Consumer-based brand equity of South African luxury fashion brands

Abstract

Purpose

South Africa is the leading market for luxury goods in Africa – a fact evident from the statistics on luxury retail and the expanding footprint of international and local luxury brands. In a market that is dominated by prominent international brands, indigenous South African brands are seldom the subject of empirical research. This study addresses this gap by analysing the consumer-based brand equity (CBBE) of South African luxury fashion brands and its outcomes on the purchase/repurchase intention of consumers of South African luxury fashion brands.

Design/methodology/approach

The study adopted quantitative research methods and utilized survey questionnaires to acquire data from 130 respondents. Structural equation modelling was used in testing the proposed alternative hypotheses.

Findings

The study affirmed the relevance of Aaker's (1991) CBBE model for luxury goods in the emerging economy of South Africa. It established perceived quality and behavioural loyalty as significant predictors of brand equity while affirming the prevalence of hedonism and behavioural loyalty in South Africa's luxury fashion market.

Research limitations/implications

The small sample size and the limited geographic scope of the study had a significant adverse impact on the broad application of the study's outcome. Furthermore, Aaker's (1991) CBBE model, while adequate, may have diminished the probability of a nuanced outcome.

Originality/value

This study advances the frontiers of interdisciplinary research by applying the marketing framework of CBBE to fashion studies in South Africa. The validated measurement scale, which emphasises the relevance of hedonism and behavioural loyalty in South Africa, may be useful for a similar study on luxury fashion brands in other emerging economies.

Citation

Appiah-Nimo, K., Muthambi, A. and Devey, R. (2023), "Consumer-based brand equity of South African luxury fashion brands", Journal of Fashion Marketing and Management, Vol. ahead-of-print No. ahead-of-print. https://doi.org/10.1108/JFMM-10-2021-0277

Introduction

The market for luxury goods in emerging economies has developed and Brazil, India, Russia, Mexico and mainland China have become important frontiers in the globalization of the luxury industry (Atwal and Bryson, 2014; Caïs, 2018; Diniz, 2014). In Africa, South Africa, Egypt, Morocco, Mauritius and Nigeria are leading destinations for luxury goods (KPMG, 2015). However, a review of the literature reveals a gap in the knowledge on the consumption of luxury goods in Africa. This study addresses this gap in luxury consumer behaviour in South Africa.

South Africa is the largest luxury goods market in Africa and boasts an impressive array of luxury hotels, private jets, yachts, automobiles and personal luxury goods such as leather products, clothing and accessories. The South African luxury goods industry is propelled by a market of some 36,500 millionaires or high-net-worth individuals and the influx of wealthy tourists and destination shoppers from Africa and other parts of the world (AfrAsia, 2021; Crosswaite, 2014). In 2017, South Africa accounted for US$2.2bn of the US$6.6bn worth of luxury goods consumed in Africa, making it an attractive launch pad for international luxury brands seeking a foothold in the emerging African luxury goods market (AfrAsia, 2018; Crosswaite, 2014; Euromonitor, 2019; Jacobs, 2013; Moorad, 2013).

Luxury is an elusive concept and difficult to define in universal terms. Across different markets, cultures and regional boundaries, it is a phenomenon that harbours subjective idiosyncrasies and has inspired extensive enquiries in marketing and consumer behaviour (Kernstock et al., 2017; Turenen, 2018). According to Turenen (2018), the subjectivity of luxury underscores the need for consumer-centric interpretations regarding the perceptions and experiences of luxury (Turunen, 2018). This research addresses the salient gap in how consumers navigate the competition between international and local brands and whether such preferences are influenced by differences in historical, cultural and socio-economic conditions (Roy and Chau, 2010).

In a market where status-seeking motivation is accomplished by appropriating the prestige of international brands, the limited literature on luxury consumption in South Africa is focused largely on the activities of international luxury brands (Euromonitor, 2019; Steinfield, 2015). However, the emergence of small but formidable independent South African luxury brands in recent years underscores the relevance of this study.

