Measurement of corporate social responsibility of automobile enterprises based on AHP-GRA model

Abstract

With the development of economic globalization, great importance has been attached to corporate social responsibility by firms, governments and social organizations. Currently, only a few domestic automobile enterprises can fully disclose their social responsibility information. Most automobile enterprises fail to establish an effective and comprehensive mechanism for social responsibility management and information disclosure, and the concept of social responsibility has not been popularized and widely accepted by automobile enterprises. In order to solve the problem of externalization of internal costs incurred from personal and property safety of consumers and environmental pollution control which should be borne by automobile enterprises, and to promote the establishment and effective implementation of a social responsibility management system in the automobile industry, this study constructs a comprehensive China automobile CSR evaluation index system from four aspects, including CSR to government, CSR to employees, CSR to consumers and CSR to environment and community. In addition, this study makes a comprehensive comparison of the CSR of China’s automobile enterprises based on the synthetical evaluation method of Analytic Hierarchy Process and Grey Relation Analysis (AHP-GRA Model). The results show that: (1) under the influence of macro-economic control by the state, automobile enterprises, irrespective of being joint ventures, state-owned or private enterprises, hold the green development of firms in great account, especially corporate environmental governance; (2) the AHP-GRA Model can offer remedies to the single evaluation model, making up for its lack of accuracy and objectivity; (3) through a trend evaluation based on the AHP-GRA comprehensive evaluation model, a simple and reliable evaluation model is provided for the government, market and consumers to understand the current situation and future development of the CSR of China’s automobile enterprises in a comprehensive and objective way. The study shows that the AHP-GRA comprehensive evaluation model is feasible in the evaluation of social responsibilities of automobile enterprises, which covers the shortage of the single evaluation method, and provides new decision-making ideas and methods for the evaluation and optimization of social responsibility of automobile enterprises.

Keywords

China automobile enterprises social responsibility AHP-GRA model

1 Introduction

Corporate Social Responsibility (CSR) refers to the social, environmental and economic responsibilities that a firm consciously undertakes when engaging in business activities. The concept of CSR was first proposed by Sheldon in 1924. Sheldon believed that a firm should also bear the responsibilities of employees, consumers and the society in connection with the firm while benefiting from their activities. As a matter of fact, CSR is comprised of moral obligations [1]. In the 20th century, capitalist countries such as the United States and those in Europe had in-depth academic research and practice on CSR, and firms and various unions, organizations, and governments have begun to accept the view that “firms should assume social responsibility”. Meanwhile, the business world and government organizations have also begun to lay emphasis on the system construction and implementation of CSR [2]. In the late 1970s, less than half of the Fortune 500 companies made reference to CSR in their annual reports. However, by the end of the 1990s, more than 90% of those companies had incorporated CSR into their annual objectives. Volkswagen Group is one of the most influential automobile enterprises in the world. As a traditional large-scale automobile enterprise, Volkswagen Group has always been concerned with its CSR performance. In fulfilling its CSR duties, the Volkswagen Group has been dedicated to incorporating CSR into its overall corporate strategy.

At present, only a handful of automobile enterprises in China are able to realize full disclosure of their CSR information. The majority of enterprises in the industry falls flat in making their CSR information public and is at a relatively underdeveloped stage. On the one hand, such a phenomenon reflects that the concept of social responsibility has not been widely accepted and recognized by the automobile enterprises, and most of those enterprises have not integrated social responsibility into their day-to-day operation and business management. On the other hand, it also reflects that automobile enterprises have not yet been able to set up a valid and well-rounded mechanism for social responsibility management and information disclosure. These enterprises disclose their information in an untimely and passive manner, and they lacks appropriate and efficient communication with the stakeholders [3]. Therefore, this study selects six listed Chinese automobile enterprises for analysis. The study is carried out from four aspects, including CSR to the government, CSR to employees, CSR to consumers and CSR to environment and community. Based on the synthetical evaluation methods of Analytic Hierarchy Process and Grey Relation Analysis, this study establishes the AHP-GRA model for a comprehensive comparison of the social responsibilities of China’s automobile enterprises.

