Development of composite sustainable supply chain performance index for the automobile industry

Abstract

The automobile companies are becoming increasingly aware about the importance of sustainability and its challenges. In order to compete in the global market, the industry has responded to these challenges by identifying the sustainability indicators. It is quite difficult to evaluate the performance of the company on the basis of large number of sustainability indicators. Integration of key sustainability indicators is quite essential for effective decision-making. This article presents an integrated method by accommodating both qualitative and quantitative approaches for the development of composite sustainable supply chain (SC) performance index. We applied an integration of fuzzy analytical hierarchy process, Liberatore score and signal-to-noise ratio approaches to compute the index. The proposed methodology demonstrates how quantitative statistical inputs can be combined with expert opinion to construct an overall index of sustainability. The index can be extensively used by SC managers to measure their organization's contribution towards SC sustainability. An application of an integrated methodology is demonstrated for an automobile SC.

Keywords:

• sustainability
• performance measurement
• supply chain
• automobile

1. Introduction

The concept of sustainability is becoming mainstream (Corbett and Klassen 2006) in the present business world. Even though the roots of sustainability were found in 1713 in the agricultural sector (Halldorsson, Kotzab, and Skjøtt-Larsen 2009), its accentuation started only from 1990 onwards. The core objective of sustainable development (SD) is ‘a development that meets the needs of the present without compromising the ability of future generations to meet their own needs’ (WCED 1987). Researchers and practitioners comprehended this view in multi-indicators which encompass (triple bottom line) economic, social and environmental implications of business (Bai and Sarkis 2010). The accountability for sustainable operations extended beyond the organization and across a supply chain (SC) context. However, researchers focus on sustainable supply chain management (SSCM) is still in its infancy (Ageron, Gunasekaran, and Spalanzani 2012; Seuring and Müller 2008b; Gold, Seuring, and Beske 2010; Hassini, Surti, and Searcy 2012). The contemporary literature has mostly focused on measuring sustainability performance at product or functional level rather than analysis of entire SC (Yakovleva, Sarkis, and Sloan 2012; Koh et al. 2012).

Indices play a major role in assessing the sustainability performance. These indices compile several indicators. An indicator is a quantitative or a qualitative measure derived from a series of observed facts that can reveal relative positions (e.g. of a country) in a given area. It can be evaluated at regular intervals and it points out the direction of change across different units and through time (OECD 2008). Conventionally, sustainability performance has been evaluated by using three (economic, environmental and social) indicators. From the studies of Seuring and Müller (2008b), Carter and Easton (2011) and Ashby, Leat, and Smith (2012), it is concluded that most of the literature on sustainability performance in the SC area is inclined towards environmental issues and little focus is given on social indicator. Understanding the social sustainability is too complex and comprises some inherent challenges to measure (Missimer et al. 2010).

Though there are notable frameworks and indices available in the field of sustainable SC performance measurement (Chaabane, Ramudhin, and Paquet 2011; Hadiguna, Jaffar, and Mohamad 2011; Seuring and Müller 2008a; Carter and Rogers 2008), while assessing the sustainability performance integration of important indicators are often missing in these tools (for example social, economic and environmental indicators all together) (Byggeth and Hochschorner 2006; Hassini, Surti, and Searcy 2012). It depicts that there is still considerable room for tools and methodologies to effectively measure sustainability performance of SC operations in an integrated approach. With the view of this, authors developed a composite index to measure sustainability performance of a SC through integrated methodology which considers fuzzy analytical hierarchy process (FAHP), Liberatore score method and signal to noise (S/N) ratio. The application of the proposed methodology is demonstrated for an automobile SC.

It is also found that the growing importance in this area confirms the need for research such as that reported in this article, based on the evidences posited and research gaps identified in the Appendix A. The aim of this article is as follows:

• Identify the measures associated with the sustainability indicators such as economic, environmental, social, technical and political in SC.

• Demonstrate the application of proposed framework and provide the managerial insights into the domain of SC sustainability.

The remainder of the article is organized into five sections. Section 2 is a review of literature on sustainable SC performance measurement. Section 3 is a description of the methodology. Section 4 is on the application of the developed methodology for evaluating sustainability performance of an automobile SC. Section 5 discusses the findings and conclusion, Finally, Section 6 concludes with limitations and thoughts on future research scope.

2. Sustainable SC performance measurement

In recent years, a number of firms realized the potential of SSCM. Initial foundation was laid down by reporting sustainability issues by the World Business Council for Sustainable Development (Schmidheiny, Chase, and Simone 1997), the Global Reporting Initiative (GRI 2002a, 2002b) and development of standards (OECD 2002); subsequently researchers have developed various frameworks and indices focusing on the individual and integrated levels of sustainability indicators. In spite of the increasing awareness of the business case for sustainability and the growing knowledge of how to integrate sustainability into business, the majority of companies have not yet moved on to implementation (Hallstedt et al. 2010), since they often lack the insight for the effective measurement of sustainability performance. Our review of the literature suggests that many of the researchers considered triple bottom line (economic, environmental and social) approach to assess the sustainability performance in SC.

Chaabane, Ramudhin, and Paquet (2011) report that the objective of economic sustainability is to minimize the total logistic cost or maximize the profit of different SC activities throughout all the product life cycle stages – purchasing, production, warehousing, distribution and recycling. Similarly, the objective of environmental and social sustainability is to reduce the environmental damages and increase the quality of life of the communities in which SC operates. The overall goal is to consider these indicators simultaneously and balance these from a microeconomic standpoint (Carter and Rogers 2008). In addition to this, the authors considered additional technical and political indicators as separate indicators from social to contemplate the case presented in Section 4 and operationalize the sustainability concept. In addition, these two indicators also play a significant role in geographically in SC operations at both intra- and inter-organizational levels in the different countries. For example information technology (IT) plays a significant role in achieving many sustainability objectives of social, economical and environmental such as managing inventories and transparency in SC (Piotrowicz and Cuthbertson 2009). IT is perceived as having both negative and positive influences on the environment (Piotrowicz and Cuthbertson 2009) and businesses. Therefore, this study has given due considerations to various technology-specific measures such as information security, IT implementation and adoption of new technologies. Political indicator plays a significant role in implementing sustainability practices. Since many SC stages operate in different geographical locations, factors such as government regulations (Seuring and Müller 2008a, Ageron, Gunasekaran, and Spalanzani 2012), carbon tax structures (Gupta and Desai 2011) and political stability influence the SCs which they operate (Bhatnagar and Sohal 2005). The objective of political indicator is to promote and support the adoption of sustainable practices in SC.