Brand equity is a popular concept in marketing and consumer behaviour and is explored as an invaluable asset with a competitive advantage (Keller, 2013). In the face of rapid globalization, prominent international luxury brands benefit from the country-of-origin effect and, thus, Europe's venerable heritage of luxury augments the prominence of European brands in emerging markets (Roy and Chau, 2010; Stoenescu et al., 2016; Turunen, 2018). The opposite is true for indigenous brands which are perceived to be of low quality and subdued brand equity (Atwal and Bryson, 2014; Cheah et al., 2020; Kapferer, 2002).

The literature on indigenous African luxury brands is nascent even in South Africa where the luxury goods industry has grown exponentially in the past few decades (Stiehler, 2016). This study addresses this gap by offering additional insight into South Africa's burgeoning luxury goods industry. It adopts Aaker's (1991) consumer-based brand equity (CBBE) model to examine the impact of CBBE variables – brand awareness, perceived quality, brand associations and brand loyalty – on brand equity and purchase/repurchase intention.

The findings of this study could impel the brand management and marketing strategies of South African luxury fashion brands towards growth and enhanced brand equity. The results of the study could equally inform the strategy of luxury fashion brands entering South Africa and other emerging economies.

Literature review

Consumer-based brand equity

Brand equity is broadly defined as the value endowed by the brand to the product (Farquhar, 1989). In other words, it is the perceptible or determinable value that is associated with the products and services of a firm by virtue of the brand (Aaker, 1991). Brand equity is explored extensively in marketing literature as it is essential to establishing market leadership, commanding price premiums and achieving highly effective marketing and communication strategies while protecting brands against unhealthy price competition from new category entrants (Aaker, 1991; Chahal and Bala, 2010; Keller, 2013; Motameni and Shahrokhi, 1998; Washburn and Plank, 2002).

Among the various approaches to determining brand equity, the consumer-centric approach of CBBE is widely adopted. Keller (1993), a leading proponent of CBBE, asserts that a brand's value is conceived in the mind of the consumer. CBBE is defined as the differential effect of brand knowledge on consumer response to the marketing of the brand (Keller, 2013). According to Aaker (1991), CBBE comprises brand assets and liabilities that are linked to a brand, its name and symbol and can improve or diminish the value of the product or service provided by a firm (Aaker, 1991).

The two foremost CBBE models were conceptualized by Aaker (1991) and Keller (1993). According to Aaker (1991), brand equity may result from five categories of brand assets (or liabilities) that include brand awareness, brand associations, perceived quality, brand loyalty and other proprietary assets (Aaker, 1991; Christodoulides et al., 2015; Washburn and Plank, 2002). Keller's (1993) brand resonance pyramid explored brand equity as the culmination of interdependent stages that include brand salience, brand performance and imagery, judgements and feelings.

Several approaches to measuring CBBE with varying degrees of specificity and scope have been suggested (Cobb-Walgren et al., 1995; Washburn and Plank, 2002; Yoo and Donthu, 2001). However, since its inception, Aaker's (1991) classic CBBE model has been widely adopted by various contemporary scholars across various product and services categories, national and cultural boundaries, marketing channels and sectors (Balaji, 2012; Buil et al., 2008; Christodoulides et al., 2006; Liu et al., 2017; Shukla and Purani, 2012).

Aaker's (1991) CBBE model was operationalized on a parsimonious scale by Yoo and Donthu (2001). The measurement scale based on the conceptualization of CBBE by Aaker (1991) and developed by Yoo and Donthu (2001) was adopted for this study as it adequately captured the psychometric properties of interest.

Model development and research hypotheses

Brand awareness and brand equity

Brand awareness is the ability of consumers to identify and recall a brand as a member of a specific product category (Aaker, 1991). According to Keller (2013), brand awareness falls within a continuum of low recognition to top-of-mind awareness (TOMA), with TOMA being critical to the success of brands (Hakala et al., 2012). However, despite being desired as the most fundamental of the four brand equity variables, brand awareness may not sufficiently predict brand equity (Aaker, 1991; Jin and Cedrola, 2017; Keller, 2013). The literature proposes a positive relationship between brand awareness and brand equity (Washburn and Plank, 2002). The following alternative hypothesis depicts this proposed relationship:

H1a.

Brand awareness of South African luxury fashion brands has a direct and significant influence on brand equity.