2 Literature review

At the moment, scholars have carried out extensive research on partner-related issues. Nelarine et al. [4] studied on organizations that do not value CSR or are unsuccessful in their CSR practices, and these organizations have been used for comparison with the small and medium-sized enterprises (SMEs) in the private sector that perform relatively better in CSR practices. Through horizontal comparison between them, they identify a number of indexes to measure the CSR, and put forward relevant suggestions. Asmussen et al. [5] carried out study on the coordination and control mechanism of multinational corporations (MNCs) in the implementation of global CSR strategy when such strategy is in conflict with the operation strategy at the subsidiary level. Palakshappa and Grant [6] researched on the association between social enterprise (SE) and corporate social responsibility (CSR). On account of the relationship between corporate size and relative organizational costs, Baumann-pauly et al. [7] compared the differences between MNCs and SMEs in terms of organizational CSR, and provided theoretical explanations correspondingly. Through large sample analysis on 3,040 U.S firms and an annual observation of 16,606 firms between 1991 and 2010, Attig et al. [8] found that there is a positive correlation between corporate internationalization and CSR ratings. In a similar manner, firms with a wide range of foreign subsidiaries in countries with well-functioning political and legal institutions also enjoy better CSR ratings. Krishna [9] proposed that micro-enterprises and large enterprises are equally motivated to participate in CRS in terms of visibility, access to resources and business scale. However, there is a U-shaped relationship between corporate size and CRS participation. Flammer [10] explored the impact of product marketing on CSR, and the research results show that trade liberalization is an important factor affecting CSR practices. Kitzmueller and Shimshack [11] defined CSR from an economic perspective and developed CSR taxonomy. Basu and PalazzoI [12] conducted research on how the patterns of interrelationships among cognitive, linguistic, and semantic dimensions can raise the profile of CSR. Turker [13] studied several methods and standards of CSR measurement, and developed a scale for measuring CSR through exploratory factor analysis. The results of the analysis provide a four-dimensional structure of CSR, including CSR to social and nonsocial stakeholders, employees, customers, and government. Agudo et al. [14] developed an empirical procedure for measuring corporate social performance (CSP) from the perspective of corporate social responsibility (CSR) and the understanding of a group of corporate managers on the importance of different aspects of proper corporate management. McWilliams and Siegel [15] analyzed the creation and acquisition of private and social value by firms adopting a CSR strategy. Harjot et al. [16] discussed the impact of board diversity on corporate social responsibility (CSR) performance. Donia et al. [17] developed a 14-item Substantive and Symbolic Corporate Social Responsibility (CSR-SS) scale. Ehsan et al. [18] used a multi-method to both quantitatively and qualitatively measure the CSR practices of more than 170 listed companies in Pakistan from 2008 to 2015. The results show that CSR information with respect to products, customers and stakeholders is mostly disclosed by Pakistan firms while relatively less information is disclosed regarding the health and education dimensions of CSR. Guo et al. [19] used event-study analysis and the indices in the 2009–2016 CSR Report by Rankins CSR Ratings (RKS) to test the response of China’s A-share market to firms that disclose social responsibility information. The research results show that there is not significant relation between the cumulative abnormal return of stocks and the CSR report. Meng et al. [20] targeted to research on the 42 listed companies in the transportation industry in 2015. Through fuzzy TOPSIS, 39 evaluation indices are weighted, and the performance evaluation model of CSR in the transportation industry is constructed. Wang et al. [21] evaluated the level of CSR performance by calculating the total CSP score of the firm based on AHP model analysis method. Pashchenko [22] used fuzzy comprehensive evaluation method and analytic hierarchy process to comprehensively evaluate CSR satisfaction of Chinese private enterprises from the perspective of employees. Tao and Leng [23] selected 22 indices from the perspectives of contractual stakeholders and public stakeholders based on the SMART principle, and constructed the CSR Evaluation System for mineral-resource enterprises with the analytic hierarchy process (AHP). The empirical analysis is conducted based pm the annual report data of Yanzhou Coal Mining Company Limited, which is taken as a sample, and has found that its CSR performance show a W-shaped change. Xu et al. [24] made modification to the existing evaluation method of CSR of forestry enterprises, and proposed an entropy evaluation method based on time sequence to simulate the data. It has been found that the consideration of time sequence can capture the CSR information missing in static analysis, which effectively makes up for the insufficiency of the existing evaluation method of CSR of forestry enterprises. Tang and Yin [25] used the CSR of 30 insurance enterprises in China as a sample to carry out multi-index dynamic comprehensive evaluation. The study has found that although the results obtained by the three weighting methods are different, they have been proved to be objective and stable, and there is no overall distribution difference in their rankings. Qi et al. [26] compared and evaluated the CSR performance of major stock market index constituents in China and the United States from four perspectives, including overall performance, ranking changes, market value impact, and industry distribution. By a brief analysis of research on CSR both at home and abroad, Zou and Li pointed out the current limitations of CSR evaluation and measurement. According to the latest research results of foreign researchers, they have mainly discussed on the feasibility of the application of item response theory model in CSR evaluation, and provided an all-round introduction of the mathematical model of CSR evaluation based on the item response theory [27]. Mao et al. used the extended balanced scorecard to divide the content and logical relationship of CSR undertaken by M firms, and then designed a strategic goal-oriented CSR evaluation index system based on the demands of various stakeholders. Finally, according to the relevant collected data of M firms, the CSR performance of M firms is evaluated with regard to quantitative method [28].

