Millennials' Response Toward Luxury Fashion Brands: The Balance Theory’s Perspective

The mushrooming growth of luxury brands has been evidenced worldwide, especially among developing countries such as Malaysia. Despite Malaysian consumer confidence slowing due to uncertainty about the economic conditions, the demand for luxury brands is still expected to rise. Interestingly, a significant rise in the demand for luxury brands has been observed in the millennials group. This study examines the relationship of electronic word-of-mouth and the country of origin’s image on millennials' attitudes toward luxury brands and the intention to purchase luxury fashion brands. The study among 333 millennials revealed that the country of origin’s image has both direct and indirect effects on the intention to purchase luxury fashion brands. Electronic word-of-mouth (eWOM) was found statistically insignificant in explaining the intention to purchase, but eWOM had a significant relationship on intention to purchase through its mediation effect of the attitude toward luxury brands. Accordingly, the findings found support for the balance theory. The results also suggest that the owners of luxury brands should carefully manage the country of origin because it will affect the overall brand image, as perceived by consumers. Limitations and future directions of the study are also included at the end of the article.


Introduction
Over the past few years, luxury brands have been recognized as one of the most rapidly growing businesses in numerous emerging markets worldwide. For instance, within the Asian market, the trade in luxury products in Malaysia and Indonesia is expected to rise significantly. Euromonitor's (2020) report stated that despite the reduction in Malaysian consumer confidence in the country's economic growth, the demand for luxury brands among consumers and tourists in Malaysia has continued to rise. In fact, in 2013, Bain & Company classified Malaysia as one of the six Southeast Asian countries that contributed to the luxury market's development (D'arpizio and Levato, 2013). Given the growing demand for luxury goods, fashion brands are considered to play a significant role in affecting consumers' lifestyles (McColl and Moore, 2011). The growth is attributed to the substantial increase in Malaysian household disposable incomes and the drastic changes to luxury brand outlets in Malaysia's retail landscape (Young, 2013). Indeed, the COVID-19 pandemic has not hampered sales of luxury brands, as more brands are reinventing and re-imaging themselves to fit into the digital revolution and face the new normal lifestyles worldwide. Many luxury brands have recently joined omnichannel, in cooperation with Amazon, to reach potential customers (Deloitte, 2020).
Interestingly, the demand for luxury fashion brands is not only popular among adults or higher income earners. It has also recorded an increasing demand from millennial consumers. According to the Luxury Goods World Market Study report cited by D'arpizio et al., (2019), young consumers, especially millennials born between early 1980 and 2000, are estimated to represent 55% of the global personal luxury goods market by 2025. Despite their youthfulness and relatively low-incomes, millennials are eager to look fashionable by dressing in the latest fashions, to project their professional images. They tend to be stylish, up-to-date, and keen to shop frequently. Furthermore, previous studies have shown that millennial consumers are highly brand conscious and willing to spend extra money on luxury products to enhance their credibility, image, and reputation (Chan, To, and Chu, 2015;Eastman, Iyer and Thomas, 2012;Giovannini, Xu and Thomas, 2015;Kapferer and Michaut-Denizeau, 2020). Concisely, this indicates that millennial consumers are among the market segment who have a high degree of demand for luxury products. Hence, it is crucial to understand how millennial consumers respond to Malaysia's luxury fashion brands, especially in their attitude and purchase behavior.
A literature review indicates that studies undertaken to understand consumer behavior toward luxury fashion brands have received considerable attention through the theory of planned behavior (Jain, Khan, and Misra, 2017;Zhang, Cude, and Zhao, 2020). However, those studies focused on general consumers and did not specifically understand millennial consumers, who are unique and behave differently. Despite the significant contribution of the millennials' cohort to the luxury brand market (D'arpizio and Levato, 2013), studies examining the factors influencing millennials' attitudes and behavior toward luxury fashion brands are rather limited, especially in the Malaysian context (Burnasheva, GuSuh, and Villalobos-Moron, 2019;Kasuma et al., 2016). Most previous studies attempted to link perceived value, functional value, individual value, and so-cial value to the purchase of luxury brands (Giovannini et al., 2015;Hashim et al., 2019;Kasuma et al., 2016). Besides, Teo, Nik, and Azman (2017) examined how fashion knowledge, brand confidence, fashion lifestyle, and fashion patronage behavior influence millennials' fashion involvement. There are a limited number of studies aimed at understanding millennials' attitudes toward luxury goods and what influences them.