Perceived quality and brand equity

Perceived quality is the consumer's evaluation of a brand's quality in comparison to other competitive brands (Aaker, 1991). Quality perceptions may result either from a consumer's personal experiences, consumer reviews or the influence of marketing media (Jung and Shen, 2011). Quality is considered critical to the evaluation of luxury brands and may influence brand equity (Kapferer and Bastien, 2017; Vigneron and Johnson, 2017). Consequently, the literature proposes a direct and significant relationship between quality perception and brand equity (Buil et al., 2008; Esmaeilpour, 2015; Fritz et al., 2017; Keller, 2013; Low and Lamb, 2000; Sharma, 2017). The following alternative hypothesis supports this proposed relationship:

H2a.

The perceived quality of South African luxury fashion brands has a direct and significant influence on brand equity.

Brand associations and brand equity

Brand associations refer to the psychological and emotional imagery that is formed about a brand in the minds of consumers (Keller, 2013). According to Kapferer and Bastien (2017), luxury branding involves the creation of many intangible brand associations. Brand associations are, however, underpinned by antecedents expressed in the literature as value dimensions (Sharma, 2017).

This study examined brand association according to the luxury value framework by Wiedmann et al. (2007) which underscored the value dimensions of price, utility, quality, uniqueness, self-identity, hedonism, prestige, materialism and conspicuousness. The literature proposes a direct relationship between brand association and brand equity (Aaker, 1991; Keller, 2013), positing that brand equity is enhanced as consumers hold robust, favourable and unique brand associations (Bakshi and Mishra, 2016). This relationship is expressed as the following alternative hypothesis:

H3a.

Brand associations of South African luxury fashion brands have a direct and significant influence on brand equity.

Brand loyalty and brand equity

Brand loyalty refers to the consumer's intense attachment to the brand and their sincere commitment to patronize it despite changes in the brand's or consumer's situational factors and the marketing efforts of competitors (Aaker, 1991). Brand loyalty can be further categorized under attitudinal and behavioural loyalty (Thakur and Kaur, 2015). According to Appiah et al. (2016), attitudinal and behavioural loyalty may reflect a linear progression of the consumer–brand relationship.

Once established, brand loyalty creates a barrier to entry for competitor firms, provides the opportunity to apply price premiums, reduces marketing costs and protects the brand from competition (Sözer et al., 2017). Consequently, a direct relationship is proposed between brand loyalty and brand equity, with brand loyalty expected to improve brand equity (Riaz et al., 2014; Sharma, 2017). The following alternative hypothesis expresses this relationship:

H4a.

Brand loyalty towards South African luxury fashion brands has a direct and significant influence on brand equity.

Purchase/repurchase intention and brand equity

Purchase intention is defined as the consumer's consideration to acquire or patronize a brand (Choi et al., 2016). Consumer purchase intent is complex and subject to numerous influencing factors (Rieke et al., 2016). However, Spears and Singh (2004) opined that purchase intentions are personal action tendencies towards a brand that are underpinned by perceived value, brand associations and brand attitudes. Repurchase intention, on the other hand, is the tendency of existing consumers to acquire the product repeatedly after a favourable post-purchase evaluation (Japutra et al., 2014). Purchase and repurchase intentions were examined as a composite phenomenon, with a proposed direct relationship between brand equity and purchase/repurchase intention.

H5a.

Brand equity of South African luxury fashion brands has a direct and significant influence on the purchase/repurchase intention of consumers.

Proposed theoretical model

Based on the preceding discussion on the relationship between brand awareness, perceived quality, brand associations, brand loyalty, purchase/repurchase intention and brand equity, the theoretical model presented in Figure 1 graphically illustrates the relationships between the constructs and variables under investigation in this study.

Research methodology

Research design

This research adopted quantitative methods. Such an approach involves the collection, analysis and interpretation of statistical data (Saunders et al., 2016). Quantitative research uses data collection techniques, such as questionnaires, to generate numeric outcomes in the form of descriptive and inferential statistics. The quantitative method was adopted for this research due to the objectives of the research question, which sought to determine a cause-and-effect relationship between the CBBE variables.

