposed frameworks in order to classify and deal with strategic issues in SCM and sustainability [29], [30], and CLSC uncertainties and sustainability relation[31], [32], [33] and in development of algorithm for simultaneous study of economic and environmental impacts of CLSC operations [34]. The other commonly used method for sustainability research in supply chain is the case study methodology since this methodology is most suitable for understanding the issues in a new research domain [35]. Researchers have applied this methodology to study different aspects of CLSC across various industry sectors [36], [37], and [38]. Another less common method used by researchers is analytical models for studying the supply chain and sustainability relationships. Literature points out a lack of multicriteria decision making (MCDM) approaches for green logistics and supply chain sustainability study [39], [14]. However quantitative studies have been done using different approaches for the study of sustainability phenomena in supply chain. This includes use of optimization concepts [40], Analytic Hierarchy Process (AHP) [41] [42], Fuzzy decision making [43]. Another less applied but promising research method suggested in literature is mixed method research [44], [45]. In the present study, researchers applied multi-methods approach or method triangulation in his study. This entailed the use of a combination of research instruments that includes interviews schedules, questionnaires guides, non-participant observation, and secondary data analysis. 2.3.1 AHP Method The AHP is a multi-criteria decision making (MCDM) methodology, developed by T.L. Saaty [46] in 1980. Owing to its easy to understand and simple to apply methodology [47], [48], researchers have applied AHP methodology across a number of situations for studying sustainability and supply chain issues including sustainable supply chain technology evaluation and selection, performance measurement system development and prioritization of environmental factors [24], [47], [48], [49], [50]. International Journal of Managing Value and Supply Chains (IJMVSC) Vol. 6, No. 3, September 2015 34 This multi-criteria methodology disintegrates a complex decision-making issue into subproblems/ sub-sub problems, which are easy to understand and diagnose. The subjective choices of experts are assigned numerical values through ranking on a scale developed by Saaty [46]. The pair-wise comparisons of various criteria generated at previous step are organized into a square matrix. The diagonal elements of matrix are 1. In order to assign relative values to various factors under consideration, Principal Eigen values and maximum Eigen values (λ max ), are calculated. Consistencies in the diagnosis of the experts are verified through calculating Consistency Index (CI) and Consistency Ratio (CR) values, applying following steps [46] CI = (λ max −n) / (n−1) (1) Where, λ max is the maximum Eigen value of judgment matrix and n is number of evaluated criteria. Consistency Ratio (CR) is calculated as: CR = CI / RI (2) RI value is fixed for a given sample size. In general, a CR up to 0.10 is regarded to be acceptable value [24] , [46]. 2.3.2. Principal Component Analysis Method Principal component analysis (PCA), a multivariate methodology, is applied when researcher needs to contract numbers of observed variables into comparatively fewer factors contributing towards appreciable amount of data variability [51]. This methodology compiles changes in the observed data, to a group of unrelated factors each of which is an aggregate of original variables and are known as principal components (PC). Researchers have applied multivariate techniques for study of supply chain sustainability [52], [53] in electronic industry. 3. RESEARCH DESIGN Our research methodology is adapted from the work of Kim et al. [54], who studied most important factors for growth of retail sector supply chain in Korea by applying multi-criteria analysis technique. This study consisted of two surveys: First an AHP survey and then a general survey. The first survey was aimed at prioritizing and assigning quantitative values to the sustainability factors in CLSC operations identified from the literature and the field work. A conceptual CLSC sustainability model was developed. A general survey was undertaken to statistically check the validity and reliability of the previously prioritized factors. Principal Component Analysis (PCA) methodology was employed to extract the most common factors. The commonality of the factors in the two methods validates the conceptual model developed through AHP technique. Prior to large scale survey of CLSC practices and sustainability performances, a pilot study was first conducted to test the suitability of proposed survey instrument prior to sending