Linking brand engagement to customer-based brand equity and role of brand experience, brand personal- ity, and brand affect: A case of automobile market of Pakistan

Article history: Received: February 2, 2020 Received in revised format: March 11 2020 Accepted: March 11, 2020 Available online: March 11, 2020 Brand engagement is relatively a new concept in marketing. The objective of this research study is to unearth the relationship between brand engagement and customer-based brand equity drivers; brand personality, brand experience, and brand affect. The data was collected from 426 automobile customers in Pakistan through multistage cluster sampling technique. PLS-SEM analysis of the data supported the relationships between brand equity drivers and brand engagement and the mediating role of brand affect. Brand experience, brand personality, and brand affect emerged as the significant predictors to brand engagement. It is recommended that brand managers design marketing plans with a greater emphasis and focus on pleasant brand experiences which result in brand affect, ultimately leading to brand engagement. © 2020 by the authors; licensee Growing Science, Canada


Introduction
Globally, the automobile industry has witnessed a shift of power from the west to east (Bernhart, Kleimann, & Hoffmann, 2011) as many countries in Asia have become major automobile manufacturers as well as these countries have a substantial share in the aggregate demand of the automobile market. The growing demand for automobiles in Asia has also identified the importance of leveraging customer-based brand equity (CBBE) amidst the burgeoning pressure on the automobile manufacturers due to increasing competition in the automobile industry (Adetunji, Rashid, & Ishak, 2018). The mounting competitive pressures on the automobile brands in Asian markets have also identified the falling retention rate of automobile customers due to various factors such as preference for higher product quality expectations, better customer experience and due to social factors like growing desire for automobile brands that raise the self-image of customers and help them look distinctive (Edmunds, 2018;Gerrits, Zhang, Klotz, Xu, & Xie, 2014). Such a preference shift of customers for automobile brands demands to design marketing strategies to raise brand engagement of automobile brands and build a long-lasting customer-brand relationship through superior brand experience (Placing customer centercity, 2014;Skier, 2017). As increasingly customers are moving away from loyalty, brand engagement has emerged as the strategy to build competitive advantage (Chan et al., 2014) and as a manifestation of enduring loyalty (Keller, 2001). Extant literature indicates the sheer paucity of studies that assessed the role of brand experience and brand personality in building brand engagement, especially in the context of automobile brands. Responding to the calls of (Lemon & Verhoef, 2016) this study aims to investigate how brand experience and brand personality influence brand engagement and whether brand affect mediates these relationships, specifically for automobile brands. Since brand experience is termed as the new battlefield for automobile brands (Skier, 2017), it is imperative for the Asian automobile manufacturing countries such as Pakistan to sense and respond to the transforming needs of customers.