From Appendix A, it was found that the majority of article on sustainable SC literature have discussed economic and environmental measures with little focus on social measures in SC study. In addition to this, Hassini, Surti, and Searcy (2012) in their study mentioned an exhaustive list of sustainable measures, and it is reported that there is no consistency in the number of measures used to assess sustainability. Hence, the selection of sub-indicators under each indicator is a challenging task and even more difficult to measure (Cohen and Roussel 2004). Keeping in mind these challenges and taking into account the theory of performance measures design (Bendoly, Rosenzweig, and Stratman 2007; Gopal and Thakkar 2012; Hassini, Surti, and Searcy 2012), the authors proposed a methodology to develop composite index considering the five indicators (economic, environmental, social, technical and political), to assess the organization's contributions towards SC sustainability, since composite indices or indicators play a significant role in assessing the sustainability performance in multi-attribute view. These indicators exhibit the competence to summarize, focus and simplify the enormous complexity of our dynamic environment (Singh et al. 2009).

Hence, the indicators and composite indicators are increasingly recognized as a useful tool for policy-making and public communication in conveying the information on countries and corporate performance in fields such as environment, economy, society or technological improvement (Singh et al. 2009). Despite having many sustainability indices at organizational level, still no comprehensive framework exists for measuring the sustainability of SC. The focus on the development of indices for measuring sustainability performance of SC is lacking in the literature. Section 3 explains the methodology to develop composite index.

3. Methodology

This research employs an integrated use of FAHP, Liberatore score and S/N to compute the sustainability index. The schematic representation of proposed methodology is presented in Figure 1.

Figure 1 Proposed methodology.

This methodology consists of five major stages:

	identification of sustainability indicators and sub-indicators (Section 3.1);
	computation of weights using FAHP (Section 3.2);
	computation of qualitative index using Liberatore score method (Section 3.3);
	computation of quantitative index using S/N ratio method (Section 3.4);
	computation of composite sustainable supply chain performance index (CSSCPI) (Section 3.5).

3.1 Identification of sustainability indicators and sub-indicators

The first stage of the framework is to identify a range of indicators and sub-indicators to be considered in the sustainability assessment. These can be obtained either from the literature or by interaction with the practitioners.

3.2 Computation of weights

After finalizing the indicators and sub-indicators, the next step is obtaining the weights of these indicators. The literature reports many methods for the computation of weights such as principle component analysis, equal weights, data envelopment analysis, conjoint analysis and analytical hierarchy process (AHP). However, among these methods, AHP has gained wider acceptability since it is designed to incorporate tangible as well as non-tangible factors, especially where the subjective judgements of different individuals constitute an important part of decision-making (Saaty 1980). Applications of AHP can be seen in a wide range of areas such as project management, environmental impact assessment (Ramanathan 2001) and performance measurement system (Suwignjo, Bititci, and Carrie 2000). In operations management field, the most addressed decision themes by using AHP technique are product and process design, and managing the SC (Subramanian and Ramanathan 2012).

Initial step in AHP is the formation of hierarchy with the selected indicators and sub-indicators to compute the weights. It can be done by breaking down the problem into a hierarchy of decision elements. Computation of weights for the indicators and sub-indicators can be done by pairwise comparison. This indicates the strength with which one indicator dominates another for the indicator under which they are compared. The reason lies in that the performance of a system is a result of the interaction of various indicators, but every indicator plays its own role and contributes to the system as a whole (Bahinipati, Kanda, and Deshmukh 2009).

However, in AHP the problem of dealing with subjective assessments remains largely unsolved (Wang, Luo, and Hua 2008). Particularly, when we measure sustainability performance it is difficult to judge in some instances. Sustainability is, indeed, quite subjective in nature because what appears unsustainable for an environmentalist may be sustainable for an economist and measurements signifying sustainability may differ for these specialists (Phillis and Andriantiatsaholiniaina 2001). Therefore, decision-makers may be certain about the preferences for alternatives, but may not be sure about precise numerical values of judgements (Deschrijver and Kerre 2003). In addition to this, the conventional AHP still cannot reflect the human thinking style (Kahraman, Cebeci, and Ruan 2004).

This has led to the inclusion of theories, such as fuzzy and grey, which are capable of dealing with subjective assessments (Dweiri and Kablan 2006; Gu and Zhu 2006). However, Saaty (2006) mentioned that the procedure of AHP is already fuzzy; hence, it is not necessary to incorporate fuzzy logic in conventional AHP. However, Kordi and Brandit (2012) stated that there are valid reasons to fuzzify it further. One reason for this could be the need for greater flexibilities in decision-makers' judgements when ranking the criteria over each other and second, using fuzzy concept decision-makers are more flexible in expressing their judgements by applying different degrees of fuzzification or uncertainty in fuzzy ratios. In addition, in recent literature, FAHP has proven to be a very useful tool and has quickly gained acceptance among a vast array of researchers (for example, Celik, Er, and Ozok 2009; Samvedi, Jain, and Chan 2012a, 2012b). Hence, to overcome this problem, the authors used FAHP in lines of the other sustainable quantification studies.