it out. The procedural steps are explained in Fig. 1. International Journal of Managing Value and Supply Chains (IJMVSC) Vol. 6, No. 3, September 2015 3.1 . The AHP survey: Prioritizing and assigning ranks to critical factors 3.1.1. Data Collection A questionnaire was prepared to evaluate and prioritize the factors identified previously and were arranged in accordance with the matrix as suggested by Saaty developed based on four dimensions of operations management and two performances related to CLSC Operations, (3) People, and (4) Structure and Infrastructure, while performance dimensions included (5) Business performance and (6) 3 literature review. A total of 25 experts and they all agreed to respond. The respondents included academics (4 nos. ), and managers (15 sustainability fields. A limited number of people, possessing thorough knowledge of the subject matter are required for evaluating identified factors and it is not necessary to involve many respondents [55]. 3.1.2. Findings and discussions of AHP Survey The hierarchy levels of AHP survey factors are shown in Figure 2. packages (i.e., Expert Choice TM) was used for making computations. The local weights of all the main criteria and subcriteria were first calculated. The global priority vector was then calculated combining all successive hierarchical levels in each matrix, as per standard AHP method. International Journal of Managing Value and Supply Chains (IJMVSC) Vol. 6, No. 3, September 2015 Figure1. Research Design . The AHP survey: Prioritizing and assigning ranks to critical factors A questionnaire was prepared to evaluate and prioritize the factors identified previously and were arranged in accordance with the matrix as suggested by Saaty [46]. The questionnaire was four dimensions of operations management and two dimensions of related to CLSC . The management dimensions include: (1) Strategy, (2) Operations, (3) People, and (4) Structure and Infrastructure, while performance dimensions included (5) Business performance and (6) 3 -BL Sustainability perfor mances, as identified in the A total of 25 experts , having relevant experience were selected for AHP surv and they all agreed to respond. The respondents included top industry leaders (6 ), and managers (15 nos.) from metal recycling supply chains and A limited number of people, possessing thorough knowledge of the subject matter are required for evaluating identified factors and it is not necessary to involve many of AHP Survey The hierarchy levels of AHP survey factors are shown in Figure 2. The commercial software packages (i.e., Expert Choice TM) was used for making computations. The local weights of all criteria were first calculated. The global priority vector was then calculated combining all successive hierarchical levels in each matrix, as per standard AHP International Journal of Managing Value and Supply Chains (IJMVSC) Vol. 6, No. 3, September 2015 35 A questionnaire was prepared to evaluate and prioritize the factors identified previously and The questionnaire was dimensions of . The management dimensions include: (1) Strategy, (2) Operations, (3) People, and (4) Structure and Infrastructure, while performance dimensions mances, as identified in the were selected for AHP surv ey top industry leaders (6 nos.), from metal recycling supply chains and A limited number of people, possessing thorough knowledge of the subject matter are required for evaluating identified factors and it is not necessary to involve many The commercial software packages (i.e., Expert Choice TM) was used for making computations. The local weights of all criteria were first calculated. The global priority vector was then calculated combining all successive hierarchical levels in each matrix, as per standard AHP International Journal of Managing Value and Supply Chains (IJMVSC) Vol. 6, No. 3, September 2015 36 Figure 2 Key factors identified in AHP survey The evaluation began by determining the relative weight of the six dimensions of CLSC selected for the study. A matrix was prepared to explain and assign comparative importance to each factor, as suggested in literature [46], [47], [48] and results are summarized (Table 1). Table 1: Prioritization of 3BL dimensions Peopl e (P) Sustainability Performance (SP) Structure & Infra (SI) Business Performance (BP) Strategy (S) Operations (O) Eigenvect or P 1 1 2 0.25 0.25 0.5 .09 SP 1 1 2 0.2 0.2 0.5 .09 SI 0.5 0.5 1 0.2 0.5 0.5 .06 BP 4 5 5 1 0.5 4 .32 S 4 5 2 2 1 1 .26 O 2 2 2 0.25 1 1 .17 CR .09 The experts’ opinions suggest that Business Performance dimension is the highest priority followed by Strategic factors and Operational factors. People factor and 3-BL sustainability performances have been equally rated. Similarly, Table 2 to Table 7 show the factors under each dimension of People, Operations, Structure