Unlike natural sciences, the core challenge of measurement in the area of corporate strategic management is that the characteristics of the firm or individual that we want to measure are implicit and not easy to be observed directly. As a consequence, the scarcity of such measurement method is so serious that some scholars call this challenge “one of the most serious threats in strategic management research” [29]. CSR is rich in content, and due to cultural differences in various regions, no such consensus has been reached with regard to the evaluation methods in the previous literature. Most of the research uses mathematical research methods such as TOPSIS model, entropy model, balanced scorecard, and the item response theory model. Nonetheless, a single evaluation method cannot ensure the systematisms and rationality of the evaluation process. In view of this, based on the characteristics of the automobile industry, this study constructs an index system that includes four dimensions, CSR to government, CSR to employees, CSR to consumers, and CSR to the environment and the community. The CSR of six listed automobile companies in China has been measured based on the AHP-GRA model, and the results of each dimension and its measurement have been analyzed, which aims to provide impetus to the development of CSR in the automobile industry.

3 Research method

3.1 Index system for CSR measurement of automobile enterprises

Based on the scientific nature, comparability and availability of the content and measurement indices of CSR of automobile enterprises, this study measures the CSR of automobile enterprises from four dimensions. Accordingly, the measurement indices are categorized into 3 levels, including 4 first-level indices, 13 second-level indices, and 27 third-level indices.

3.1.1 CSR to government

The evaluation indices in this section mainly cover four parts, including those regarding taxation, corporate economic benefits, employment, and compliance with laws and regulations. The main indices are as follows: (1) total annual tax payment (RMB 10,000); (2) tax contribution rate (%), i.e. total annual tax payment/total annual operating income; (3) total annual operating income (RMB 10,000); (4) Enterprise asset-liability ratio (%); (5) Employment contribution ratio (%), i.e. the total annual employee compensation/annual corporate assets; (6) whether there have been any violations of laws and regulations during the year [31].

3.1.2 CSR to employees

The evaluation indices in this section mainly cover three parts, which are in respect to compensation and benefits, management training and safety in production [32]. The main indices are as follows: (1) total annual salary of employees (RMB 10,000); (2) annual per capita income of employees (RMB 10,000/person); (3) growth rate of employee wages (%); (4) per capita management expenses (RMB 10,000/person); (5) employee training expense ratio (%), i.e. the total annual employee training expenses/corporate annual net profit; (6) per million man-hour injury rate (%); (7) total investment in safety production (RMB 10,000).

3.1.3 CSR to consumers

The evaluation indices in this section mainly cover three parts, including those concerned with R&D investment, product and service quality, and customer stability [33]. The main indices are as follows: (1) the proportion of R&D expenditure invested in the current year in the total operating income of the same year (%); (2) the proportion of R & D personnel in the total number of employees (%); (3) the number of annual market recall events (4) the number of patents licensed annually; (5) the total annual investment in after-sales service (RMB 10,000); (6) the annual sales volume of new energy vehicles; (7) the sales growth rate.

3.1.4 CSR to environment and community

The evaluation indices in this section mainly cover three parts with regard to environmental governance, energy consumption and waste disposal, and responsibility to community. The main indices are as follows: (1) the total annual investment in environmental governance (RMB 10,000); (2) whether it has passed the environmental management system certification; (3) whether there have been major environmental accidents in the year; (4) COD emission reduction rate; (5) comprehensive energy consumption reduction rate per RMB 10,000 output value; (6) amount of social donations; (7) number of new employees.

3.2 AHP-GRA model for CSR measurement of automobile enterprises

3.2.1 Determination of the weight of factors to measure

The weight indicates the quantitative distribution of the importance of an index. The methods for determining the weight that are commonly used include Delphi method, expert investigation method and analytic hierarchy process. This chapter mainly determines the weight of each factor based on the analytic hierarchy process [34]. In order to facilitate the presentation, Ω can be used here to represent the weight set of the first index level to the target level, Ωi represents the weight set of the second index level to the first index level, and Ω_ij represents the weight set of the third index level to the second index level, which can be expressed as: $Ω = {w_{1}, w_{2}, w_{3}}$ (1) $Ω = {α_{t}^{(i)}}$ (2) $Ω_{ij} = {β_{s}^{(j)}}$ (3)

Of which: i represents the number of evaluation indices (i = 1,2,3) in the first index level, t stands for the number of centralized evaluation indices (t = 1, 2, ... , n₁) corresponding to the second index level, j is the number of evaluation indices (j = 1, 2, ... , n₂) in the second index level, and s is the number of centralized evaluation indices (s = 1, 2, ... , n₃) corresponding to the third index level.

Focusing on the CSR measurement of Chinese automobile enterprises, we use the analytic hierarchy process to obtain the weight according to the model of hierarchical structure established in this study, which is detailed as follows:

(1) Single-level priority

In the light of the results of the questionnaire filled by senior experts in the industry, this study constructs a judgment matrix for each index after synthetic analysis, and carries out a consistency check accordingly with regard to V, the CSR of Chinese automobile enterprises, the obtained values are shown in Table 1.

