The effects of corporate social responsibility on customer-based brand equity: Spanish hypermarket case

Abstract Hypermarkets have experienced substantial restructuring in the recent past because their situation has generated the necessity to develop new models that include different new aspects. The question under consideration here is whether the adoption of measures of corporate social responsibility (C.S.R.) by hypermarkets in Spanish markets can solve their continuing problems by establishing a fair degree of consumer-based brand equity. The purpose of the present study is twofold: (a) to investigate the effects of C.S.R. on hypermarket brand equity; and (b) to explore the dimensions of both these variables by using and testing property scales. As such, it is necessary to consider the possibility of integrating the management of C.S.R. into the global strategy of hypermarkets and to analyse its possible effects on the variables that influence consumer-based brand equity. It may be interesting to dedicate resources and efforts to strengthening hypermarkets’ links with their consumers.


Introduction
Hypermarkets have opened all over the world, but they have experienced substantive restructuring in the recent past, and they have generated the necessity to develop new models of management and strategies that need to take into account aspects related to the improvement of brand value. In Spain, as in the rest of Europe, hypermarkets have appeared to be a commercial format for success and, from their first establishment to the present day, the number of hypermarkets has been growing to reach a figure that was inconceivable for this type of establishment until a few years ago. This growth in the number of hypermarkets has not been reflected in their market share, that has been falling since 2001 (see Table 1).
The current situation indicates an entry into a 'saturation stage' (Cuesta & Gutiérrez, 2010), although a decrease in the number of hypermarkets has not yet been apparent in their evolution. Despite the difficulty of entering a saturated market, new hypermarkets are continuing to open. In recent years, different growth strategies have been followed, based on costly acquisition of other companies, business partnerships and entry into other commercial formats in search of the necessary diversification required to offer new and complementary possibilities for local traders.
Thus, the growth strategy followed until now, consisting of opening new hypermarkets, is already insufficient and, if hypermarkets want to avoid entering a stage of decline, they will need to develop new growth strategies. It would therefore seem reasonable to consider the possibility of integrating the management of corporate social responsibility (C.S.R.) into the global strategy of hypermarkets. This could be very important in Spain, where retail brands have recently been growing and have attained a 38.7% market value share in all sections (Nielsen, 2014). Retail brand identifies the goods and services of a retailer and differentiates them from those of competitors. A retailer's brand equity is exhibited in consumers responding more favourably to its marketing actions than they do to competing retailers (Keller, 2003).
Thus, with the additional goal of trying to contribute to the literature in the right way, the present research has the general aim of determining the influence that the economic, ethical and legal, and discretionary dimensions of C.S.R. have on brand equity. This article offers a review of the relevant theoretical literature and a description of the hypotheses of the study before it describes the methodology and rationale for measuring C.S.R. and customer-based brand equity.

Evolution of C.S.R. as a concept
The debate on the content of socially responsible behaviour has been gaining force since the 1960s, when McGuire (1963) supported a company in search of the development of the community, education and the happiness of its own employees. Since this first moment, it has been revealed that companies must answer not only to their shareholders, but to a broader range of the public. Thus, the stakeholder theory annotates its field of action by affirming that companies have responsibilities towards all the individuals or groups that, directly or indirectly, could be affected by its activity; fundamentally, employees, shareholders or investors, clients, suppliers, government and the community in which it acts (Clarkson, 1995). In the most modern definitions, C.S.R. acquires strategic importance, joining in the mission, vision and organisational principles of companies (Hildebrand, Sen, & Bhattacharya, 2011;Porter & Kramer, 2011;Wood, 2010).
Concerning C.S.R. 's dimensions, and considering the definitions of social responsibility that allude to the idea that companies must obtain benefit, obey the law and exceed the law, Carroll (1979) elaborated his model of four dimensions -economic, legal, ethical and philanthropic -a theory that has been widely accepted in the literature on social responsibility (Carroll, 1991;Lewin, Sakano, Stephens, & Victor, 1995;Maignan, Ferrell, & Hult, 1999;Wartick & Cochran, 1985). Along the same line another model can be found; Graafland, Eijffinger, and Smid (2004) considered only three dimensions: economic, related to the commercialisation of products; social, linked with the ethics inside and outside the organisation; and ecological, which includes the responsible behaviour towards the environment. Finally, we can emphasise the stream of studies that relate social responsibility and the development of marketing activities to a social dimension (Davis & Blomstrong, 1975;Drumwright, 1996;Galbreath, 2010;Handelman & Arnold, 1999), since they included aspects like the protection of the environment, investment in the community, conservation of resources and altruistic donations (Quazi & O'Brien, 2000). Nevertheless, these expositions seem to focus on specific issues and forget the important multidimensional character of this concept (Derwall, Koedijk, & Horst, 2011;Maignan & Ferrell, 2000, 2001Seifert, Morris, & Bartkus, 2003).