Sampling and sample size

The study targeted luxury consumers in Johannesburg – South Africa's wealthiest city – which has a vibrant luxury retail industry (Crosswaite, 2014; Euromonitor, 2019; KPMG, 2015). The study made use of quota sampling; this technique ensures that the sample is equitably distributed and representative of particular categories of respondents (Clow and James, 2014; De Vos et al., 2017). Quota sampling enabled the sampling of an almost equal number of respondents who were patrons and non-patrons of South African luxury fashion brands. This was premised on the need to acquire insights into the attitudes of both segments towards South African luxury fashion brands.

A total of 130 responses were acquired from the initial 220 disseminated surveys. These responses represented a sample realization rate of 65 per cent. Despite the shortfall in the sample realization rate, the acquired sample was deemed sufficient in the context of this often discreet and inaccessible luxury consumer segment. Furthermore, the final sample size was sufficient to conduct inferential statistical analysis.

Data collection

The study utilized electronically distributed self-administered surveys. The measurement scale and medium of distribution were cost-effective and suitable for reaching South African luxury brand consumers. The questionnaire was piloted and refined before dissemination. It comprised 53 items grouped under six subscales and measured the demographics of the respondents, the four predictor variables of brand equity – brand awareness, perceived quality, brand association and brand loyalty – and the outcome variable purchase/repurchase intention.

For example, brand awareness was measured by the following three items: 5.1 To what extent do you know South African luxury fashion brands? 5.2 To what extent can you identify South African luxury fashion brands among competing brands? and 5.3 I can identify the symbols and/or logos of South African luxury fashion brands.

Each item evaluated the respondent's knowledge of local brands as well as their ability to recognize and recollect the brands' associated logos and aesthetics along an extent scale of 1 to 5, where 1 implied “To no extent” and 5 implied “To a full extent”. The measurement scale was adapted from pre-tested scales from works by Christodoulides and De Chernatony (2010), Washburn and Plank (2002), Yoo and Donthu (2001) and Zhang and Kim (2013). The scale yielded Cronbach's alpha scores of above 0.75, suggesting the satisfactory reliability of the measuring instrument (Buil et al., 2008; Christodoulides et al., 2015; Dodds et al., 1991; Pappu et al., 2005; Washburn and Plank, 2002; Yoo and Donthu, 2001; Zhang and Kim, 2013).

The data collected from the respondents were also subjected to a common method bias test. The common method bias test seeks to determine whether the relationship observed between constructs is biased due to being measured using the same methodology (Jordan and Troth, 2019). In the study, Harman's single-factor test was used to test for common method bias. In Harman's single-factor test, factor analysis is run using all the items used in the model and thereafter forced into a single factor from the set of items (Aguirre-Urreta and Hu, 2019; Jordan and Troth, 2019).

According to Aguirre-Urreta and Hu (2019), if the percentage of variance for this single factor is less than 50 per cent, the common method bias did not affect the data. In the study, all the items in the model were subjected to Harman's single-factor test, and the percentage of variance amounted to 46 per cent when the items were forced into a single factor. With the percentage of variance lower than 50 per cent, common method bias is not a cause for concern for the data set used in the study for the inferential statistics.

The data collected from the respondents were also evaluated for a non-response bias to determine if the responses of the respondents who answered the questionnaire first are similar to the respondents who answered the questionnaire later. According to Bańkowska et al. (2015), non-response bias arises when the data collected from the respondents would have differed if the individuals who did not respond to the survey had responded.

To evaluate the data set for the non-response bias, the respondents were split into two groups. For the first group, the respondents were split in half and, for the second group, the first 30 respondents and the last 30 respondents were selected. Thereafter, the demographic characteristics of the two groups were analysed to evaluate non-response. No significant association between the early and late respondents was found; the responses flowed in consistently throughout the period during which the electronic questionnaire was open for completion.

Demographics of the respondents

In Table I, the demographics of the respondents are presented.

Data analysis

IBM SPSS Version 26 and AMOS Version 26 were used to analyse the data. Before determining the most appropriate inferential statistical tests for the study, a test of normality was conducted using the Shapiro–Wilk and Kolmogorov–Smirnov tests. The results of the Kolmogorov–Smirnov statistics – ranging between 0.117 and 0.250 (p < 0.05) – and Shapiro–Wilk test scores – between 0.896 and 0.939 (p < 0.050) – indicated a skewed data distribution which necessitated the adoption of non-parametric comparative tests including the Chi-square, Mann–Whitney U and Kruskal–Wallis tests (Woodrow, 2014).