and Infrastructure, Strategy, Sustainability Performance and Business Performances, and their evaluation results by experts, respectively. Table 2: Prioritization of sub-factors under People Dimension HR1 HR2 HR3 HR4 Eigenvector Training and skill development 1 1 2 0.25 0.25 Involvement 1 1 2 0.2 0.2 Technology absorption capability 0.5 0.5 1 0.2 0.5 Documented procedures 4 5 5 1 0.5 CI 4 5 2 2 1 RI 2 2 2 0.25 1 CR .09 International Journal of Managing Value and Supply Chains (IJMVSC) Vol. 6, No. 3, September 2015 37 Table 3: Prioritization of sub-factors under Operations Dimension OM1 OM2 OM3 OM4 Eigenvector Lean and green manufacturing 1 1 2 0.25 0.25 Green purchasing 1 1 2 0.2 0.2 Reverse Logistics 0.5 0.5 1 0.2 0.5 Improving recycling capabilities 4 5 5 1 0.5 CI 4 5 2 2 1 RI 2 2 2 0.25 1 CR .09 Table 4: Prioritization of sub-factors under Structure and Infrastructure Dimension SI 1 SI 2 SI 3 SI 4 Eigenvector Organizational procedures and controls 1 .5 5 5 .33 Stakeholders' involvement 1 8 5 .52 Capability to implement societal projects 1 1 .07 safe work environment 1 .08 CI .02 RI .90 CR .02 Table 5: Prioritization of sub-factors under Strategy Dimension ST1 ST2 ST3 Eigenvector Integration of supply chain processes (ST1) 1 1 .5 .23 Supply chain collaborations (ST2) 1 .25 .19 Integration of marketing & supply chain processes (ST3) 1 .58 CI .02 RI .58 CR .03 Table 6: Prioritization of sub-factors under Strategy Dimension ROI MI VM ER HF EO Eigenvector ROI improvement (ROI) 1 1 2 .33 .4 1 .11 Margin Improvement( MI ) 1 2 .2 .2 .5 .08 virgin metal consumption (VM) 1 .2 .125 .5. .05 Reduction in environment release (ER) 1 .5 2 .28 Hazard free working environment (HF) 1 2 .36 Employment opportunities for local communities (EO) 1 .12 CI .025 RI 1.24 CR .02 International Journal of Managing Value and Supply Chains (IJMVSC) Vol. 6, No. 3, September 2015 38 Table 7: Prioritization of sub-factors under Business Performance Dimension SIM CR NCA PI NP D Eigenvect or Sales Improvement (SIM) 1 .2 .1 10 10 .3 Cost reduction (CR) 1 2 .5 2 .19 New customer addition (NCA) 1 1 .25 .19 Employees' productivity improvement (PI) 1 3 .17 New Product Development (NPD) 1 .14 CI .02 RI 1.12 CR .02 The local and global ratings of CLSC factors important for sustainable value creations are presented in Table 8. The calculations highlight that Integration of marketing and supply chain functions (0.151), Sales improvement (.097), Green Purchasing (.09) and Cost reduction (.061) are considered highest rating factors. The analysis also highlights that sale improvement is approximately as important as green purchasing while it is 1.59 times more important than cost reduction. Experts also considered ROI improvement as important as creating job opportunities for the local communities (.01 and .011 respectively). Table 8: Local and Global rating summary of CLSC sustainability factors International Journal of Managing Value and Supply Chains (IJMVSC) Vol. 6, No. 3, September 2015 The second survey aimed at investigating the CLSC operations management strategic and operational levels was raw material suppliers, and reverse logistics from the data-base of professional bodies INLZDA and ISRI). Raw material suppliers and linkages with India metal recycling industry; either through their offices in India or abroad. Content validation technique was app research issue under investigation. secondary metal manufacturing (such as raw material ingots, billets), units performing additional processing of the secondary metal manufactured in the first phase (such as wire drawing, rolling into channels and sections etc.), 15 raw material supplying into engineering and automotive components, who are part of automotive and construction industry closedloop supply chains. For the first survey, manufacturing companies, 134 valid responses returned as the recipients w ere completely filled (4%). The survey and professional software SPSS w are able to account for the observed relationship, was employed on the 42 CLSC operations and sustainability research stages [56]. Principal component factors . Orthogonal rotation method ( deriving factor loading due to its simplicity From the original 42 variables which were used in the questionnaires, only 16 were related to each other in order to form dimensions. All variables in this research, based on their mean scores are valued high by the participants. Verification on the majori and validity through statistical method is necessary to minimize the potential distortion in accordance with the propensity of the responder gauge the internal consistency of the instrument [57], [58]. Measurement validity the constructs it is designed to measure and Sphericity are performed to test suggest that factor analysis was appropriate satisfactory as suggested in literature 