Table 1

A	V1	V2	V3	V4	V5	Intensity of importance
V1	1	23/31	82/89	52/63	1	0.263
V2	31/54	1	1/14	4/25	31/54	0.226
V3	17/82	14/15	1	53/61	17/82	0.320
V4	11/52	25/29	23/66	1	11/52	0.191

Whereas λ_max = 4.2300, CI = 0.077, RI = 0.9, CR = 0.085 < 0.1.

With respect to V1, the CSR to government, the obtained values are presented in Table 2.

Table 2

V1	V11	V12	V13	V14	Intensity of importance
V11	1	44/89	48/53	33/76	0.215
V12	2/3	1	10/37	31/34	0.266
V13	5/48	37/84	1	50/61	0.195
V14	10/33	3/31	11/50	1	0.325

Whereas λ_max = 4.2470, CI = 0.082, RI = 0.9, CR = 0.091 < 0.1.

In respect to V2, the CSR to employees, the calculated values are displayed in Table 3.

Table 3

V2	V21	V22	V23	Intensity of importance
V21	1	3/16	48/83	0.347
V22	16/35	1	9/13	0.218
V23	35/48	4/9	1	0.435

Whereas λ_max = 3.1039, CI = 0.0520, RI = 0.58, CR = 0.089 < 0.1.

With reference to V3, the CSR to consumers, the values obtained are detailed in Table 4.

Table 4

V3	V31	V32	V33	Intensity of importance
V31	1	63/88	27/50	0.351
V32	25/63	1	13/60	0.519
V33	50/127	60/253	1	0.13

Whereas λ_max = 3.0033, CI = 0.0016, RI = 0.58, CR = 0.0028 < 0.1.

In relation to V4, the CSR to environment and community, the values obtained are listed in Table 5.

Table 5

V4	V41	V42	V43	Intensity of importance
V41	1	33/61	74/75	0.563
V42	61/155	1	4/5	0.296
V43	75/224	5/14	1	0.141

Whereas λ_max = 3.0843, CI = 0.04215, RI = 0.58, CR = 0.073 < 0.1.

By applying the same calculation method, the author has also determined the weight of the third-level indicators [35]. The specific results are shown in Table 6.

Table 6

Weight of overall hierarchical priority of CSR of Chinese automobile enterprises

First-level index	Weight	Second-level index	Weight	Third-level index	Weight	Single-level priority	Overall hierarchical priority
CSR to government	0.263	Taxes	0.215	Total annual tax payment	0.637	0.137	0.036
				Tax contribution rate	0.363	0.078	0.021
		Corporate economic benefits	0.266	Total annual operating income	0.671	0.178	0.047
				Enterprise asset-liability ratio	0.329	0.088	0.023
		Employment	0.195	Employment contribution rate	1	0.195	0.051
		Compliance with laws and regulations	0.325	Whether there have been any violations of laws and regulations during the year	1	0.325	0.085
CSR to employees	0.226	Compensation and benefits	0.347	Total annual salary of employees	0.452	0.157	0.035
				Annual per capita income of employees	0.368	0.128	0.029
				Growth rate of employee wages	0.18	0.062	0.014
		Management training	0.218	Per capita management expenses	0.775	0.169	0.038
				Employee training expense ratio	0.225	0.049	0.011
		Safety in production	0.435	Per million man-hour injury rate	0.768	0.334	0.076
				Total investment in safety production	0.232	0.101	0.023
CSR to consumers	0.32	R&D investment	0.351	The proportion of R &D expenditure invested in the current year in the total operating income of the same year	0.753	0.264	0.085
				The proportion of R &D personnel in the total number of employees	0.247	0.087	0.028
		Product and service quality	0.519	Number of annual market recall events	0.372	0.193	0.062
				Number of patents licensed annually	0.372	0.193	0.062
				Total annual investment in after-sales service	0.165	0.086	0.027
				Annual sales volume of new energy vehicles	0.091	0.047	0.015
		Customer stability	0.13	Sales growth rate	1	0.130	0.042
CSR to environment and community	0.191	Environmental governance	0.563	Total annual investment in environmental governance	0.443	0.249	0.048
				Whether it has passed the environmental management system certification	0.326	0.184	0.035
				Whether there have been major environmental accidents in the year	0.231	0.130	0.025
		Energy consumption and waste discharge	0.296	COD emission reduction rate	0.577	0.171	0.033
				Comprehensive energy consumption reduction rate per RMB 10,000 output value	0.423	0.125	0.024
		Responsibility to community	0.141	Social donations	0.534	0.075	0.014
				Number of new employees	0.466	0.066	0.013