Automobile Market in Pakistan
The automobile sector in Pakistan is the second biggest revenue generator after oil and gas (Arifeen, 2018). It is also estimated that the automobile sector contributes to the country's economy by offering 40,000 direct jobs and 360,000 indirect jobs through thousands of vendors that serve the needs of the automobile manufacturers in the country (Hashmi, 2014). But the automobile market in the country has witnessed a major shift in the customers' preference for the reconditioned car brands imported from countries like Japan over the locally produced global brands. Essentially, the automobile customers in Pakistan have shown a growing tendency to choose the reconditioned cars brands due to their growing dissatisfaction with the locally produced global brands (Amir & Asad, 2018). In last 10 years, the demand for the reconditioned cars have significantly increased and only in the year 2016-2017, the number of reconditioned cars imported into the country was 65,000 in comparison with 185,000 locally produced cars in the same time frame.
Owning to the rapidly growing preference for the imported reconditioned car brands in Pakistan, the market share of the locally produced global brands reduced to 73% in 2018 from 92% in 2011. According to a conservative estimate, only in the year 2017, the worth of the imported reconditioned car brands was around 16 billion PKR which represents the lost sales of the locally produced cars (Khan, 2017). Even though, over the years, the government has introduced stringent policies to curb the import of the reconditioned car brands into the country, the preference for the reconditioned imported car brands has emerged a challenge for the local manufacturers of the global automobile brands in the country. It essentially demonstrates that customers who were earlier dependent on the locally produced global brands, now see a higher value in the imported reconditioned car brands and as a result, have demonstrated dwindling brand loyalty to the locally produced global automobile brands.
Even though, a serious scarcity of research literature exists on the subject, a few studies in the past have contended that the automobile customers in Pakistan prefer the imported reconditioned car brands for the higher quality, style, image, safety, and overall better experience that the reconditioned car brands offer that locally produced global brands lack (Mehmood, 2015;Zia & Sohail, 2016). Likewise, the mounting influx of the reconditioned automobile brands have posed serious challenges to the local automobile manufacturers and the country's economy (Khan, 2017;Zaheer, 2012). For instance, it is well acknowledged that purchasing personal vehicles such as cars are emotions-ladder decisions for customers in Pakistan (Flop gear, 2015;Hanan, 2016) and elsewhere in Asia (The heart of, 2014). Emotions are at the center of the decision making of cars purchase in Pakistan for personal vehicles not only influence the social status (Agence France Presse [AFP], 2015) but also reflect the struggle one has to endure to finally have access to the personal vehicle (Flop gear, 2015). Nonetheless, extremely few studies in the past have examined how emotions play out in the decisions to purchase and repurchase personal vehicles in Pakistan and in South Asia. Despite the grave consequences of the declining preference and resulting sales of the locally produced automobile brands in the country, little attention has been paid by the earlier researchers to examine the potential factors behind this phenomenon. It demands research studies on the subject to reveal the factors that contribute to the growing preference of imported reconditioned automobile brands and the slump in the loyalty of the locally produced global brands in Pakistan, also known as semi knocked down kits (SKDs). Thus, the primary motivation of this study lies in the sheer paucity of the empirical studies to identify and the potential factors that contribute to the declining brand loyalty of the locally produced global automobile brands in Pakistan. Determining such factors is of high significance to understand what factors may influence the perception of the customers who prefer the reconditioned car brands that has caused the SKDs losing market share and the country's economy losing millions of US dollars. Since a handful of studies exist on this subject, this research is guided by the CBBE theory to explore and test the possible form the customer-based brand equity of the automobile brands sold in Pakistan. Accordingly, CBBE theory explains the role of image (personality) and consumer perception (Keller & Lehman, 2006) in brand engagement.

Theoretical Background
CBBE is the most influencing brand concept in the past two decades. The base of CBBE is found in the economics of information and cognitive psychology (Aaker, 1992;Keller, 1993). Many marketing experts have helped evolve the development of the conceptualization of the CBBE theory. Amongst these are Farquhar (1989), Srivastava and Shocker (1991), Aaker (1991Aaker ( ,a,b, 1996, Keller (1993) Lassar, Mittal, and Sharma (1995) and Yoo, Donthu, and Lee, (2000). Farquhar (1989) viewed CBBE as the incremental value that the name of a brand adds to a product or service which is embodied in the customer's response towards the brand and the final performance of the brand. Keller (1993Keller ( , 2001 extended the same perspective of CBBE theory and conceptualized CBBE on the basis of brand and brand associations. The author argued that CBBE is the favorable association of customers with brands in response to marketing efforts of a brand in comparison with the marketing efforts of competing brands. Kellerr (2001) expended the CBBE theory by placing its bases on six building blocks namely, brand salience (identity), brand performance, brand imagery, brand judgments, brand feelings, and brand resonance. Brand resonance entails the highest degree of brand loyalty also known as brand engagement, according to Keller (2016). Keller (2001) viewed brand engagement as a volunteer behavior of customers towards brands embodied in investing time, money and other resources beyond purchase and consumption. In essence, brand engagement is the behavioral manifestation of customers beyond brand purchase reflected through customers' active participation in brand-related activities, interacting with fellow brand users and advocating the brand ( Doorn et al., 2010;Keller, 2013). The foundation of brand engagement also lies in the relationship marketing theory (Vivek et al., 2012) which primarily embodies the interaction and interplay of customers and brands and the long-term customer-brand relationship. Similar to CBBE, relationship marketing emphasizes to leverage a firm's resources to inculcate brand loyalty (Thurau, Gwinner, & Gremler, 2002;Yanfei & Yafeng, 2012) which ultimately results in brand engagement behavior (Reichheld, 2003). Moreover, the relationship marketing theory emphasizes the role of various factors that stimulate brand engagement which holds significance for marketing managers who strive to invest customer -brand relationships (Ashley, Noble, Donthu, & Lemon, 2011). Alternatively, brand engagement the measurement of the close relational bond between brands and customers manifested by relational exchanges which often results in customers' participation in various brand-related activities and in continuation of transactions (Sashi, 2012). This study views brand engagement is the behavioral manifestation of customers towards brands by contributing their time, money, energy, and other resources beyond the purchase process, as a consequence of motivational drivers and loyalty (Doorn et al., 2010;Keller, 2003). Brand engagement may also be viewed as the brand-customer convergence in which both collaborate and create value for each other (Prahalad & Ramaswamy, 2004) in the context of the relationship marketing theory. The technological advancement such as web 2.0 has a major role in the co-creation behavior of customers (Harrison & Barthel, 2009) as the internet has enabled customers to engage with their favorite brands from the comfort of their home. For instance, online brand communities are the virtual places which enable customers to become a member of the brand community, participate and express their opinion to the fellow brands enthusiasts (Fırat & Dholakia, 2006). In essence, brand engagement is the behavioral response of customers towards brands that extends beyond, brand awareness, purchase, satisfaction, emotional connection, and even behavioral loyalty. Brand engagement represents the extension of the marketing concept from marketing orientation to relationship marketing (Sashi, 2012).