Customer-based brand equity
Since the term 'brand equity' emerged in the 1980s, there has been a growing interest in the subject among marketing academicians and practitioners. In order to try to understand this concept, Aaker (1991) defines it as a set of assets (and liabilities) linked to a brand's name and symbol that add to (or subtract from) the value provided by a product or service to a firm and/or that firm's customers. It represents the difference between overall brand preference and multi-attributed preference based on objectively measured attribute levels (Park & Srinivasan, 1994) and overall quality and choice intention (Agarwal & Rao, 1996). Customer-based brand equity is defined from the perspective of the customer and is based on consumer knowledge, familiarity, and associations with the brand (Dollatabady & Amirusefi, 2011;Washburn & Plank, 2002). Proponents contend that for a brand to have value, it must be valued by customer. Then, Keller (1993) defines customer-based brand equity as 'the differential effect that brand knowledge about the consumer or how customers respond to the marketing of that brand' and when customers respond more favourably to a product whose brand is identified, the brand will have a positive customer-based brand equity which exists the customer develops a high level of awareness and familiarity and strong, favourable, and unique brand associations (Keller, 2002) and it was established six core brand values: brand salience; brand performance; brand imagery; brand judgements; brand feelings; and brand resonance. Recent definitions of brand equity have evolved, and include the added value of a name, which can expand to include a broad set of attributes that drive customer choice (Faircloth, Capella, & Alford, 2001) and the relationships between consumer brand perception, brand preference and brand choice (Kim, Jin-Sun, & Kim, 2008). Customer-based brand equity is a multidimensional concept and a complex phenomenon. Aaker (1991Aaker ( , 2011 grouped it into five categories: perceived quality; brand loyalty; brand awareness; brand association; and other proprietary brand assets such as patents, trademarks and channel relationships, which have been widely adopted to measure customer-based brands. Recently, Keller (1993Keller ( , 2002 focused on two of these components: awareness and association. Other authors (Barwise, 1993;Yoo & Donthu, 2001;Yoo, Donthu, & Lee, 2000) have also investigated the formation of brand equity and suggested the same dimensions and stated that they are positively related to brand equity. Konecnik and Gartner (2007) suggest that these dimensions have an influence on conative, affective and cognitive components. Finally, Stahl and Heitmann (2012) measured four 'pillars' that capture the awareness/familiarity and brand association constructs encompassed by Keller's theory: knowledge; relevance; esteem; and differentiation.

Methodology
The present study is part of a larger investigation of the relationships between C.S.R. and consumer-based brand equity effects. We can identify this idea in some of the research (Abdolvand & Charsetad, 2013;Chen, 2001;Klein & Dawar, 2004) or by using similar concepts like corporate societal marketing (Hoeffler & Keller, 2002).
The bibliographical review of the variables that constitute the model included many possible relations based on the perception and behaviour of consumers in Spanish hypermarkets. In terms of consumer-based brand equity, the initial survey instrument was developed to incorporate a pool of 19 items compiled from the literature. This finally generated a survey with 15 items, four of which were included in the assessment of brand loyalty, three for perceived quality, five for brand associations and three for brand awareness.
With regard to C.S.R., as discussed previously, this work involves looking for the possible relations that allow us to determine economic, legal, ethical and philanthropic dimensions. This approach has been widely accepted in the literature on social responsibility but after analysing the survey there are 14 items and three C.S.R. dimensions, namely economic, ethical and legal, and discretionary.
Based on the above definition and the suggested relationship of C.S.R. and brand equity in the literature, the following hypotheses can be formulated (see