The fundamental concept of FAHP is based on fuzzy set theory which was introduced by Zaddeh (1965). Fuzzy set is a class of objects with a continuum of grades of membership. Such a set is characterized by a membership (characteristic) function, which assigns to each object a grade of membership ranging between zero and one. The decision-maker can specify preferences in the form of natural language expressions about the importance of each indicators and sub-indicator with respect to other; for this, linguistic scale is considered. This scale is ‘absolutely more important’, ‘very strongly more important’, ‘strongly more important’, ‘weakly more important’ and ‘equally important’. Hence, in the FAHP procedure, the pairwise comparisons in the judgement matrix are fuzzy numbers.

Initially, Van Laarhoven and Pedrycz (1983) introduced fuzzy logic into hierarchical problems. Subsequently, researchers such as Buckley (1985), Boender, Grann, and Lootsma (1989), Chang (1996) and Cheng (1996) derived different methods to solve the FAHP problem. However, the most widely used method in the literature is Chang's extent analysis, since computational requirement is very low and it does not involve additional operations (Buyukozkan, Kahraman, and Ruan 2004). Hence, for this study, the authors used Chang's extent analysis to solve the FAHP problem. For obtaining linguistic values for the linguistic variables, authors used triangular membership function since the triangular membership functions are simple (Phillis and Andriantiatsaholiniaina 2001) and in addition to this, Chang's extent analysis method allows only triangle fuzzy numbers (Buyukozkan, Kahraman, and Ruan 2004).

After collecting pairwise comparison judgements from all three experts, the judgements are translated into the corresponding pairwise comparison judgement matrices (PCJMs). After forming PCJMs of three experts, the authors converted linguistic opinions into fuzzy values by using triangular membership function. As suggested by Saaty (1980), the geometric mean approach, instead of the arithmetic approach, is used to combine the individual PCJMs to obtain the consensus PCJMs for the experts.

Similarly, PCJMs are computed for the sub-indicators. After obtaining the judgement matrices, weights are calculated using Chang's extent analysis algorithm.

3.3 Computation of qualitative index

The purpose of computing this index is to measure the total sustainability performance in a qualitative approach. For this, Liberatore score method is considered. The method is selected for its ability to evaluate each indicator based on previous performance on a linguistic rating scale. Initially, the use of this method was reported by Liberator's (1987). Subsequently, it has been used in the literature for various applications (Liberatore, Nydick, and Sanchej 1992; Singh et al. 2007). In this method, the rating scale is used to measure the performance of each indicator instead of comparison. The objective of the method is to calculate Liberatore scores by considering the rating values of each indicator for a particular year which is given by the evaluator in linguistic scale of outstanding (O), good (G), average (A), fair (F) and poor (P). The priority weights of these linguistic variables (scale) can be determined using pairwise comparisons. The relative importance between two adjacent linguistic variables is constant at two times. Finally, after obtaining the pairwise comparison matrix of rating scale, we can compute weights by eigenvalue problem procedure. The final weights of the rating scales outstanding (O), good (G), average (A), fair (F) and poor (P) are 0.513, 0.261, 0.129, 0.063 and 0.034. Now, we can compute the Liberatore score of each sub-indicator by multiplying the global weights of the indicators with the rating values of each sub-indicator. Finally, the obtained scores are normalized on a 10-point scale. The equations used to obtain qualitative index are listed as follows:

(1)

where

	Liberatore score (LS) = LRV × GW,
	maximum score (LMAX) = 0.513 × GW,
	local weight (LW) = weights obtained by FAHP,
	global weight (GW) = LW × respective indicator weight,
	overall rating (LR) = rating given by the expert,
	Liberatore rating value (LRV) = obtained by pairwise comparison of rating scale.

Similarly, sub-indices for remaining the indicators can be computed.

3.4 Computation of quantitative index

For computing quantitative index, authors used S/N ratio method. The concept of S/N ratio was initially reported in quality engineering area by Genichi Taguchi. The objective of this method is to measure the functionality of the system by considering the performance characteristics of measures (Taguchi and Jugulum 2002). There are several types of S/N ratios available based on performance characteristics. However, in this study, the authors considered ‘smaller the better’ and ‘larger the better’ S/N ratios, since the performance indicators are classified into either maximization or minimization. For example, performance objective of ‘cost associated with environmental compliance’ indicator is maximized, since the assumption of this indicator is that more a company spends in trying to achieve compliance, the more likely it will yield an effective sustainable SC (Olugu, Wong, and Shaharoun 2011). Similarly, other indicators are classified based on the characteristic of indicators.

Once the classification of indicators has been decided, the next step is to compute the quantitative index. The procedure to compute the quantitative index is as follows:

• Collect the indicators' performance values for a particular year (quarterly or half yearly). Whenever, quantitative assessment is not possible, assess the indicators qualitatively based on 1–9 scale.

• Calculate S/N ratios by using the following equation. (These equations are adopted from Taguchi quality engineering literature.)

  For smaller the better

  (2)

  For larger the better

  (3)
  where y is the performance characteristic (sub-indicator values) and n is the number of data points.

• Perform the normalization of the data by using the following equation.

  (4)
  where UB represents upper bound and LB represents lower bound.

  By using the above-mentioned equation all S/N ratio values (by taking smaller the better and larger the better values individually) can be converted into 1–10 scale.

• Calculate weighted normalized S/N ratio values (global weight of each indicator multiplied by S/N ratio values of corresponding indicator).

• After obtaining the weighted normalized values, compute the average of weighted S/N ratio values.

Similarly, sub-indices for remaining indicators can be computed by using the above-mentioned procedure.

3.5 Computation of CSSCPI

To obtain the CSSCPI value of each indicator, multiply the sub-index values obtained from the qualitative and quantitative methods of each indicator. The final composite index can be computed by using life cycle assessment (LCA) polygon technique (radar chart) (Singh et al. 2007), which describes impact categories in a radial system of axis. By using CSSCPI values of each indicator, a polygon can be constructed. The area of the polygon gives the final CSSCPI value of SC. The area of the polygon is calculated dividing the total area of the polygon into triangles. Then, using the formula (1.2^*a^*b^*sin(360/5)), the area of each triangle is computed. The following assumptions are made for computing the area of the polygon. (1) The number of fields for each data-set is the same (denoted by n). Here, n = 1. (2) On the radar chart, two adjacent points are connected by a straight line. The larger the area better is the sustainability performance of the SC.