Research Framework and Hypotheses Development
Building upon the preceding discussion of the CBBE and the relationship marketing theories, the research framework of this study has been developed accordingly, as illustrated in Fig. 1. The research framework of this study presents the brand equity drivers; brand experience and brand personality as antecedents to brand engagement through the mediating role of brand affect. For instance, prior research studies have reported that brand personality and brand experience are significantly connected to brand engagement behavior through the mediating role of brand affect (Lee, Back, & Kim, 2009;Lyu, Mao, & Hu, 2018). Similarly, brand affect also significantly leads to brand engagement behavior (Samala & Singh, 2019;Verleye, Gemmel, & Rangarajan, 2014). Thus, in the current study, brand personality and brand experience represent the cognitive factors, brand affect as the emotional factor, and brand engagement as the behavioral response of the cognitive and emotional factors. The main argument that the research framework of this study resides on that brand experience and brand personality alone cannot guarantee brand engagement behavior. Instead, brand experience and brand personality are influential in inciting the brand engagement behavior through the mediation effect of brand affect (positive emotional response), as has been proposed by earlier studies (Franzak, Makarem, & Jae, 2014;Pansari & Kumar, 2017). Similarly, brand affect has been argued to better predict the behavior compared to solely cognition (Allen, Machleit, & Kleine,1992). Hence, this study contends that brand affect as a mediator can better explain the earlier identified relationship between brand engagement and brand equity drivers; brand experience and band personality. Especially, in the context of the automobile customers in Asia and Pakistan where the emotional trigger is an important force that shapes their purchase decisions (Nielsen, 2014; The heart of, 2014)

Relationship between Brand Personality and Brand Affect
Brand personality is the "human traits attached to brands (Keller, 2013). Supporting the same conceptualization of brand personality, Azoulay and Kapferer (2003) defined it as "the human personality characteristics are practical and relevant for brands." Various studies in extant literature indicate a logical relationship between brand personality and brand affect. Conceptualizing brand personality as one of the essential dimensions of the customer-based brand equity, Aaker (1991) proposed various dimensions of brand personality. It is also viewed as a major influencer towards customers' purchase of automobile brands (Brunello, 2015). Supporting the same, Keller (2001) argued that brand personality represents the brand meaning which eventually leads to emotional (affection) response of customers towards brands. Kim and Zhao (2014) evaluated if brand personality predicted the variability in brand affect and then in behavioral loyalty. The authors reported a significant relationship between brand personality and brand affect, ultimately leading to the loyalty behavior. Assessing the same relationship, Sung and Kim (2010) investigated if various brand personality types such as ruggedness, excitement, competence, sincerity, and sophistication explained brand affect and resulting brand loyalty for various product categories. The researchers concurred that the types of brand personality such as ruggedness and sophistication significantly explained brand affect and brand loyalty. Thus, this study hypothesizes: H1: Brand personality influences brand affect for the automobile brands in Pakistan.