The demand for luxury fashion brands actually could be underpinned by the attitude toward luxury brands. In short, an attitude refers to a person's tendency to react in consistently favorable or unfavorable ways toward specific issues (Hill, Fishbein, and Ajzen, 1977). It is assumed that the more favorable the attitude is, the higher the demand for luxury fashion brands is. Thus, attitude relates very closely to oneself, which leads to the question of what influences one's attitude? Previous studies were based on the theory of planned behavior. As a result, the most discussed determinants are limited to the predictors of that theory: attitude, subjective norms, and perceived behavioral control. This study is based on a basic attitude formation theory, namely the balance theory, which aims to understand what influences a person's attitude (i.e., millennials). The balance theory proposed by Heider (1958) suggests that a person's attitude toward a particular system depends on three main elements: person (P), other person (O), and object (X).
Beauloye (2020) stated that "millennials value luxury brands that celebrate their passions and individuality in a way that can be visually shared on social media." Besides, there is also a strong stigma among Asian consumers of "Western stuff is hot" among the young generation (Heine, Atwal, and He, 2019). Based on these statements, it seems that the country of origin (COO) and use of social media could explain millennials' attitudes toward luxury brands. However, there have been mixed results in the past, and the predictors seem to have been neglected in understanding millenials' attitudes. For instance, Daswani (2015) claimed that millennial consumers are not concerned about country of origin (COO) but might appreciate design and innovation. Hence, it is crucial to examine how these predictors could explain the millennials' attitudes. Based on the balance theory, millennials are the "person," how they perceive electronic word-of-mouth (eWOM) could be considered as "other," and the COO's image could serve as the "object." In order to form a favorable attitude, the relationship between these elements must be balanced. Aiello et al., (2009) studied the international perspective on luxury brands and the effect of the country of origin. The scholars further added that the impact of COO might be varied between countries and the causal model, in that the integration of the brand and COO should be evaluated by understanding the attitude toward luxury brands. Mohd, Nasser, and Mohamad (2007) indicated that the COO would affect higher order consumers' brand building behavior as in customer-based brand equity (CBBE) model which is related to brand loyalty and brand association as compared to the lowest level i.e. brand salience. Scholars such as Phau and Leng (2008) asserted that local and foreign luxury brands (COO elements) affect attitudes toward luxury fashion differently.
More importantly, Chu, Kamal, and Kim (2013) suggested that social media usage would also affect millennials' intentions to purchase luxury products. Burnasheva et al., (2019) also highlighted that social media us-age is different and unique for each country. Besides, in the context of today's digital marketing era and the high dependency of millennials on social media interaction, Burnasheva et al., (2019) and Abubakar, Ilkan, and Sahin (2016) stressed that eWOM significantly affects a consumer's attitude and purchase decision. Meanwhile, Sijoria, Mukherjee, and Datta (2018) added that eWOM plays a significant role in strengthening brand equity. More importantly, peer communication via social media supports brand-related information sharing (Delafrooz, Rahmati and Abdi, 2019).