Factor analysis was conducted on the subscales measuring the four CBBE predictor variables. A series of factor reduction procedures, exploratory factor analysis (EFA) and confirmatory factor analysis (CFA), reduced the subscales in the proposed model to six factors that were renamed as brand awareness, brand association (hedonic), perceived quality, brand loyalty (attitudinal), brand equity and (re)purchase intention.

Lastly, structural equation modelling (SEM) was undertaken to test the hypothesis. SEM was adopted for its ability to capture errors and ensure accuracy in estimating relationships, both obvious and latent, among variables (Byrne, 2013). SEM included various factor reduction procedures – EFA and CFA, tests of reliability and validity, regression analysis and tests of model fit using the Chi-square difference test (χ²).

The EFA and CFA procedures reduced the items in the measurement model from 42 to 29 by eliminating items with low R² (<0.4) and those with a high contribution to skewness. The new measurement model was examined with the maximum likelihood-based standardized root mean squared residual (ML-SRMR) and supplemented with the root mean squared error of approximation (RMSEA) (Hu and Bentler, 1999). The initial proposed model (χ² = 700.905; df = 436; p = 0) displayed an unsatisfactory fit index of normed fit index (NFI) = 0.782, non-normed fit index (NNFI) = 0.888, comparative fit index (CFI) = 0.902 and RMSEA = 0.069, prompting further remedial procedures that yielded a final measurement model of χ² = 327.177, df = 194, p = 0, with fit indices of χ²/df = 1.686, NFI = 0.856, NNFI = 0.923, CFI = 0.935 and RMSEA = 0.077.

The fit indices for the final model were greatly improved, albeit, exhibiting an RMSEA value greater than 0.06. However, the discrepancy in the RMSEA value was due to its sensitivity to the small sample sizes as RMSEA tends to over-reject true population models with small sample sizes (Hu and Bentler, 1999). However, Browne and Cudeck (1992) suggest that in circumstances where the sample size of a true population model is small, an index of RMSEA ≤ 0.08 is indicative of an acceptable fit.

Hypotheses testing

The study used robust statistics to determine the significance of the hypothesized relationships between the factors. The first phase of the analysis examined the relationship between the four predictor variables – brand awareness, perceived quality, brand association (hedonic) and brand loyalty (behavioural) – and the intermediate factor – brand equity. The robust test results indicated that the relationship between two of the four predictors and brand equity was significant. The two factors were perceived quality and brand loyalty (behavioural), with robust statistics of 5.435@ and 6.549@, respectively.

The second phase of the analysis examined the relationship between the four predictor variables (brand awareness, perceived quality, brand association (hedonic) and brand loyalty (behavioural)), the intermediate variable (brand equity) and the outcome variable ((re)purchase intention). The robust results indicated that the relationships between three variables and (re)purchase intention were significant. These were brand awareness (4.571@), perceived quality (2.284@) and brand equity (2.545@). The robust test results are presented in Table II.

Table III summarizes the structural model, presenting the direction of relationships and path coefficients. The path coefficients indicate the strength of correlations and the level of significance of the proposed relationship.

Validity and reliability

The new constructs from the factor analysis were subjected to reliability and convergent and discriminant validity tests. Cronbach's alpha coefficient for the new factors ranged between 0.59 and 0.93, suggesting satisfactory reliability. The research also adopted the average variance extracted (AVE) approach to determine convergent validity, with scores ranging between 0.53 and 0.98, indicative of adequate convergent validity among the constructs. Furthermore, the condition for discriminant validity was fulfilled, with the AVE being greater than the maximum shared variance (MSV) and average shared variance (ASV) (AVE > MSV and ASV) (Hair et al., 2019).

Ethics

This study addressed the fundamental ethical concerns of acquiring informed consent, protecting respondents from harm and maintaining confidentiality. The research proposal was scrutinized by the ethics committee of the Faculty of Architecture, Design and Arts of the University of Johannesburg to ensure that ethical concerns were adequately addressed. The electronic survey also included a synopsis that explained the purpose of the study and provided the respondents with the right to withdraw from the study had they wished to do so. Furthermore, the data collected were anonymous, thus, guaranteeing the confidentiality of the respondents.