Relationship between Brand Experience and Brand Affect
CBBE emphasizes that brand equity is the outcome of brand associations and the strength of brand associations is realized by brand experience (Aaker, 1991). Additionally, brand experience, according to Keller (2001), is the distinctive personal experience offered by brands to customers. The author explained that brand experience may stem from the personal accounts of interacting, using a brand or from the brand experience of friends and relatives which eventually builds brand affect. This study views brand experience as "subjective, internal response of customers prompted by the brand-related stimuli such as brand's identity, packaging, design, environment and communications (Brakus, Schmitt, & Zarantonello, 2009)". In the quest that if brand experience evokes positive and negative emotional response (affect), Lyu et al. (2018) conducted a research study to examine this relationship. Collecting the data from Chinese tourists, the authors reported that cognitive experience is critical in evoking brand affect in customers. Likewise, Huang (2017) also investigated if the CBBE driver; brand experience had an impact on brand affect and the following loyalty behavior in Taiwan. The results of the study indicated the sensory, intellectual and behavioral experiences significantly explain brand affect. The positive impact of brand experience in building brand affect was also reported by Pullman and Gross (2004) who examined the trio of experience -affect -loyalty. The authors contended that brand experience largely helps shape the emotional response of customers -brand affect. Thus, this study hypothesizes that: H2: Brand experience influences brand affect for the automobile brands in Pakistan.

Relationship between Brand Affect and Brand Engagement
Brand affect represents the positive emotional response of customers towards a brand. Chaudhuri and Holbrook (2001) defined brand affect as the ability of a brand to evoke positive emotions of customers after the use of the brand. The CBBE theory conceptualized by Keller (2001) referred to the same "emotional response" of customers as feelings that may be intensive, mild, negative and positive emotions. In fact, brand affect is a comprehensive construct that represents brand love (Albert & Merunka, 2013), brand passion (Füller, Matzler, & Hoppe, 2008), emotions (Lin & Lee, 2012) and affection (Halaszovich & Nel, 2017). Similar to the definition of brand affect by Chaudhuri and Holbrook (2001), feelings were explained as "emotional response of customers" by Keller (2013) and was argued to function as an essential factor that leads to brand engagement. Thus, brand affect represents the concept of feelings in CBBE theory (Lin & Lee, 2012). Brand affect is defined as the ability of a brand to elicit positive emotional responses in customers as a consequence of using the brand. Previous studies have explored the relationship between brand affect and brand engagement. For instance, Samala and Singh (2019) reported that Millennials developed brand affect towards fashion brands which ultimately shapes their brand engagement behavior. The researchers argued that brand is a significant predictor of brand engagement behavior. Placing brand affect in an integrative research framework of antecedents and brand engagement behavior, Xie et al. (2015) found that brand affect significantly influences the brand engagement behavior. In the same vein, (Verleye et al., 2014). conducted a research study in the service sector and reported that brand affect positively instills brand engagement behavior such as feedback, helping other customers, positive word of mouth and cooperation. Based on this discussion, it is hypothesized that: H3: Brand affect influences brand engagement for the automobile brands in Pakistan.

Brand Affect as a Mediator
The CBBE model of Keller (2001) emphasized both cognitive and emotional factors that collectively determine the highly in sync customer-brand relationship manifested in brand engagement. The author has concurred that perceived brand performance (cognitive judgments) leads to brand affect (positive emotional response) which ultimately results in brand engagement (true brand loyalty) (Keller, 2013). Importantly, brand personality and brand experience are the customer-based brand equity drivers that build brand affect and brand engagement. Brand personality is an essential building block of CBBE and reflects the perceived meaning of a brand a customer has which results in affective evaluations related to the brand and in turn raises the relationship between brand and customers to the highest degree of loyalty -brand engagement (Keller, 2001). A few studies have empirically supported the relationship between brand personality and brand affect. For instance, Matzler, Pichler, and Hemetsberger (2007) examined if the brand personality such as 'extraversion' and 'openness' develop brand affect in customers and if brand affect develops the brand engagement behavior -as a sequential relationship. The researchers reported a significant relationship between brand personality and brand engagement through the mediating role of brand affect. The mediating effect of brand affect has also been validated by Roy, Khandeparker, and Motiani, (2016). The researchers found that the relationship between brand personality and brand engagement behavior is mediated by brand affect. Hence, this study posits that: H4: Brand affect mediates the relationship between brand personality and brand engagement.
Brand experience reflects the personal brand experience or the experience of the social contacts of customers. Another essential building block of CBBE, brand experience is a major factor that results in brand affect. Several studies in the past have explored the relationship between brand experience and brand engagement but only a handful of studies have explored the presence of mediators to explain how this relationship occurs. For example, Park and Park J. (2015) explored if brand affect mediated the relationship between various dimensions of brand experience and the brand engagement behavior. The results of the study validated the mediating role of brand affect in linking most dimensions of brand experience with brand engagement. In the same vein, Verleye et al. (2014) investigated if brand experience significantly predicted the brand engagement in the service sector through the mediating role of brand affect. The findings of this study also revealed that brand affect functions as a linking force (mediator) between brand experience and brand engagement.
H5: Brand affect mediates the relationship between brand experience and brand engagement.