Hypothesis 2d. Brand loyalty is a significant component of consumer-based brand equity.
Hypothesis 3. C.S.R. has a significant positive direct effect on consumer-based brand equity.

Sample and data collection
The data collection was realised through a study of the customers of Spanish hypermarkets undertaken with a sample survey conducted by interviewers, who had previously been informed about the aims of the study and the technical questions regarding the distribution of the questionnaire. In respect of targeting, Consumers were between 18 and 82 years of age and, trying to cover a representative part of the territory of Spain, questionnaires were distributed in 18 cities of different population sizes and with different brands of hypermarkets in their region. The total sample was formed by 667 individuals who were responsible for the purchases of their households and had bought items in this type of retail store on some occasions. The dimension of this sample is considered to be appropriate for research that uses the questionnaire method, for which the recommended minimum is 500 individuals (Malhotra, 2009). Items were developed from existing scales to measure the five constructs on a five-point Likert scale (1 = strongly disagree to 5 = strongly agree).
Partial least squares (P.L.S.), a structural equation-modelling (S.E.M.) tool, was used to perform the analyses. S.E.M. enables researchers to examine the structural component (path model) and measurement component (factor model) simultaneously in one model (Gefen, Straub, & Boudreau, 2000). In P.L.S., the results are presented in two stages: the measurement model, which includes an assessment of the reliability and validity of the measures; and the structural model, which tests (1) the amount of variance explained, (2) the significance of the relationships, and (3) the model's predictive relevance.

Constructs
This study examines the relationships among nine measured variables (see Figure 1). There are three C.S.R. dimensions, namely economic, ethical and legal, and discretionary. The C.S.R. concept includes different items from the literature, but three dimensions and 14 items were found. The brand equity latent variable used Aaker's original scale for this construct and fifteen items (see Figure 1).
The survey questionnaires were designed to measure these two second-order constructs with three (C.S.R. construct) and four dimensions (consumer-based equity brand equity construct). The questionnaire was composed of two main sections. In the first part, data were collected about the demographic characteristics and behaviour of the respondents. The second part examined these C.S.R. dimensions (economic, legal and ethical, and discretionary dimensions) with 14 items. Brand equity dimensions (perceived quality, brand awareness, brand loyalty and brand associations) were composed of 15 items (Appendix 1). The scale used to measure this dimension was a five-point Likert scale (1 = strongly disagree; 2 = disagree; 3 = neutral; 4 = agree; 5 = strongly agree).
Factor analyses were used to validate the measurement of these constructs separately. Factors were extracted using covariance matrices and the method of principal components.
Varimax rotations were used to help interpret the initial factor patterns. The factor loadings (see Appendix) provide evidence of the factorial validity of the scales.