Results and discussion

Results of hypotheses testing

The final stage of the data analysis examined the relationships between the factors included in the theoretical model using SEM. The outcomes of the correlation coefficients, significant at p = 0.05, were used to test the alternative hypotheses. The results are summarized in Table IV and discussed thereafter.

Discussion of empirical findings

H1a tested the relationship between brand awareness and brand equity. Despite brand awareness having a positive correlation with brand equity, this relationship was not significant at β = 0.01, p > 0.05; hence, the alternative hypothesis was rejected. This outcome indicates that despite being fundamental, brand awareness is not a sufficient determinant of brand equity in the sample population. It is consistent with the observation in the literature that brand awareness was insufficient to establish brand equity (Aaker, 1991; Hakala et al., 2012; Konecnik and Gartner, 2007).

Several factors may explain this result. First, the consumption of luxury goods requires deeper knowledge structures and discretionary purchasing power (Celik and Erciş, 2018). Consequently, the progression from brand awareness to brand loyalty is protracted and mere awareness may not guarantee patronage or establish brand equity. Second, the pervasive phenomenon of classifying European brands in South Africa as luxury brands undermines the appreciation of local South African luxury brands, resulting in limited awareness (Stiehler and Tinson, 2015). Mutunku (2016) observed this phenomenon among Paris-based Congolese sapeurs who were adamant about associating luxury with brands of African origin.

H2a tested the correlation of perceived quality with brand equity. This relationship was significant at β = 0.38, p < 0.05. Hence, the alternative hypothesis was accepted. The observed significant relationship is consistent with the findings of Novotová (2016); Sharma (2017) and Yoo and Donthu (2001) who proposed a strong correlation between brand luxuriousness, quality perceptions and brand equity. According to Pappu et al. (2005), perceived quality delivers value by differentiating a brand from the competition and encouraging patronage, thus, underscoring the positive correlation between perceived quality and brand equity.

H3a examined the relationship between brand association and brand equity. A positive correlation was observed between brand association (hedonic) and brand equity. This relationship was nonetheless insignificant at β = 0.11, p > 0.05. Hence, the alternative hypothesis was rejected. This result indicates that brand association is a complex construct to measure because it varies among individuals of different socio-cultural and economic contexts.

Hedonism, consistent with the pleasure-seeking orientation of luxury consumers, emerged as the most relevant value dimension to South African luxury fashion consumers (Choo et al., 2012). The literature emphasizes the emotional dimension, such as sensory pleasure and gratification, aesthetic beauty or excitement, of luxury consumption (Vigneron and Johnson, 2017; Wiedman et al., 2007). While the literature supports a weak correlation between brand association and brand equity (Washburn and Plank, 2002), the correlation between hedonism and brand equity is not explored. However, hedonism as a subjective construct prioritizes self-gratification above loyalty and may explain the observed outcome.

H4a examined the relationship between brand loyalty and brand equity. The result established a positive correlation between brand loyalty and brand equity with a path coefficient of β = 0.62. This relationship was significant at p < 0.05. Hence, the alternative hypothesis was accepted. The observed outcome is consistent with the literature, with behavioural loyalty emerging as the dominant attribute among the sample population. Aaker (1991) and Yoo et al. (2000) propose a positive correlation between brand loyalty and brand equity, suggesting that a consumer's regular patronage of a brand discourages the tendency to switch. Bahri-Ammari et al. (2016) observe that behavioural loyalty is a critical antecedent to the brand equity of high-value brands and guarantees premium-related outcomes such as price.

H5a examined the relationship between brand equity and (re)purchase intention. The results established a positive relationship between brand equity and (re)purchase intention. This relationship was significant (β = 0.48, p < 0.05). Hence, the alternative hypothesis was accepted. Brand equity was established in the literature as dependent on the four predictor variables of CBBE according to Aaker's (1991) CBBE model (Yoo and Donthu, 2001). While improving brand equity is prescribed, its outcomes in terms of behaviours such as patronage and purchase intentions are even more desirable (Cobb-Walgren et al., 1995).