Instrument and Measurements
To collect the data for this study, a questionnaire was adapted from previous studies. The construct of brand experience was adapted from Brakus et al. (2009) , brand personality was adapted from Geuens et al. (2009), brand affect was adapted from Chaudhuri and Holbrook (2001) and from Xie et al. (2015), and brand engagement was adapted from Bergkvist and Larsen (2010). The questionnaire comprised two parts; constructs measurements and demographic questions. For the 4 constructs in the study, a total of 32 items were employed. Specifically, brand experience was measured with 11 items, brand personality was measured with 10 items, brand affect was measured with 5 items, and brand engagement was measured with 6 items. Later one item related to brand personality was removed due to low outer loading. The items of the constructs were anchored by 5 points Likert scale: strongly agree = 5, agree = 4, neutral = 3, disagree = 2, strongly disagree = 1.

Sample
As the target population of this research was the automobile owners (passenger vehicles) and the total number of passenger cars in Pakistan were 2,889,500 until 2017-18, a sample of 538 responses was collected from three major cities of the country; Karachi, Lahore, and Islamabad. The sample size was primarily determined following the guidelines of Krejcie and Morgan (1970) and Salkind (2012). The respondents were contacted through mall intercept and were administered the questionnaire in person. Following the multistage sampling technique, the cities, malls and the data collection points at the malls were selected as per the directives of Sudman (1980) who contended that the data collection from malls through their prescribed method is similar to cluster sampling (Sudman & Blair, 1999). Only those vehicle owners were handed out the questionnaire who owned reconditioned car brands. After assessing the collected responses for missing values, invalid responses and outliers, a total of 426 responses were included in the final results.

Results
To assess the data collected from the response, this study has used PLS-SEM, a second -generation multivariate analysis technique. PLS-SEM is an effective and adaptable quantitative data analysis technique for predictive modeling (Cheah, Ting, Ramayah, Memon, Cham, & Ciavolino, 2018). SEM offers superior capabilities to estimate moderation and mediation effects that a regression modeling is not able to examine (Preacher & Hayes, 2004). It is recommended for the objective of predicting relationships and explanation of the variance (Hair, 2017). PLS-SEM technique has been applied through SmartPLS software. A major advantage of PLS-SEM is that it has the ability to adequately measure hierarchical -component models (HC) (also known as second-order constructs) (Becker, Klein, & Wetzels, 2012;Johnson, Rosen, Chang, Djurdjevic & Taing, 2012;Ringle, Sarstedt, & Straub, 2012). As the research framework of this study involves two second-order constructs: brand personality and brand experience and two low-order constructs; brand affect and brand engagement, PLS-SEM allows to adequately measure the second -order and low order constructs integrated in a single framework. Due to the unequal number of items in the constructs, the disjoint two-stage process was adopted for the measurement model as recommended by Backer et al. (2012) and Sarstedt, Hair, Cheah, Becker, and Ringle (2019). Before measurement and structural models assessment, common method bias and normality of the data were assessed. Normality was examined through skewness and kurtosis. The skewness values ranged between ± 0.016 and ± 1.497 while the kurtosis values were between ± 0.258 and ± 1.14, suggesting normality of the data as the threshold value of ± 3 is suggested by Bollen (1989) and Kumar, Lee, & Kim, (2009). For bias in the customers' responses, Harman's single factor test was employed as proposed by Mattila and Enz (2002) and Podsakoff and Organ (1986). The test showed the maximum variance below 50% using single factor constraints, suggesting the absence of bias in response. Lastly, the data was examined for multicollinearity through Variance Inflation Factors (VIF). The VIF values appeared between 1.301 and 2.316, indicating the VIF values well below the threshold value of 5 .