3.2.2 Determination of the index value sequence of the measured object

Suppose there are m objects to be measured, and n indices are selected as the lowest-level evaluation indices in the evaluation system. Then we have m sequences, X_i = {X_i(1), X_i(2), ... , X_i(n)}, i = 1,2, ... , m, of which X_i(j) is the value of index j of the ith evaluated object. The original index values of the third-level indices in the CSR measurement system of China’s automobile enterprise should be composed of the following matrices: $X = [\begin{matrix} X_{i} (1) & X_{1} (2) & \dots & X_{1} (n) \\ X_{2} (1) & X_{2} (2) & \dots & X_{2} (n) \\ ⋮ & ⋮ & ⋮ \\ X_{m} (1) & X_{m} (2) & \dots & X_{m} (n) \end{matrix}]$ (4)

3.2.3 Selection of the optimal reference sequence

Suppose X₀ = X₀(1), X₀(2), ... , X₀(n), where X₀ (k)(k = 1, 2, ... , n) is the optimal value of the kth index among the evaluated objects. Of all the indices, if an index is positive, the maximum value of the index in each scheme is taken; if an index is negative, the minimum value in each scheme is taken. The optimal reference sequence should be: $\begin{matrix} X_{0} = {752071, 8, 3.1, 75641617, 65.54, 2.13, 1, \\ 1000618, 6.97, 42.49, 23.56, 18.78, 0.08, 71000, \\ 7.27, 14.42, 1, 1722, 1788694, 507000, 69.85, \\ 25245.84, 1, 1, 27.9, 16, 6720, 8467} \end{matrix}$

3.2.4 Normalization of index data

Due to different dimensions and orders of magnitude between indices and in order to facilitate the pair-wise comparison between indices and the calculation of the relational degree, the sequence needs to be normalized. The normalized formula is as follows: $λ_{ik} = \frac{X_{i} (k) - min_{i} X_{i} (k)}{max_{i} X_{i} (k) - min_{i} X_{i} (k)}$ (5)

Where i = 1, 2, ... , m; k = 1, 2, ... , n. λ_ik represents the normalized value of the kth index X_i(k) of the ith evaluated object. The following matrix can be obtained after normalization: $λ = [\begin{matrix} λ_{01} & λ_{02} & \dots & λ_{0 n} \\ λ_{11} & λ_{12} & \dots & λ_{1 n} \\ \dots & \dots & \dots & \dots \\ λ_{m 1} & λ_{m 2} & \dots & λ_{mn} \end{matrix}]$ (6)

3.2.5 Determination of the relational coefficient matrix

If we take the normalized sequence of λ₀ =λ₀₁, λ₀₂, ... , λ_0n as the reference sequence, and λ_i =λ_i1, λ_i2, ... , λ_in (i = 1, 2, ... , m) as the comparative sequence, then the relational coefficient can be calculated by the following formula: $ξ_{ik} = \frac{min_{i} min_{k} | λ_{0 k} - λ_{ik} | + ρ max_{i} max_{k} | λ_{0 k} - λ_{ik} |}{| λ_{0 k} - λ_{ik} | + ρ max_{i} max_{k} | λ_{0 k} - λ_{ik} |}$ (7)

Where i = 1,2, ... , m; k = 1,2, ... , n, ρ is the distinguishing coefficient and ρ∈ [0,1]. Generally, if we take ρ= 0.5, the incidence coefficient matrix can be obtained as follows: $E = [\begin{matrix} ξ_{11} & ξ_{12} & \dots & ξ_{1 n} \\ ξ_{21} & ξ_{22} & \dots & ξ_{2 n} \\ ⋮ & ⋮ & ⋮ \\ ξ_{m 1} & ξ_{m 2} & \dots & ξ_{mn} \end{matrix}]$ (8)

Where ξ_ik(i = 1,2, ... , m; k = 1,2,, ... , n) is the incidence coefficient between the k-th index and the k-th optimal index of the i-th evaluated object. The relational coefficients are shown in Table 7.