4. Case study

The automotive sector has received a lot of attention in the context of sustainability from both regulators and consumers. The maturity stage of automobile industry, intensification of competition and rapid changes in climate have led to considerable pressure on automobile manufacturers to consider sustainable issues. Most importantly, processes such as designing, sourcing, producing and distributing products in the global markets play a central role, as these activities account for a bulk of the resources consumed and the environmental impact (Gupta and Desai 2011). From the literature, it is evident that various issues on sustainability performance of automobile SC have been addressed such as LCA, carbon management, green practices and energy management (Orsato and Wells 2007; Zhu and Sarkis 2007; Lee 2011; Mayyas et al. 2012; Kagawa et al. 2013). These studies have addressed a specific issue of sustainability performance rather than assessing SC performance in an integrated approach.

To illustrate the developed methodology, the authors selected an automobile company located in the southern part of India. Indian automobile industry is complex and diverse in structure (Jharkharia and Shankar 2006) and different from the automotive industry in the West (Saad and Patel 2006) in the adoption of SC practices. This implies that the SC should be custom designed for places such as India, and industries should not depend on the readily available SC solutions which will enforce changes in the structural, logical and informational blocks (Kulakarni and Ramdasi 2004). The proposed method is applied to the case company's SC to assess the sustainability performance of SC for the year 2012.

Step 1: Identification of sustainability indicators and sub-indicators

As we discussed in Section 2, five sustainability indicators such as economic, environmental, social, technical and political are selected. Sub-indicators for under each indicator are selected from the literature and by interaction with a team of three experts with more than 20 years of experience in handling SC activities from case company. It was found that several indicators, which focused on SC issues, were incorporated into the company's overall system of sustainability. However, during the process of selection of sub-indicators some of the sub-indicators are eliminated. This results in 42 sub-indicators which are directly addressed SC issues related to upstream, midstream and downstream. These are listed in Table 1.

Table 1 List of sustainability indicators and sub-indicators.

Economic (EC)	Environmental (EN)		Social (SO)	Technological (TE)	Political (PO)
EC1 – Cost associated with environmental compliance (million rupees)	EN1 – Total waste (ton)	EN10 – Availability of standard operating procedure for the collection of ELV (scale)	SO1 – Green product warranty (scale)	TE1 – IT implementation (%)	PO1 – Government regulations on environmental aspects (scale)
EC2 – Total SC cost (Crores).	EN2 – Utility utilization (GJ)	EN11 – Availability of collection centres (scale)	SO2 – Customer interest in green products (scale)	TE2 – Adoption of new technologies (scale)	PO2 – Political stability (scale)
EC3 – Cost associated with energy consumption (crores)	EN3 – Implementation of environmental regulations (scale)	EN12 – Reduction of air emissions (scale)	SO3 – Customer satisfaction from green products (scale)	TE3 – Upgradation of existing technologies (scale)	PO3 – Tax structure (scale)
EC4 – Improvement in sales revenue (scale)	EN4 – Recycled material in products (kg/vehicle)	EN13 – Supplier commitment on overall environmental aspects (scale)	SO4 – Customer involvement in returning the ELVs (scale)	TE4 – Level of Information security (scale)
EC5 – Cost associated with processing of recyclables (scale)	EN5 – Product to be disposed to landfill or incinerated (%)	EN14 – Improvement of resource efficiency (%)	SO5 – Conduction of training and development activities(scale)
EC6 – Cost of disposal for hazardous and waste (lakhs)	EN6 – Biodegradable content in products (%)	EN15 – Percentage of suppliers having ISO certification (%)	SO6 – Employee well-being (occupational health and safety) (scale)
	EN7 – Market share controlled by green products (%)	EN16 – Level of management effort to motivate suppliers (scale)	SO7 – Child labour involvement (scale)
	EN8 – Availability of environmental evaluation schemes (scale)	EN17 – Availability of environmental reward systems (scale)	SO8 – Involvement in charitable activities (scale)
	EN9 – Environmental management initiatives (scale)	EN18 – Percentage decrease in scrap and rework (%)	SO9 – Management commitment on stakeholder welfare (scale)
			SO10 – Corruption (scale)
			SO11 – Disclosure of environmental initiative to the public (scale)

Source: Gunasekaran, Patel, and Tirtiroglu (2001), Schultmann, Jochum, and Rentz (2001), Zhu and Sarkis (2004), Hervani, Helms, and Sarkis (2005), Clemens (2006), Sarkis (2006), Searcy, McCartney, and Karapetrovic (2007), Markley and Davis (2007), Vachon and Klassen (2008), Vachon and Mao (2008), Olugu, Wong, and Shaharoun (2011), Fabbe-Costes and Jahre (2008), Paulraj (2011), Ageron et al. (2012), Wittstruck and Teuteberg (2011).

Step 2: Computation of weights

To obtain the weight of each indicator and sub-indicator, a questionnaire posited to three experts, to provide pairwise comparisons of indicators and sub-indicators used in the AHP hierarchy. For this, a nine-point scale is used as suggested by Saaty (1980). These responses are converted into fuzzy scale.

The final PCJM are computed by taking the geometric mean of three experts. Table 2 indicates the PCJM of indicators on fuzzy scale.

Table 2 Final PCJM matrix for the indicators.