Measurement Model Assessment
In PLS-SEM, the measurement model assesses if the manifested variables (items) significantly explain the unobserved -latent variables (constructs); and if the variables are valid. Measurement model assessment establishes the reliability and validity of the constructs (Ramayah, Chow, & Chyaw, 2011b). Table 1 displays the 1st and 2nd stage assessment of the measurement model. Outer loading values are recommended to be ≥ 0.70 (Hair et al., 2014). However, it is also recommended to retain the items that show outer loadings ≥ 0.40 if they contribute to the recommended threshold of AVE ≥ 0.50 (Byrne, 2016). Following these guidelines, all the items that loaded ≥ 0.60 were retained as they provided the AVE value ≥ 0.50 for convergent validity. For this, the item, PRSEMOT10 was removed. The outer loadings of the retained items for the lower-order constructs ranged between 0.683 and 0.891 while for the second-order constructs outer loadings ranged between 0.683 and 0.891. Likewise, the composite reliability (CR) of the lower and higher order constructs ranged between 0.722 and 0.895, satisfying that recommended threshold of ≥ 0.70 (Hair et al., 2014). Table 2 represents the 1st -stage discriminant validity of the constructs. While convergent validity examines if items in each construct show desirable variance, discriminant validity represents the desirable dissimilarity amongst the constructs in a research framework (Duarte & Raposo, 2010;Hair et al., 2014). Crossloadings and HTMT values established the discriminant validity of constructs. Cross-loadings of the items for a construct should be higher than its loadings with other constructs in a study (Chin, 1998;Hair et al., 2014). Thus, first, cross-loadings of the first stage of the measurement model were satisfied and then the cross-loadings of the second stage of measurement model were established. Appendix A and B display the cross-loading values of the first and second stage of measurement models. Next, the HTMT criterion was assessed to further strengthen the discriminant validity of the measurement model. As displayed in Table 2, the HTMT values ranged between 0.23 and 0.82. Table 3 shows the HTMT values between 0.604 and 0.896. Since HTMT values are recommended to be ≤ 0.90, the discriminant validity of the constructs in this study is established (Gold, Malhotra, & Segars, 2001). The Fig. 1 below depicts the second stage (final) measurement model.

Structural Model Assessment
The structural model, also referred as inner model in PLS-SEM, explains the nature of the relationship between latent constructs and if the relatonships are significant. Alternatively, the structural model assessment reveals if the hypothesized relationships are significant and if they are negatively or positively linked to each other. Each path between the constructs represents a hypothesis earlier presented in the methodology section.
The hypothesis 1 of this study predicted a relationship between brand personality and brand affect. The results of PLS-SEM path modeling provided substantial evidence of brand personality influencing brand affect amongst the automobile customers (β = 0.379, t = 9.58, P < 0.00). Hypothesis 2 of the study predicted a relationship between brand experience and brand affect.
The results also substantiated it by furnishing evidence of a significant relationship between experience and brand affect (β = 0.538, t = 14.79, P < 0.00). Moreover, the relationship between brand affect and brand engagement, hypothesis 3, also appeared as significant (β = 0.638, t = 17.11, P = < 0.00). In contrast, hypothesis 4 and 5 predicted the indirect relationships between brand engagement and brand personality and brand experience through the mediating role of brand affect. The bootstrapping procedure was employed to assess the mediating affect as advised by Hayes and Preacher (2010). For hypothesis 4, PLS-SEM results supported this relationship and identified the significant influence of brand affect to link brand personality and brand engagement (β = 0.242, t = 8.170, P = 0.000). Similarly, the mediating role of brand affect between brand experience and brand engagement emerged significant (β = 0.343, t = 10.713, P = 0.000) in the PLS-SEM modeling results for hypothesis 5.

Fig. 2. Structural Model with Mediator
Further, the coefficient of determination -R 2 has been assessed for the structural model as advised by Hair, Sarstedt, Ringle, and Mena (2011). and Hair et al. (2012). Essentially, R 2 explains the degree of variance in an endogenous variable (mediator and dependent variable) in PLS-SEM by the exogenous variable (independent variables) (Hair, Black, & Babin, 2010;Elliott & Woodward, 2007). Table 4 displays the R 2 for brand engagement is 0.408 and for brand affect is 0.646, suggesting the coefficient of determination (R 2 ) is between weak and moderate for brand engagement and is between moderate and substantial for brand affect, as per the guidelines provided by Henseler et al. (2009). In addition, the effect size f 2 has also been measured to identify the individual contribution of each exogenous variable to coefficient of determination R 2 . As shown in Table 4, f 2 value of brand experience is 0.594 and of brand personality is 0.295 for brand affect, suggesting the effect size is large for brand experience and is between low and medium for brand personality, as per the guidelines given by Cohen (1988). Lastly, predictive relevance (Q 2 ) of the model was assessed through blindfolding procedure in SmartPLS. Q 2 explains the significance of exogenous variables in predicting the endogenous variables in PLS-SEM (Hair et al., 2014). Any value > 0 of Q 2 is considered to show its significant power to predict the constructs (Henseler et al., 2009). Hence, the cross-validated redundancy values 0.381 for brand affect and 0.201 for brand engagement reflect the predictive power of brand personality and brand experience towards brand engagement.