Numerous studies have been conducted to understand COO and consumers' attitudes on overall consumer behavior, specifically in Malaysia. For instance, comparative analyses based on Malaysia versus China have been conducted to explore branding (Lew and Sulaiman, 2014;Sarwar et al., 2013), halal country of origin (Rios, Riquelme and Abdelaziz, 2014), attitude toward international products (Asshidin, Abidin, and Borhan, 2016), and COO's image (Adenan, Kasuma, and Abang Abdul Rahman, 2018). However, few studies have explored the effect of COO on attitude and the intention to purchase luxury brands, especially among millennial consumers. Malaysian consumers, especially youngsters, are truly brand conscious (Kasuma et al., 2016). Besides, there is a shred of evidence for a significant difference in attitude formation across generations (Lambert-Pandraud and Laurent, 2010).
To understand the effect of COO on attitude formation and intention to purchase eWOM has been introduced into the existing body of knowledge. According to Chao and Rajendran (1993), the impact of COO on attitude formation could be reduced when it is tested with multiple cues. Millennials are identified as socially active in searching for and exchanging information online, and are thus referred to as "digital natives" (Sozio, 2010). Hence, it is worth including eWOM in the present model. Specifically, this study examines the millennials' responses, in terms of their attitude formation and how it subsequently affects their intention to purchase luxury fashion brands. Based on the extensive literature and mixed results linking the proposed direct relationship between COO and eWOM, the attitude toward luxury fashion brands is also proposed to mediate the relationship between COO, eWOM, and intention to purchase luxury fashion brands.

Literature Review
Different generations buy different brands (Lambert-Pandraud and Laurent, 2010); elderly consumers buy older brands, younger consumers tend to be more brand conscious but they are not brand loyal (Little, 2012). Millennial consumers are also known as consumers with a high level of self-esteem (Hashim et al., 2019). They prefer reputable and luxury brands (especially foreign brands) to secure social recognition. Besides, millennials are also a group of consumers who have a high dependency on their peers during purchase decisions (Fernandez, 2009). Chadha and Husband (2006) refer to luxury fashions as fashionable and high-quality goods made by reputable luxury brands. Fashion products portray an ideal context to evaluate millennials' attitudes toward luxury brands and purchase intentions. Fashion apparel is considered a highly self-expressive product, and hence the possession of luxury-branded clothing is viewed as a vital status indicator for millennials (Casidy, Nuryana, and Hati, 2015;Sierra and Hyman, 2011;Koksal, 2014). Correspondingly, other relat-ed studies have suggested that luxury fashion brands enable consumers to demonstrate their self-concepts, improve their self-esteem, and make a good impression on their peers (Deeter-Schmelz, Moore and Goebel, 2000;Taylor and Cosenza, 2002;Piacentini and Mailer, 2004;Husic and Cicic, 2009;Casidy et al., 2015).
This research fits within the above theoretical perspectives, considering the recent developments in social media involvement among millennials (especially eWOM) for influencing consumers' decision-making processes. This study attempts to add to the existing knowledge on how the COO's image and eWOM influence millennials' attitudes toward luxury brands and their purchase intention, which is somewhat limited (Godey et al., 2012).

Intention to Purchase Luxury Brand
Based on the marketing literature, the relationship between intention and behavior is best explained through the theory of planned behavior (TPB) (Ajzen, 1991). The TPB assumes that people's behavior is best predicted by examining one's intention toward the behavior. Consumer purchase intention is deemed the gesture or indication of a consumer's actual purchase (Mamat, Noor, and Noor, 2016). The intention is presumed to develop the determination factors which affect the behavior. It demonstrates the amount of effort an individual is planning to employ, or how determined the person is to execute the behavior (Ajzen, 1991). Prior studies reflect that there is a positive and significant relationship between intention and actual purchase behavior. Hence, this study attempts to examine consumers' intentions to purchase rather than examining the consumers' actual purchases, to understand the phenomena at hand. Besides, Ajzen (1991) suggested that intention and behavior may be influenced by several others predictors, such as the attitude toward the behavior, subjective norms, and perceived behavioral control.