3.1.3. Findings and discussions of General Survey SPSS presents the Eigen values standard methodology through SPSS software. identified within the data set. Table International Journal of Managing Value and Supply Chains (IJMVSC) Vol. 6, No. 3, September 2015 investigating the CLSC operations management practices at was conducted on the secondary metal manufactur ing firms raw material suppliers, and reverse logistics (RL) providers. The respondent list was prepared professional bodies represent ing metal recyclers in India (ALUCA Raw material suppliers and RL providers were chosen based on their linkages with India metal recycling industry; either through their offices in India or abroad. technique was app lied to ensure that chosen factors adequately research issue under investigation. In this phase, the questionnaires were mailed to a total of 2 secondary metal manufacturing (such as raw material ingots, billets), units performing additional processing of the secondary metal manufactured in the first phase (such as wire drawing, rolling into channels and sections etc.), 15 raw material supplying firms, and 10 RL service providers, into engineering and automotive components, who are part of automotive and construction loop supply chains. For the first survey, out of 250 postal surveys to manufacturing companies, 134 valid responses (54%) were received,15 questionnaires were ere no longer at that address(i.e. 6%), 10 surveys were not The survey s were administered between January 2013 and October 2013 w as used for data analysis. In order to identify fewer factors that are able to account for the observed relationship, exploratory factor analysis (EFA) CLSC operations and sustainability practices identified in the previous Principal component s having Eigen values more than one were for deriving . Orthogonal rotation method ( Varimax) with Kaiser Normalization was deriving factor loading due to its simplicity [60]. Only absolute values over 0.4 were considered. From the original 42 variables which were used in the questionnaires, only 16 were related to each other in order to form dimensions. All variables in this research, based on their mean scores are valued high by the participants. Verification on the majori ty of survey results for reliability and validity through statistical method is necessary to minimize the potential distortion in accordance with the propensity of the responder [54]. Cronbach's alpha test was performed to the applied instrument and the resultant value of 0.71 validates Measurement validity indicates how closely an instrument measures the constructs it is designed to measure and Kaiser-Meyer-Olkin (KMO) and Bartlett Test of test the validity in PCA methodology. The test results suggest that factor analysis was appropriate for these data sets. Factor loadings values were satisfactory as suggested in literature [59]. Table 9: KMO and Bartlett’s test of General Survey SPSS presents the Eigen values related with each PC (factor) were determined by methodology through SPSS software. Before extraction, 16 linear components were identified within the data set. Table 10 depicts the list of factors with coding. International Journal of Managing Value and Supply Chains (IJMVSC) Vol. 6, No. 3, September 2015 39 practices at ing firms , their respondent list was prepared ing metal recyclers in India (ALUCA ST, providers were chosen based on their linkages with India metal recycling industry; either through their offices in India or abroad. address the In this phase, the questionnaires were mailed to a total of 2 25 secondary metal manufacturing (such as raw material ingots, billets), units performing additional processing of the secondary metal manufactured in the first phase (such as wire drawing, rolling firms, and 10 RL service providers, into engineering and automotive components, who are part of automotive and construction of 250 postal surveys to (54%) were received,15 questionnaires were no longer at that address(i.e. 6%), 10 surveys were not between January 2013 and October 2013 identify fewer factors that exploratory factor analysis (EFA) technique practices identified in the previous for deriving Kaiser Normalization was applied for over 0.4 were considered. From the original 42 variables which were used in the questionnaires, only 16 were related to each other in order to form dimensions. All variables in this research, based on their mean scores ty of survey results for reliability and validity through statistical method is necessary to minimize the potential distortion in test was performed to instrument and the resultant value of 0.71 validates indicates how closely an instrument measures Bartlett Test of results (table 9) Factor loadings values were were determined by applying Before extraction, 16 linear components were