Conclusion
This study has leveraged customer-based brand equity theory to investigate the customers' preference for the reconditioned car brands over the locally produced global brands in Pakistan. Building upon the CBBE model proposed by Keller (2001), this study assessed if brand personality and brand experience influenced the brand engagement behavior (true brand loyalty) of the automobile customers in Pakistan. The data obtained from the automobile customers in the country revealed that the brand equity drivers: brand personality and brand experience are the significant predictors of brand engagement behavior for the automobile brands in the country. The analysis also substantiated the mediating role of brand affect between the brand personality, brand experience, and brand engagement. The findings of the study make significant contributions to CBBE and the relationship marketing theories as well as fill several empirical gaps identified in the extant literature related to band engagement and its antecedents. CBBE is the proponent of raising and aligning the brand perception which greatly hinges on the positive brand experience, image and feelings, eventually resulting in brand engagement (Keller, 2003). This is one of the first studies that provide theoretical and empirical support to establish that brand experience and brand personality are the antecedents of brand engagement behavior. Likewise, the relationship marketing concurs customers reciprocate to the marketing activities of businesses that benefit customers (Grönroos, 1994), corroborating the Keller's (2001) view of brand engagement as a 'highly sync' relationship between customers and brands. Additionally, the relationship marketing theory explains that response of customers towards brands originates from the perceived value they derive from brands (Blackwell et al.,1999). On the same lines, the benefits customers draw from the brand become the core relational factors that determine the long-term relationship between customers and brands, eventually leading to brand engagement behavior (Bowden, 2009). Thus, the empirical findings of this study make a significant contribution to the theory by aligning the brand engagement behavior with the relationship marketing theory and empirically providing the largely neglected role of the brand equity drivers as predictors to the bran engagement behavior. Especially, how brand affect (positive emotional response) influences the brand engagement behavior and its mediating effect appears to be largely neglected in the extant marketing literature. Further, the past studies that examined brand engagement in the context of automobile brands are strikingly few in number. Specifically, such studies are scarce in the automobile market of Pakistan. The findings supported that brand experience and brand personality play significant role in shaping brand engagement behavior of automobile customers. The results also substantiated the significant impact of brand affect (feelings and emotions) to build a highly sync customer-brand relationship in the context of automobile brands. Evidently, automobile customers are motivated by the pleasant brand experience and brand affect to stay truly loyal to automobile brands. Additionally, brand personality and brand experience play a major role in developing a close customer-brand relationship in the automobile market of Pakistan. Whereas, brand affect also holds significance for the automobile customers in Pakistan who are evidently driven by their emotions in the purchase and repurchase of automobile brands. Thus, the SKD brands in Pakistan are recommended to enhance the personality and experience of their brands which leads to brand affect and eventually helps in creating brand engagement, the declaration of brand loyalty.

Limitations and Recommendations
Research studies, in spite of the due care and scientific methods, are likely to subject to some limitations. Only a single mediator has been examined in this research that links the customer-based brand equity drivers to brand engagement. It is expected that other variables may be identified in future studies to propose a more comprehensive research framework that enhances brand equity and relationship marketing theories. Due to various constraints, the data of this study was obtained through a cross-sectional design. Future studies may adopt a longitudinal research design and compare the results. Likewise, this study's scope was limited to the automobile customers in Pakistan. Future studies may test the research framework of this study for other product categories and may examine the automobile market of other countries in Asia and other parts of the world. 0.422 *AFF= brand affect; ENG = brand engagement; EXPBHV = behavioral experience; EXPCOG = cognitive experience; EXPSNS = sensory experience PERSACT = personality -activity; PERSRSP = personality -responsibility; PERSAGG = personality-aggressiveness; PRSSIMP = personality -simplicity.