Table 7

Incidence coefficient value

	Enterprise A	Enterprise B	Enterprise C	Enterprise D	Enterprise E	Enterprise F
λ ₁	0.366266	0.333333	0.375845	0.3505	1	0.394709
λ ₂	0.359455	0.447983	1	0.536896	0.333333	0.489559
λ ₃	0.361825	0.333333	0.343669	0.33655	1	0.360116
λ ₄	0.900039	1	0.333333	0.41569	0.683368	0.387721
λ ₅	0.855856	0.487179	0.418502	0.333333	0.519126	1
λ ₆	1	1	1	1	1	1
λ ₇	0.408745	0.334734	0.355827	0.333333	1	0.37668
λ ₈	0.333333	1	0.335807	0.340866	0.706112	0.391492
λ ₉	1	0.348037	0.349541	0.333333	0.426982	0.54877
λ ₁₀	0.333333	1	0.334446	0.5278	0.587562	0.368402
λ ₁₁	1	0.398516	0.333333	0.333455	0.342204	0.334187
λ ₁₂	0.542373	0.680851	1	0.881543	0.886427	0.333333
λ ₁₃	0.365695	0.339579	0.333333	0.335976	1	0.351683
λ ₁₄	0.508891	0.483508	0.551882	1	0.333333	0.426544
λ ₁₅	0.411602	0.333333	1	0.547794	0.541818	0.919753
λ ₁₆	1	0.333333	0.333333	1	0.4	0.8
λ ₁₇	0.589823	0.333333	0.359155	0.353186	1	0.494186
λ ₁₈	0.344429	0.335078	0.338408	0.333333	1	0.345123
λ ₁₉	0.374267	0.335911	0.333333	0.338833	0.343155	1
λ ₂₀	1	0.333333	0.485737	0.372725	0.348966	0.462247
λ ₂₁	1	0.334104	0.333333	0.344802	0.339521	0.343406
λ ₂₂	1	1	1	1	1	1
λ ₂₃	1	1	1	1	1	1
λ ₂₄	1	0.333333	0.744578	0.732227	0.544974	0.452416
λ ₂₅	1	0.666667	0.333333	0.488155	0.382022	0.723404
λ ₂₆	0.396719	0.336593	1	0.333333	0.594891	0.337999
λ ₂₇	0.865704	0.333333	1	0.365024	0.809276	0.417749

3.2.6 Calculation of the grey relational degree

Calculation of single-level relational degree. The calculation formula of grey relational degree is R = (r_i)_l ×m = W×E^T, where W is the matrix composed of the weight of each index of a certain level relative to the target of the upper level, and E is the aforesaid relational coefficient matrix.

Calculation of the final relational degree of a multi-level evaluation system. For a multi-level evaluation system, the calculation method of the final relational degree is as follows: the relational coefficient of indices in the l_th level is synthesized to obtain the relational degree among indices in the l-1_th level to which they belong, and then the relational degree obtained in this level is taken as the original data to continue to synthesize to obtain the relational degree of indices in the l-2_th layer, and so on until the relational degree of indices in the highest level is obtained. Finally, the relational degree of the highest index level V can be obtained. The specific results are shown in Table 8.

Table 8
Social responsibility measurement of 6 listed automobile enterprises

Enterprise A B C D E F

Measured value 0.688982 0.548821 0.578301 0.610403 0.693965 0.573977

Enterprise	A	B	C	D	E	F
Measured value	0.688982	0.548821	0.578301	0.610403	0.693965	0.573977

4 Results analysis and discussion

The measured values of the first-level indices and second-level indices in China automobile CSR measurement system are presented in Tables 9 and 10.

Table 9
Measured values of the first-level indices in the automobile CSR measurement system

First -level index Listed automobile enterprise

A B C D E F

CSR to government 0.713454 0.647624 0.626638 0.576353 0.82749 0.710472

CSR to employees 0.492603 0.652162 0.561098 0.550569 0.786792 0.368833

CSR to consumers 0.654301 0.373296 0.474127 0.666054 0.552578 0.579228

CSR to environment and community 0.945753 0.584568 0.706631 0.634849 0.637146 0.619982

First -level index	Listed automobile enterprise
CSR to government	0.713454	0.647624	0.626638	0.576353	0.82749	0.710472
CSR to employees	0.492603	0.652162	0.561098	0.550569	0.786792	0.368833
CSR to consumers	0.654301	0.373296	0.474127	0.666054	0.552578	0.579228
CSR to environment and community	0.945753	0.584568	0.706631	0.634849	0.637146	0.619982

Table 10

Measured values of second-level indices in the automobile CSR measurement system

Second -level index	Listed enterprise
	A	B	C	D	E	F
Taxes	0.363793	0.374951	0.602414	0.418162	0.758	0.42914
Corporate economic benefits	0.538897	0.552667	0.340268	0.362587	0.895828	0.369198
Employment	0.855856	0.487179	0.418502	0.333333	0.519126	1
Compliance with laws and regulations	1	1	1	1	1	1
Compensation and benefits	0.48742	0.581946	0.347328	0.336105	0.788706	0.413107
Management training	0.483333	0.864666	0.334196	0.484072	0.532357	0.360703
Safety in production	0.501384	0.601676	0.845333	0.754971	0.912776	0.33759
R&D investment	0.48486	0.446415	0.662567	0.888305	0.384829	0.548366
Product and service quality	0.682303	0.333856	0.343776	0.589219	0.717027	0.629383
Customer stability	1	0.333333	0.485737	0.372725	0.348966	0.462247
Environmental governance	1	0.705008	0.704667	0.709747	0.707408	0.709129
Energy consumption and waste discharge	1	0.474333	0.570622	0.628985	0.476045	0.567044
Responsibility to community	0.615266	0.335074	1	0.348101	0.694794	0.375163

According to Table 8, the study establishes the relational order of the evaluation objects, that is, the order of the social responsibility of the six listed automobile enterprises is: $γ E > γ A > γ D > γ C > γ F > γ B$

Enterprise E has the highest value in its social responsibility measurement. The high scores of its CSR measurement come from its CSR to the government and to the employees. Its taxes payment and the economic benefits are significantly higher than those of the other five enterprises, which indicate that the enterprise attaches particular importance to employees’ compensation and benefits, safety in production and operation, and responsibility to community. Noticeably, its R&D investment is weaker than that of enterprises C, D, and F, which has a strong correlation with its development background of being an industry leader and joint venture.