Results
In order to assess the measure reliability, we examined how each item relates to the latent constructs. Using the rule of thumb of accepting items with loadings of 0.707 or more, we noticed that only 4 of the 29 indicators did not reach this level of acceptable reliability. However, as pointed out by Chin (1998) and Barclay, Higgins, and Thompson (1995), loadings of at least 0.5 might be acceptable if other questions measuring the same construct have high reliability scores. Falk and Miller (1992) proposed as a rule of thumb retaining manifest variables with loadings that exceed 0.55, i.e., 30% of the variance of the manifest variable is related to the component. All of the loadings exceed 0.60 for these items, and load more highly on their own construct than on others. These results provide strong support for the reliability of the reflective measures (see Figure 2).
Finally, we assessed the reflective second-order molecular constructs (Chin, 2010). The previous discussion provides support for the measures of components. The statistics for all the dimensions were as expected. As indicated in Figure 2, the indicators for C.S.R., economic dimension (0.64), legal and ethical (0.75) and discretionary (0.85) suggest that they are major reflectors of C.S.R. The validity of the measurement scales was successfully tested; the estimation of the causal relationships was performed with the statistical software SmartP.L.S. The results confirm the proposed relations between the first three model concepts and C.S.R. Consequently, Hypothesis 1 (and sub hypotheses) is not rejected. Economic dimension could be acceptable because loadings exceed 0.60. Nevertheless, this does not mean to say that the discretionary dimension should require less work to realise, but it is where actions have major repercussions for C.S.R., since it is also where the perception of realised actions is lower for hypermarkets, leaving an open field for improvement activities.
In other case, perceived quality (0.82), brand awareness (0.84), brand associations (0.86) and brand loyalty (0.86) suggest that they are very good reflectors of consumer-based brand equity. The results confirm the proposed relations between four constructs and brand equity and C.S.R. Again, Hypothesis 2 (and sub hypotheses) is not rejected because all loadings are over 0.707.
Internal consistency is assessed using two measures: Cronbach's alpha and composite reliability. Nunnally and Bernstein (1994) suggested 0.70 as a benchmark for 'modest' reliability applicable in the early stages of research and 0.80 as a 'stricter' interpretation of reliability applicable in basic research. The composite reliability of each set of reflective measures for each component exceed 0.80 (less than brand awareness but is very close). About Cronbach's alpha, three constructs are very close to the traditional acceptable value. Four other constructs met the recommended cut-off value. As a result, all of the constructs were acceptable and 29 items were retained for the seven latent variables in the study.
Discriminant validity was assessed in two ways. First, we examined the average variance extracted (A.V.E.), which indicates the amount of variance that is captured by the construct in relation to the variance due to measurement error. The values for the A.V.E. should exceed 0.50. As the statistics indicate, all the A.V.E. values are greater than this value. Second, we compared the square root of the A.V.E. (i.e., the diagonal in Table 2) with the correlations among the constructs (i.e., the off-diagonal elements in Table 2). The square root of the A.V.E. for all of the reflective constructs exceeds 0.5 and each is greater than the correlation between the constructs; in order to demonstrate discriminant validity, the diagonal elements should be greater than the off-diagonal elements (Fornell & Larcker, 1981). These statistics suggest that each construct relates more strongly to its own measures than to the measures of other constructs; that is, all the constructs share more variance with their own measures than with the others. These two sets of findings provide strong evidence of discriminant validity among the constructs.
Collectively, these results provide support for the overall quality of our measures. In particular, the statistics suggest that our component measures are reliable, are internally consistent and have discriminant validity.

Model assessment
A model estimated through P.L.S. algorithms can only be analysed if it is placed within a larger model that incorporates the consequences of the latent variable in question. According to our model statistics, our index explains a relatively large amount of variance in this C.S.R.; the model's R 2 values, the main criteria by which the model fit is assessed in P.L.S. analysis (Chin, 1998), are 0.30, 0.57 and 0.69 for economic, ethical and legal, and discretionary respectively. The situation is better in brand equity dimensions perceived quality (0.67), brand awareness (0.70), brand associations (0.74) and brand loyalty (0.74). In addition, the Stone-Geisser statistics (Q 2 ) are 0.24, 0.50 and 0.65 for economic, ethical and legal, and discretionary respectively, and 0.59, 0.57, 0.67 and 0.69 for perceived quality, brand awareness, brand associations and brand loyalty; values greater than zero indicate that the model has predictive relevance.
To provide evidence of external validity, the C.S.R. in hypermarkets should be significantly correlated with the other construct that theory suggests should be associated with it (Bagozzi, 1994). As indicated earlier and depicted in Figure 1, we included other construct in the study (brand equity) that theory suggests should be related to C.S.R. Consistent with the literature, we estimated a model in which C.S.R. is an antecedent for this construct and it has a significant positive direct effect on brand equity (see Figure 2). Figure 2 shows the positive direct effects between the latent variables. This figure indicates significant path coefficient (at p < 0.001) that was estimated using a P.L.S. bootstrapping procedure utilising 500 resamples, an amount that provides reasonable standard error estimates (Chin, 2010).
The result of this research is the existence of a strong relation between C.S.R. and consumer-based brand equity (0.61). It is necessary to remember that C.S.R. and brand equity are composed of three and four dimensions respectively. These dimensions are the reflection of the same thing in different measures. The model was globally tested to ensure that it fitted properly to the data independently of the differences among customers. Results also show the value of a significant positive direct effect between C.S.R. and brand equity in hypermarkets, then Hypothesis 3 is not rejected.