DMNS	EC			EN			SO			TE			PO
EC	1.00	1.00	1.00	0.74	0.71	0.87	1.98	2.51	3.04	2.10	2.62	3.12	0.18	0.79	1.14
EN	1.14	1.40	1.35	1.00	1.00	1.00	1.18	1.44	1.73	1.77	2.62	2.79	1.35	0.81	2.35
SO	0.32	0.69	0.85	0.58	0.69	0.85	1.00	1.00	1.00	1.32	1.58	1.88	1.77	2.28	2.79
TE	0.31	0.38	0.47	0.36	0.38	0.56	0.53	0.63	0.76	1.00	1.00	1.00	1.50	2.00	2.50
PO	0.87	1.26	5.55	0.42	1.23	0.74	0.36	0.44	0.56	0.40	0.50	0.66	1.00	1.00	1.00

Similarly, PCJMs are computed for the sub-indicators. After obtaining the judgement matrices, the next step is to calculate the weights. For this, Chang's (1996) extent analysis method is used. The mistakes pointed out in this method by Wang, Luo, and Hua 2008 and Zhu, Shang, and Yang 2012 are valid. Adequate care has been taken to avoid mistakes committed by Chang in his illustrative example. Throughout our calculation, the elements in main diagonal were kept as (1, 1, 1) and the values below the main diagonal are reciprocal of the transposes in the upper half of the matrix. Table 3 shows final weights of indicators and sub-indicators after applying Chang's extent analysis.

Table 3 Indicators and sub-indicator weights.

Economic (0.269)		Environmental (0.258)		Social (0.196)		Technical (0.099)		Political/legal (0.177)
EC1	0.198	EN1	0.026	SO1	0.038	TE1	0.247	PO1	0.587
EC2	0.219	EN2	0.026	SO2	0.031	TE2	0.153	PO2	0.154
EC3	0.237	EN3	0.053	SO3	0.077	TE3	0.224	PO3	0.258
EC4	0.078	EN4	0.053	SO4	0.077	TE4	0.374
EC5	0.078	EN5	0.026	SO5	0.051
EC6	0.043	EN6	0.105	SO6	0.111
		EN7	0.079	SO7	0.155
		EN8	0.026	SO8	0.129
		EN9	0.053	SO9	0.136
		EN10	0.026	SO10	0.111
		EN11	0.026	SO11	0.077
		EN12	0.079
		EN13	0.105
		EN14	0.079
		EN15	0.026
		EN16	0.053
		EN17	0.053
		EN18	0.105

Step 3: Computation of qualitative index (Liberatore score method)

To obtain the qualitative index for each index initially, overall performance ratings are obtained. For this, performance rating values are given by the expert in linguistic scale of outstanding (O), good (G), average (A), fair (F) and poor (P) for the year 2012. The final qualitative sub-index for the each indicator is computed by using Equation (1). Table 4 explains the computation of qualitative sub-index for the economic indicator.

Table 4 Calculation table for sub-index of economic dimension based on Liberatore method.

IndicatorEconomic weight (0.269)	Local weights (LW)	Global weights (GW)	Overall rating (LR)	Liberatore rating value (LRV)	Liberatore score (LS)	Maximum score (LMAX)
Sub-indicator
Cost associated with environmental compliance (EC1) (rupees)	0.198	0.053	G	0.261	0.013	0.027
Total SC cost (EC2) (rupees)	0.219	0.058	G	0.261	0.015	0.029
Cost associated with energy consumption (EC3) (rupees)	0.237	0.063	G	0.261	0.016	0.032
Green cost per revenue (EC4) (rupees)	0.078	0.020	F	0.063	0.001	0.010
Cost associated with processing of recyclables (EC5) (rupees)	0.078	0.020	G	0.261	0.005	0.010
Cost of disposal for hazardous and unprocessed waste (EC6) (rupees)	0.043	0.011	A	0.129	0.001	0.005
					Sum = 0.047	Sum = 0.113

Note: Qualitative index from L-score method is 10^*(∑LS/∑LMAX) = 4.159.

The qualitative sub-indices are economic (4.159), environmental (4.783), social (3.565), technological (7.916) and political (4.318) (). These computation tables are listed in Appendix for readers' reference.

Step 4: Computation of quantitative index (S/N ratio method)

Initially, all the sub-indicators are classified based on their performance characteristic such as smaller the better and larger the better. The next step is to compute these values by using Equations (2) and (3). Further normalization of the vales can be obtained by using Equation (4). Normalized S/N ratio values are computed as follows:

	For Larger the better S/N ratio values
	41.437, 16.902
	Equation for 41.437 is
	For 16.902 is

Similarly, the smaller the better values are computed. Table 5 shows the computation of quantitative index for economic indicator.

Table 5 Calculation table for sub-index of economic indicator based on S/N ratio method.

		Indicatiors values
	Unit of measurement	Qu1	Qu2	Qu3	Characteristic of measure	S/N ratio	Normalized values	Weighted normalized values
[Sub-indicatiors]
Cost associated with environmental compliance (EC1)	Million rupees	115.63	110.00	131.26	Larger the better	41.437	10.00	0.53
Total SC cost (EC2)	Crores	14,535	15,000	14,000	Smaller the better	− 83.237	1.00	0.058
Cost associated with energy consumption (%) (EC3)	Crores	284.8	250.5	301.25	Smaller the better	− 48.932	5.42	0.341
Green cost per revenue (EC4)1	Scale	7	7	7	Larger the better	16.902	1.00	0.020
Cost associated with processing of recyclables (EC5)1	Scale	5	4	5	Smaller the better	− 13.424	10.00	0.2
Cost of disposal for hazardous and unprocessed waste (EC6)	Lakhs	50	50	50	Smaller the better	− 33.97	7.35	0.08
								1.229

Note: Quantitative index from S/N ratio method = (1.229/6) = 0.204.

The quantitative sub-indices are economic (0.204), environmental (0.081), social (0.135), technological (0.120) and political (0.139). These computation tables are listed in Appendix for readers' reference.