B has the lowest CSR measurement value. In 2016, the enterprise was adversely affected by the macro-economic downturn and the policy adjustment of new energy vehicles. The decline in the sales of passenger cars and chassis (by its headquarters), the downward market demand for pickup trucks (by its branches) as well as the delayed launch of new products led to the decline in customer stability and economic benefits of enterprise B, which in turn result in the decrease in tax payment. In addition, the enterprise’s number of recall events in 2016 was significantly higher than other automobile enterprises, and it also underperformed other enterprises in terms of external donations.

Among the six enterprises, Enterprises A and F are representative private automobile enterprises. These two companies are in relatively higher rankings in terms of employment, R&D investment, product and service quality, and customer stability, which suggest that private enterprises pay more attention to product R&D and product and service quality improvement as compared to joint ventures or foreign-funded enterprises, which are more advantageous in advanced technology. Private enterprises are more dedicated to ensure market stability and fully promote the development of enterprises. In the meantime, these private enterprises are also pillar enterprises in offering more employment opportunities for the local population.

5 Conclusions

CSR has become an important issue worldwide. In order to measure Chinese automobile CSR, the AHP-GRA synthetical evaluation model is constructed in this study by combination of the analytic hierarchy process and grey relational analysis to make up for the deficiency of a single evaluation method, and the applicability of the AHP-GRA synthetical evaluation model is verified through analysis on six selected examples of Chinese listed automobile enterprises. The following conclusions are drawn in this study:

Under the influence national macro policies, both joint-venture, state-owned, and private automobile enterprises take seriously the green development of enterprises, especially the governance of environment, the efforts invested in the development of which is equivalent by enterprises irrespective of the scale of business operations and the amount of economic income.

The AHP-GRA comprehensive evaluation model can better complement the lack of accuracy and objectivity of a single evaluation model, and it can evaluate the social responsibility of Chinese automobile enterprises in a more scientific and reasonable way.

The AHP-GRA synthetical evaluation model is adopted for trend evaluation, which provides a simple and reliable evaluation model for the government, market and consumers to have a comprehensive and objective understanding of the current situation and future development of China automobile CSR.

The research on automobile CSR measurement is a multi-angle and multi-level complex issue. This study uses the AHP-GRA synthetical evaluation model to carry out related research, which is of great practical significance. However, there still exist many problems that need to be further analyzed and explored, for example, the monitoring of the development trend of industry CSR. This study only builds an evaluation model of industry CSR trend, which requires further improvement to cover the future dynamic development of the industry. Using computer technology to establish simulation model is a method worthy of application and research, which would be the interest of future research so as to better promote the integration of research theory and practice in a more effective way.

References

Chilembwe

, Airlines, travel agencies, tour operators, car rentals corporate social responsibility (CSR) practices in Malawi, Scholedge International Journal of Management & Development7 (2015), 24–38.

Ostendorf

, From HMMS to segment models: Stochastic Modeling for CSR, In Automatic speech and speaker recognition, Springer, Boston, MA, (1996), 185–210.

Zhong

H.W.

, Wang

, et al., Research report on corporate social responsibility of automotive enterprises. Beijing, Economy & Management Publishing House, (2016).

Nelarine

, Mathew

, Shaheena

, Adrian

and James

, Corporate social responsibility and the social enterprise, Journal of Business Ethics81 (2008), 355–370.

Asmussen

C.G.

and Fosfuri

, Orchestrating corporate social responsibility in the multinational enterprise, Strategic Management Journal40 (2019), 894–916.

Palakshappa

and Grant

, Social enterprise and corporate social responsibility, International Journal of Entrepreneurial Behavior & Research24 (2018), 606–625.

Baumann-pauly

, Wickert

, Spence

L.J.

and Scherer

A.G.

, Organizing corporate social responsibility in small and large firms: size matters, Journal of Business Ethics115 (2013), 693–705.

Attig

, Boubakri

, El Ghoul

and Guedhami

, Firm internationalization and corporate social responsibility, Journal of Business Ethics134 (2016), 171–197.

Krishna

, Corporate social responsibility and firm size, Journal of Business Ethics83 (2018), 167–175.

10.