The literature on consumer behaviour and theories such as the reasoned-action theory suggest that favourable brand equity will positively influence a person's behavioural intentions towards the brand (Rieke et al., 2016). Consequently, the direct and significant correlation between brand equity and (re)purchase intention affirms the convention in extant literature (Aaker, 1991; Hung et al., 2011; Yoo and Donthu, 2001).

Conclusions and implications

The typical respondent in this study was a Black African female aged between 25 and 34 years, possessing an undergraduate degree and earning a disposable income between R20,000 and R39,000. This supports observations in contemporary literature such as the prominence of female luxury consumers and the emergence of a relatively younger luxury consumer demography (Stokburger-Sauer and Teichmann, 2013). The findings have significant implications for the marketing of luxury goods in terms of identifying the appropriate communication channels and emphasizing the right messages that resonate with the youthful demography. For example, influencer marketing on social media may prove most effective in reaching this demography.

The study failed to establish a significant correlation between the ethnicity and level of education of respondents and their patronage of South African luxury fashion brands, thereby discouraging the need for targeted communication or marketing based on ethnicity or level of education. It was observed, however, that age, gender and disposable income had the most significant impact on the patronage of South African luxury fashion brands.

The literature has emphasized conspicuousness or status-seeking as the primary motivation for luxury consumption in South Africa – ranked among the most unequal economies in the world (Goldberg, 2011; Stiehler, 2016). However, this study identified the hedonistic or pleasure-seeking motive as the value dimension most relevant to South African luxury fashion consumers. This observation is consistent with the pleasure-seeking orientation of luxury consumers (Choo et al., 2012), more so for the Millennial and Gen Z consumer demography who love to indulge as a form of self-reward (Cho et al., 2022). This is a significant finding because it challenges the previous perception of conspicuousness to adequately inform the strategy of both indigenous and international luxury brands. Consequently, brands should rather emphasize engaging narratives, products and services that stimulate a pleasurable experience for customers.

Additional conclusions were drawn from the results of the hypothesis testing. For example, despite the positive correlation between brand awareness, perceived quality, brand association, brand loyalty and brand equity, only the influences of perceived quality and brand loyalty on brand equity were significant. The significance of perceived quality on brand equity underscores the need for the highest standards of craftsmanship and product quality. It is equally essential for brands to employ strong visual language in communicating product quality and craftsmanship. Quality perceptions may equally be enhanced with the right price and brand and product positioning strategies (Aaker, 1991; Keller, 2013). To this end, a premium pricing strategy is prescribed in the literature (Kapferer and Bastien, 2017; Miller and Mills, 2017; Vigneron and Johnson, 2017).

The weak correlation between brand awareness, brand associations and brand equity may be the result of inadequate or weak marketing and communication strategies and the limited retail footprint of indigenous luxury brands in the face of the closures of leading multi-brand stores and stockists of South African luxury brands. South African luxury fashion brands can deepen brand awareness by pursuing an elaborate and cohesive communication strategy that is targeted at the ideal consumer demography.

Lastly, the brand equity of South African luxury fashion brands directly influences consumer purchase intention. This conclusion also has significant implications for the operational strategy of South African luxury fashion brands, with efforts towards brand building likely to be rewarded with (re)purchase intention, resulting from behavioural loyalty. South African luxury fashion brands are encouraged to invest in long-term strategies that build brand equity by enhancing the equity variables of brand awareness, perceived quality, brand associations and brand loyalty.

Limitations and recommendations for future studies

The choice of the theoretical model constituted a major limitation for this study. First, the CBBE model conceptualization by Aaker (1991) and its operationalization by Yoo and Donthu (2001) are generic and not specific to luxury brands. However, this limitation was addressed by adapting the brand association dimension of Wiedmann et al. (2007) to incorporate the luxury value framework. Conventional CBBE measurement scales are also criticized for evaluating the equity of a brand as though it were immune from competition – hardly the case in the real world (Porto, 2018).

Furthermore, the recent conceptualization of CBBE suggests a dynamic and sufficiently complex phenomenon which differs from the classic conceptualizations of CBBE as a hierarchical and linear construct (Chatzipanagiotou et al., 2016). The classic CBBE model by Aaker (1991), while relevant, may be inadequate for evaluating the brand equity of luxury goods in emerging markets.