Step 5: Computation of CSSCPI

After obtaining the qualitative and quantitative sub-index values of each indicator, we should multiply these values to obtain final sub-index values of each indicator. These values are listed in Table 6. A graph (Figure 2) is drawn with the final sub-index values to analyse the overall sustainability performance of case company SC for the year 2012. The sustainability evaluation is carried out based on the evaluation of area of the polygon. The point where the axes meet corresponds to a value of 0. The value corresponding to the edges of the circle is normalized maxima with a value of 1. The larger the area better is the sustainability performance of the company. The area of the polygon is calculated dividing the total area of the polygon into triangles. Then, using the formula (1.2^*a^*b^*sin(360/5)) area of each triangle is computed. Polygon area is 0.9749.

Table 6 Sub-indices of each indicators.

Indicators of sustainability	Sub-index based on Liberatore score method (SILS)	Sub-index based on S/N ratio method (SIS/N)	Final indices (SILS × SIS/N)
Economic (EC)	4.159	0.204	0.848
Environmental (EN)	4.783	0.081	0.387
Social (SO)	3.565	0.135	0.481
Technological (TE)	7.916	0.120	0.949
Political/legal (PO)	4.318	0.139	0.6

Figure 2 SSCP of a case company.

Therefore, the CSSCPI value of case company is 0.9749.

For benchmarking the sustainability performance of different years, case company can evaluate CSSCPI values year-wise. If a graph is drawn for CSSCPI with respect to year, the slope of the line indicates the incremental growth/decline in the sustainability performance of SC.

5. Discussion and conclusion

As suggested by Yakovleva, Sarkis, and Sloan (2012), one effective way for managers and policy-makers to improve the sustainability of SCs is to complete a benchmarking exercise to determine how well specific SCs perform. In response to this, this study has focused on how to measure and benchmark SC sustainability of an organization. The conventional evaluation approaches available at organizational level are inappropriate at SC level, since SCs are characterized by complexity and uncertainty. To compensate this, an integrated qualitative and quantitative sustainable performance index is proposed using the concepts of FAHP, S/N ratio and Liberatore score method with respect to automobile SC.

For exhibiting the model results, the authors considered case company expert opinions while development of weights for indicators and sub-indicators and the Liberatore score method for the qualitative evaluation of sustainability performance. Furthermore, for computing S/N ratio, data are collected from the case company.

Selection of indicators and sub-indicators and their weights form a basis for prioritizing the sustainability issues in SCs. Based on the weights, targets are set and action plans are made for each indicator for achieving SD in the SC. The proposed methodology intends to support the SC mangers in the assessment and comparison of the sustainability measures. This can be used as a basis to convince the top management for appropriating the necessary actions to improve sustainability practices in use. In the era of globalization, organizations are looking for a sustainable competitive advantage. This has forced the majority of the organizations to incorporate sustainability practices in their chains. This brings the benefits for organizations by reducing both indigenous and global SC risks and hence any kind of disruption to the SC. The proposed approach not only sensitizes the organizations about their SC sustainability performance but also directs them to adopt an appropriate set of SC risk mitigation measures.

This study has the following novel features:

• It has addressed the ambiguity (in expert opinions) while deciding the weights of each indicator.

• Characteristics of sub-indicators are considered during the computation of quantitative index.

• For computing the quantitative sub-indices in the literature, researchers have used the Z-score method; however, the Z-score method has some of the following limitations (Wiesen 2006): (i) loss of meaningfulness of raw scores after normalization, (ii) loss of meaning of standard deviations and (iii) magnification of small differences. In this article, the authors have taken care of some of these issues by using the S/N ratio method (see Table 5).

• Comparing the case company's different years of CSSCPI is simple, since the maximum value of CSSCPI of the respective year gives the best performance. If there is any decrease in the sustainability performance, the SC managers can analyse the causes.

6. Limitations and future scope of work

Although certain sub-indicators of the data unavailable, with the existing information it can be concluded that in Indian context few sustainability practices such as collection of end of life vehicles (ELVs) and green product warranty are at very infancy stage. Political factors cause a significant disruption to the automobile SC. This has been reflected even in the decision of Tata to move the small car Nano plant from Singur, Hooghly district, West Bengal to Sanand, Gujarat in 2008. Hence, this article included political factor as an indicator of sustainability.

There are many region- and country-specific indicators/sub-indicators which can influence the value of CSSCPI. The proposed approach to evaluate the sustainability of the SC demands a consensus among the potential users and stakeholders of the framework. The rating of sustainability indicators on a 1–9 scale requires detailed expert knowledge on operations and impacts of specific SCs. To some extent, the experts with varied experience may induce a bias, and their opinions may affect the final scores. In order to protect the results from bias, companies and policy-makers need to incorporate stakeholders and experts from other stakeholder groups (not just industry) in determining the relative weights of indicators. For the computation of quantitative index, the authors considered the characteristics of only ‘smaller the better’ and ‘larger the better’ indicators. Furthermore, the characteristics of ‘target the best’ indicators can be incorporated. The sub-indicators that limit the sustainability growth need an evolutionary study and understanding.

However, the computed value of CSSCPI should not be generalized. In order to make the application of the proposed methodology more holistic, there is a need to understand the extensive empirical studies for determining the key indicators/sub-indicators and their inter-relationships for the various industry segments. In addition, we recommend a broader application of the proposed methodology, case studies and development of better data acquisition systems in the future to overcome some of the limitations of this study. However, the proposed methodology, along with the case example, provides a strong foundation upon which to build.

To obtain the weights of the indicators and sub-indicators instead of FAHP, other multi-criteria decision-making techniques such as ANP, multiple-methodological linkages to optimization tools such as goal programming, and a ‘what–if’ analysis can be deployed. The sustainability indicators considered in the analysis are dynamic in nature and their individual value may change over time. This shift becomes more evident as SC managers are seeking ways to reduce environmental burdens at the cost of social and economic considerations. An integrated approach to evaluate SC sustainability on five indicators such as the approach proposed here can help to more effectively evaluate these trades-offs.

Acknowledgements

I thank Mr T. Srinivasa Babu and Mr S.V.N. Prasad for their immense support during data collection.