Flammer

, Does product market competition foster corporate social responsibility: Evidence from trade liberalization, Strategic Management Journal36 (2015), 1469–1485.

11.

Kitzmueller

and Shimshack

, Economic perspectives on corporate social responsibility, Journal of Economic Literature50 (2012), 51–84.

12.

Basu

and Palazzo

, Corporate social responsibility: A process model of sense making, The Academy of Management Review33 (2008), 122–136.

13.

Turker

, Measuring corporate social responsibility: a scale development study, Journal of Business Ethics85 (2016), 411–427.

14.

Valiente

J.M.A.

, Ayerbe

C.G.

and Figueras

M.S.

, Social responsibility practices and evaluation of corporate social performance, Journal of Cleaner Production35 (2012), 25–38.

15.

McWilliams

and Siegel

D.S.

, Creating and capturing value: strategic corporate social responsibility, resource-based theory, and sustainable competitive advantage, Journal of Management37 (2011), 1480–1495.

16.

Harjoto

, Laksmana

and Lee

, Board diversity and corporate social responsibility, Journal of Business Ethics132 (2015), 641–660.

17.

Donia

M.B.

, Tetrault Sirsly

C.A.

and Ronen

, Employee attributions of corporate social responsibility as substantive or symbolic: validation of a measure, Applied Psychology66 (2017), 103–142.

18.

Ehsan

, Nazir

M.S.

, Nurunnabi

, Raza Khan

, Tahir

and Ahmed

, A multimethod approach to assess and measure corporate social responsibility disclosure and practices in a developing economy, Sustainability10 (2018), 2955.

19.

Guo

, Su

C.Z.

and Zhang

, Does corporate social responsibility disclosure improve the company’s market performance?, System Engineering Theory and Practice39 (2019), 881–892.

20.

Meng

, Shen

S.Y.

and Kuang

H.B.

, Corporate social responsibility evaluation model based on fuzzy TOPSIS: Taking transportation industry as an example, Business Review31 (2019), 191–202.

21.

Wang

X.D.

and Deng

K.Y.

, Construction of the evaluation system of corporate social responsibility, Statistics & Decision10 (2019), 174–177.

22.

Pashchenko

, An empirical study on employee satisfaction of social responsibility in Chinese private enterprises based on fuzzy comprehensive evaluation, Enterprise Reform and Management3 (2019), 5–6.

23.

Tao

and Leng

W.W.

, Evaluation on mineral resources–oriented enterprises’ corporate social responsibility: A case study of Yangzhou coal mining company limited, Journal of Southwest University (Social Science Edition)43 (2018), 108–114.

24.

, Xiong

L.C.

and Cheng

B.D.

, The evaluation method of corporate social responsibility based on entropy method with time series, Ecological Economy34 (2018), 218–222.

25.

Tang

and Yin

S.T.

, Comparison of subjective weighting methods for panel data with non-normal distribution: based on the dynamic comprehensive evaluation of social responsibility of insurance companies dynamic [J], Journal of Shihezi University (Philosophy and Social Science)33 (2019), 27–34.

26.

, Liu

T.Y.

and Qi

X.L.

, Corporate social responsibility evaluation of listed company under the background of green development: Contrastive analysis based on SSE 50 and Dow Jones industrial index, Finance Teaching and Research3 (2019), 3–13.

27.

Carroll

R.J.

, Prim

D.M.

and Richter

B.K.

, Using item response theory to improve measurement in strategic management research: An application to corporate social responsibility, Strategic Management Journal37 (2016), 66–85.

28.

Mao

S.Z.

, Le

G.L.

and Wu

Y.H.

, Enterprise social responsibility evaluation based on balanced scorecard, Finance and Accounting8 (2017), 56–58.

29.

Godfrey

P.C.

, The problem of unobservable in strategic management research, Strategic Management Research16 (1995), 519–533.

30.

Frigant

, CSR in the auto industry: do the first-tier suppliers have Stakeholders, Post-Print9 (2009), 377–393.

31.

Crifo

, Diaye

M.A.

and Pekovic

, CSR related management practices and firm performance: An empirical analysis of the quantity–quality trade-off on French data, International Journal of Production Economics171 (2016), 405–416.

32.

Alex

C.V.

, The decoupling of productivity and compensation in the Mexican auto industry: CSR vs. the moral economy of inequality. https://sase.confex.com/sase/2016am/webprogram/Study5154.html, 2016-06-24.

33.

Cheng

E.W.L.

and Li

, Analytic hierarchy process, Measuring Business Excellence5 (2016), 30–37.

34.

Saaty

T.L.

, How to make a decision: the analytic hierarchy process, European Journal of Operational Research48 (1990), 9–26.

35.

Lopez-Cozar-Navarro

and Benito-Hernandez

, Corporate social responsibility in manufacturing sector: social and economic dimensions, Dyna92 (2017), 308–313.