At a sample realization rate of 65 per cent, the findings of the study cannot be generalized to South Africa's luxury goods market. The small sample size was occasioned by time constraints coupled with the difficulty in acquiring qualified respondents from the discrete luxury consumer demography niche. The low sample realization rate is also typical of studies that adopt respondents from a non-student population. Nevertheless, the acquired data was significant enough to undergo the requisite statistical analysis and offered sufficient insight into the dynamics of luxury consumption in South Africa.

Despite the aforementioned limitations, opportunities for further research on local luxury fashion brands in South Africa and other emerging markets around the globe exist. Since this study covered a small population, a comprehensive study on the South African luxury goods market with a wider geographical scope and an increased sample size is recommended. Future research could evaluate the CBBE of competing brands – both local and international – in specific segments, mediated by status-seeking behaviour, hedonism and consumer ethnocentrism for a holistic insight into brand performance in a real-world competitive marketplace.

Lastly, the methodology for future studies could incorporate CBBE conceptualizations by Chatzipanagiotou et al. (2016), Christodoulides and De Chernatony (2010) and Wang and Finn (2014) to capture the additional cognitive and emotional dimensions associated with the consumption of luxury goods.

Figures

Figure 1.

CBBE model

Table I.

Demographics of the respondents

Respondent characteristics	Category	Valid percentage (%)
Age group	15–24	24
	25–34	47
	35–49	26
	50+	3
Sex	Male	38
	Female	62
Ethnicity	Black African	47
	White	32
	Coloured	9
	Other	12
Education	Postgraduate degree	24
	Undergraduate/bachelor's degree	52
	Diploma/Certificate	18
	Grade 12 and lower	6
Disposable income	Less than R10,000	19
	R10,000–R19,999	15
	R20,000–R39,999	25
	R40,000–R59,999	24
	R60,000 and above	8
	Missing	9

Table II.

Robust test results for SEM

F5 = F5 = 0.008*F1. + 0.362*F2 + 0.113*F3 + 0.761*F4
	0.051	0.070	0.078	0.115
	0.161	5.157@	1.449	6.626@
	(0.047)	(0.067)	(0.075)	(0.116)
	(0.173)	(5.435@)	(1.500)	(6.549@)
+ 1.000 D5
F6 = F6 = 0.464*F5 + 0.357*F1 + 0.241*F2. + 0.156*F3
	0.147	0.073	0.095	0.089
	3.153@	4.851@	2.538@	1.739
	(0.182)	(0.078)	(0.105)	(0.081)
	(2.545@)	(4.571@)	(2.284@)	(1.928)
0.159*F4 + 1.000 D6
0.153
–1.043
(0.184)
(−0.868)

Notes: Statistically significant at the 5 per cent level are marked with @; Robust statistics is given in parentheses ()

Table III.

SEM with path coefficient and p-values

Structure paths			Path coefficient	Level of significance
Brand equity	←	Brand awareness	0.01	–
Brand equity	←	Brand quality	0.38	*
Brand equity	←	Brand association – HV	0.11	–
Brand equity	←	Brand loyalty – BL	0.62	*
(Re)purchase intention	←	Brand equity	0.48	*
(Re)purchase intention	←	Brand awareness	0.53	*
(Re)purchase intention	←	Brand quality	0.26	*
(Re)purchase intention	←	Brand association – HV	0.15	–
(Re)purchase intention	←	Brand loyalty – BL	−0.14	–

Table IV.

Results of hypotheses testing

Alternative hypotheses		Robust statistics	Finding
H1a	Brand awareness of South African luxury fashion brands has a direct and significant influence on brand equity	0.17	Rejected
H2a	The perceived quality of South African luxury fashion brands has a direct and significant influence on brand equity	5.44	Accepted
H3a	Brand associations of South African luxury fashion brands have a direct and significant influence on brand equity	1.50	Rejected
H4a	Brand loyalty towards South African luxury fashion brands has a direct and significant influence on brand equity	6.55	Accepted
H5a	Brand equity of South African luxury fashion brands has a direct and significant influence on the purchase/repurchase intention of consumers	2.55	Accepted