Notes

^1. Due to non-availability of the data some of the sub-indicator values are taken from expert opinions to compute quantitative index (S/N ratio method) instead of the case company's published reports.

Appendix A. Researchers focus on sustainable SC indicators

Appendix A Researchers focus on sustainable SC indicators

Author	Economical	Environmental	Social
Clift (2003)	√	√	√
Markley and Davis (2007)			√
Linton, Klassen, Jayaraman (2007)		√
Hutchins and Sutherland (2008)			√
Beamon (2008)		√
Vachon and Mao (2008)		√	√
Seuring and Müller (2008a)	√	√	√
Seuring and Müller (2008b)	√	√	√
Halldorsson, Kotzab, and Skjøtt-Larsen (2009)	√	√	√
Vermeulen and Seuring (2009)	√	√	√
Pagell and Wu (2009)		√	√
Gold, Seuring, and Beske (2010)	√	√	√
Mann et al. (2010)	√	√
Hadiguna, Jaffar, and Mohamad (2011)	√	√
Carter and Easton (2011)	√	√	√
Gupta and Desai (2011)		√
Hassini, Surti, and Searcy (2012)	√	√	√
Ageron, Gunasekaran, and Spalanzani (2012)	√	√	√
Wittstruck and Teuteberg (2011)		√	√
Ashby, Leat, and Smith (2012)	√	√	√
Bai et al. (2012)		√
Hall et al. (2012)	√	√	√
Yakovleva, Sarkis, and Sloan (2012)		√
Koh et al. (2012)		√

Appendix B. Calculation sheets of CSSCPI for economic, environmental, social, technical and political indicators

Table

CSSCPI based on Liberatore method							CSSCPI based on S/N ratio method
[Indicator]Economic weight (0.269)	Local weight (LW)	Global weight (GW)	Overall rating	Liberatore rating value	Liberatore score (LS)	Maximum score LMAX	Unit of measurement	Indicator values			Characteristic of measure	S/N ratio	Normalized values	Weighted normalized values
								Qu1	Qu2	Qu3
[Sub-indicators]
Cost associated with environmental compliance (EC1)	0.198	0.053	G	0.261	0.013	0.027	Million Rupees	115.63	110.00	131.26	Larger the better	41.437	10.00	0.53
Total SC cost (EC2)	0.219	0.058	G	0.261	0.015	0.029	Crores	14,535	15,000	14,000	Smaller the better	− 83.237	1.00	0.058
Cost associated with energy consumption (%) (EC3)	0.237	0.063	A	0.129	0.008	0.032	Crores	284.8	250.5	301.25	Smaller the better	− 48.932	5.42	0.341
Improvement in sales revenue (EC4)	0.078	0.020	G	0.261	0.005	0.010	scale	7	7	7	Larger the better	16.902	1.00	0.020
Cost associated with processing of recyclables (EC5)	0.078	0.020	G	0.261	0.005	0.010	Scale	5	4	5	Smaller the better	− 13.424	10.00	0.2
Cost of disposal for hazardous and unprocessed waste (EC6)	0.043	0.011	A	0.129	0.001	0.005	Lakhs	50	50	50	Smaller the better	− 33.97	7.35	0.08
Qualitative index from L-score method10^*(∑LS/∑LMAX) = 4.159				Sum	0.047	0.113	Quantitative index from S/N ratio method = 0.204							1.229

Table

CSSCPI based on Liberatore method							CSSCPI based on S/N ratio method
[Indicator]Environmental performance: 0.258	Local weight (LW)	Global weight (GW)	Overall rating	Liberatore rating value	Liberatore score (LS)	Maximum score LMAX	Unit of measurement	Indicators value			Characteristic of measure	S/N ratio	Normalized values	Weighted normalized values
								Qu1	Qu2	Qu3
[Sub-indicators]
[sub-indicators]Total waste (EN1)	0.026	0.006	G	0.261	0.001	0.003	ton	24,906	20,923	27000	Smaller the better	− 87.750	3.70	0.022
Utility utilization (EN2)	0.026	0.006	O	0.513	0.003	0.003	GJ	1,300,000	1,393,922	1,487,844	Smaller the better	− 122.898	1.00	0.006
Implementation of environmental regulations (EN3)	0.053	0.013	G	0.261	0.003	0.006	Scale	8	8	9	Larger the better	18.376	5.13	0.066
Recycled material in products (EN4)	0.053	0.013	A	0.129	0.001	0.006	kg/vehicle	17	17	17	Larger the better	24.60	6.53	0.084
Product to be disposed to landfill or incinerated (EN5)	0.026	0.006	A	0.129	0.007	0.003	%	8	5	5	Smaller the better	− 15.797	9.24	0.055
Biodegradable content in products (EN6)	0.105	0.027	A	0.129	0.003	0.013	%	2	2	2	Larger the better	6.02	2.35	0.063
Market share controlled by green products (EN7)	0.079	0.02	A	0.129	0.002	0.01	%	40	40	45	Larger the better	32.356	8.28	0.165
Availability of environmental evaluation schemes (EN8)	0.026	0.006	A	0.129	0.007	0.003	Scale	8	8	8	Larger the better	18.061	5.06	0.03
Environmental management initiatives (EN9)	0.053	0.013	G	0.261	0.003	0.006	Scale	8	9	9	Larger the better	18.716	5.21	0.067
Availability of standard operating procedure for the collection of ELVs (EN10)	0.026	0.006	P	0.034	0.0002	0.003	Scale	2	2	2	Larger the better	6.02	2.35	0.014
Availability of collection centres (EN11)	0.026	0.006	P	0.034	0.0002	0.003	Scale	1	1	1	larger the better	0	1.00	0.006
Reduction of air emissions (EN12)	0.079	0.02	F	0.063	0.001	0.01	Scale	2	2	2	Smaller the better	− 6.020	10.00	0.2
Supplier commitment on overall environmental aspects (EN13)	0.105	0.027	G	0.261	0.007	0.013	Scale	9	8	9	Larger the better	18.716	5.21	0.140
Improvement of resource efficiency (EN14)	0.079	0.02	G	0.261	0.005	0.01	%	8	7	8	Larger the better	17.639	4.96	0.099
Percentage of suppliers having ISO certification (EN15)	0.026	0.006	O	0.513	0.003	0.003	%	100	100	100	Larger the better	40	10.00	0.06
Level of management effort to motivate suppliers (EN16)	0.053	0.013	G	0.261	0.003	0.006	Scale	9	8	9	Larger the better	18.716	5.211	0.067
Availability of environmental reward systems (EN17)	0.053	0.013	A	0.129	0.001	0.006	Scale	8	8	9	Larger the better	18.376	5.135	0.066
Percentage decrease in scrap and rework (EN18)	0.105	0.027	G	0.261	0.007	0.013	%	5	4	4	Smaller the better	− 12.787	9.47	0.255
Qualitative index from L-score method10^*(∑LS/∑LMAX) = 4.783				Sum	0.0574	0.12	Quantitative index from S/N ratio method = 0.0813							1.465