Discussion and conclusions
The results of this study have contributed evidence that confirms the conclusions of previous studies with regard to the multidimensionality of the concept of C.S.R., and shows that hypermarket consumers generally tend to display responsible behaviour and that it is possible to see this reflected in the economic, ethical and legal and discretionary dimensions (the last two to a greater extent). The results underline existing perceptions of C.S.R. and its various dimensions, and illustrate its effects on consumer-based brand equity and the image the retailer creates in the minds of consumers. They show the importance of this relationship and suggest ways to improve retail brands, along with a new strategy that hypermarkets can use when they find it difficult to get better results. Thus, brand equity is reflected in traditional categories: perceived quality; brand loyalty; brand awareness; and brand association.
In terms of its implication for management, this study contributes evidence about how particular factors can help create better brand equity by emphasising social responsibility, so that Spanish consumers' perception of the economic, ethical and legal, and discretionary C.S.R. aspects of the principal hypermarkets has a direct and significant influence on consumer-based hypermarket brand equity and thoughts about well-known dimensions. The final influence can be considered relevant, and it may be interesting that retailers may improve retail brands by using this social option, but it is also necessary to know which factors it originates from so that it is possible to manage them more effective and efficiently.
The gathering of such information can motivate companies and other organisations to dedicate their resources and efforts to strengthening their links with consumers and to achieving a better state of identification with their own organisation. Since this increases the proportion of consumers who positively identify with the company, it is expected that their commercial activity and sales figures will be less sensitive to changes in the environment and the marketing activities of the competition.
While this study has contributed to the branding literature, it does have certain limitations. It is necessary to emphasise that the scale used for measuring perceived C.S.R. and costumer-based brand equity in this research achieves the minimal criteria established for scales in the initial phases of development (suggested by Bagozzi and Yi [1988]), which allowed us to analyse the results with steadily increasing interest. It has also been possible to provide new ideas about the components, in order to stimulate deeper research into the concepts, efforts and investment that companies have to contribute in their efforts to become socially responsible. As regards limitations, the shoppers were from hypermarkets and the results may thus not be indicative of other types of retail store formats, such as convenience, speciality and department stores. Furthermore, the very transverse nature of the research prevents the identification of the variations in consumer perceptions which are studied across time, which suggests the need for expansion through a longitudinal study. It is also necessary to remember that the matching of the proposed model was only performed in a socio-cultural context, so it is may be necessary to carry out new forms of matching in other contexts.
Finally, it seems that, in spite of the limitations that have been mentioned, this study sheds new light on the body of knowledge that forms the basis of marketing and management and contributes positively to both disciplines. The results of the study also suggest possibilities for future research opportunities which could be carried out in other geographic locations and for other types of retail stores. Also, segmentation by income group should be a criterion for trying to find different forms of behaviour and identify casual relationships in new categories.
In short, a company can use C.S.R. as a competitive resource, since it involves favourable responses from consumers. We can conclude here that these repercussions also have a strategic character, in that they can generate greater competitiveness for hypermarket retailers.

Disclosure statement
No potential conflict of interest was reported by the authors.