Table

CSSCPI based on Liberatore method							CSSCPI based on S/N ratio method
[Indicator]Social performance: 0.196	Local weight (LW)	Global weight (GW)	Overall rating	Liberatore rating value	Liberatore (LS) score	Maximum score LMAX	Unit of measurement	Indicator values			Characteristic of measure	S/N ratio	Normalized values	Weighted normalized values
								Qu1	Qu2	Qu3
[Sub-indicators]
Green product warranty (SO1)	0.038	0.007	P	0.034	0.0002	0.003	Scale	2	2	2	Larger the better	6.02	1.00	0.007
Customer interest in green products (SO2)	0.031	0.006	A	0.129	0.0007	0.003	Scale	5	6	6	Larger the better	14.968	7.16	0.042
Customer satisfaction from green products(SO3)	0.077	0.015	A	0.129	0.001	0.007	Scale	5	6	6	Larger the better	14.968	7.16	0.10
Customer involvement in returning the ELVs (SO4)	0.077	0.015	F	0.063	0.0009	0.007	Scale	7	6	7	Larger the better	16.408	8.15	0.107
Conduction of training and development activities (SO5)	0.051	0.01	G	0.261	0.002	0.005	Scale	8	8	8	Larger the better	18.061	9.29	0.092
Employee well-being (occupational health and safety) (SO6)	0.111	0.021	O	0.513	0.010	0.010	Scale	9	9	9	Larger the better	19.084	10.00	0.21
Child labour involvement (SO7)	0.155	0.03	P	0.034	0.001	0.015	Scale	1	1	1	Smaller the better	0	10.00	0.3
Involvement in charitable activities (SO8)	0.129	0.025	G	0.261	0.006	0.012	Scale	8	7	8	Larger the better	17.639	9.00	0.225
Management commitment on stakeholder welfare (SO9)	0.136	0.026	G	0.261	0.006	0.013	Scale	9	9	9	Larger the better	19.084	10.00	0.26
Corruption (SO10)	0.111	0.021	A	0.129	0.002	0.010	Scale	4	3	3	Smaller the better	− 10.543	1.00	0.021
Disclosure of environmental initiative to the public (SO11)	0.077	0.015	G	0.261	0.003	0.007	Scale	9	9	9	Larger the better	19.084	10.00	0.15
Qualitative index from L-score method10^*(∑LS/∑LMAX) = 3.565				Sum	0.0328	0.092	Quantitative index from S/N ratio method = 0.137							1.514

Table

CSSCPI based on Liberatore Method							CSSCPI based on S/N ratio method
[Indicator]Technical weight (0.269)	Local weight (LW)	Global weight (GW)	Overall rating	Liberatore rating value	Liberatore score (LS)	Maximum score LMAX	Unit of measurement	Indicator values			Characteristic of measure	S/N ratio	Normalized values	Weighted normalized values
								Qu1	Qu2	Qu3
[Sub-indicators]
IT implementation (TE1) (%)	0.247	0.024	G	0.261	0.006	0.012	%	96	96	96	Larger the better	23.780	10.00	0.24
Adoption of new technologies (TE2)	0.153	0.015	G	0.261	0.003	0.007	Scale	8	8	8	Larger the better	18.061	1.00	0.015
Upgradation of existing technologies (TE3)	0.224	0.022	O	0.513	0.011	0.011	Scale	9	9	9	Larger the better	19.673	3.537	0.077
Level of information security (TE4)	0.374	0.037	O	0.513	0.018	0.018	Scale	9	9	9	Larger the better	20	4.051	0.149
Qualitative index from L-score method10^*(∑LS/∑LMAX) = 7.916				Sum	0.038	0.048	Quantitative index from S/N ratio method = 0.120							0.481

Table

CSSCPI based on Liberatore method							CSSCPI based on S/N ratio method
								Indicator values
[Indicator]Political weight (0.269)	Local weight (LW)	Global weight (GW)	Overall rating	Liberatore rating value	Liberatore score (LS)	Maximum score LMAX	Unit of measurement	Qu1	Qu2	Qu3	Characteristic of measure	S/N ratio	Normalized values	Weighted normalized values
Government regulations on environmental aspects (PO1)	0.587	0.103	G	0.261	0.026	0.052	Scale	4	4	6	Larger the better	12.930	1.00	0.103
Political stability (PO2)	0.154	0.027	G	0.261	0.007	0.013	Scale	8	8	9	Larger the better	18.376	10.00	0.27
Tax structure (PO3)	0.258	0.045	A	0.129	0.005	0.023	Scale	2	3	2	Smaller the better	− 7.533	1.00	0.045
Qualitative index from L-score method10^*(∑LS/∑LMAX) = 4.318				Sum	0.038	0.088	Quantitative index from S/N ratio method = 0.139							0.418