The study by Botha and Reyneke (2016) among 338 millennials in South Africa found that ceertain factors, such as social value, personal value, functional value, self-congruity, brand consciousness, and social media significantly affected millennials' purchase intention of a luxury brand. Materialism plays a significant impact. In Malaysia, the study by Kasuma et al., (2016) of 200 millennial respondents indicated that financial value, functional value, individual value, and social value are critical determinants for the intention to purchase luxury handbags. Besides, Ali and Shafique (2015) studied Pakistani intentions to purchase luxury fashion goods by examining three main predictors, namely: materialism, social comparison, and fashion innovativeness, with the attitude toward luxury fashion as a mediating variable. This study, with 112 respondents, found that only social comparison statistically affects the attitude toward luxury fashion. Materialism, fashion innovativeness, and attitude toward luxury fashion do not significantly influence the intention to purchase luxury fashion brands.

The Balance Theory
As mentioned earlier, the TPB has been widely used to understand consumers' behavior toward luxury brands. In the TPB, attitude, the subjective norm, perceived behavioral control, and intention are used to predict behavior. This study investigates in-depth how millennials form attitudes toward luxury fashion brands by employing the basic attitude formation theory, namely the balance theory, and extending it into the intention to purchase. The balance theory assumes that individuals will form their attitudes based on how they perceive things and how they perceive the views of other people toward a certain object (Heider, 1946). According to Heider (1946), the balance theory builds upon a triadic relationship between: a person (P), another person (O), and an object (X). In order to form a positive attitude, the relationship between P, O, and X must be balanced. In this case, P are the millennials, O is how eWOM is perceived by other people, and X is the COO's image of luxury fashion brands. Hence, to understand the phenomenon, namely luxury fashion brands among millenials, Heider's balance theory, which incorporates the importance of interpersonal relationships, is relevant to underpin the proposed model. According to Heider (1958), emotions and social relationships, or perceptions between people on certain subjects, must be harmonized to create cognitive consistency (likes or dislikes). Every person has their own opinion, which possibly traps a person in a situation where their opinion conflicts with that of others. Such an imbalance or disagreement could lead to a feeling of discomfort. Considering that millenials are known to have robust social networks (Park, Chen and Cheng, 2021), the opinion of others through eWOM also perhaps helps shape their attitude. In this context, even though millenials could have a positive perception toward the COO's image, if they received negative feedback from eWOM, they might have a negative attitude toward luxury fashion brands due to an imbalance or perceived disagreement.
The balance theory is widely used in psychology and marketing, primarily for understanding the effect of brand and celebrity endorsements (such as Yakut, 2017). The balance theory states that if people like certain celebrities, and those celebrities use specific brands, then there is a greater tendency for people to have a positive attitude toward the brand. Similarly, in the context of luxury fashion brands, it is postulated that the more positive the perceptions are that consumers have toward the origin of the brand (country of origin image) and it is positively reviewed by others (eWOM), the more favorable the consumers' attitudes will be toward the luxury fashion brand.

Relationship Between The Coo's Image, Ewom, and The Attitude Toward Luxury on Intention to Purchase Luxury Fashion Brand
Country of origin refers to the country where a manufacturer's brand or product is associated with and where the products are manufactured or assembled (Mohd et al., 2007;Valaei and Nikhashemi, 2017). The country of origin reflects the product/ service's perceived image. Roth and Romeo (1992) state that the country of origin's image refers to all the consumers' established perceptions about the products from that particular country, based on their existing perceptions of that country's manufacturing and marketing strengths and weaknesses. The most valuable recorded luxury fashion brands originate from developed countries such as Italy (Giorgio Armani, Versace), France (Louis Vuitton, Gucci), the United Kingdom (Chanel, Burberry), and the United States of America (Calvin Klein, Coach) (Deloitte, 2019). However, in today's emerging economies, it is common to have luxury brands originating from, and produced in, developing countries, such as China, Vietnam, and Bangladesh. For instance, luxury brands such as Louis Vuitton, Gucci, and Burberry are also produced in China (Xu, 2020). Besides, luxury brands also originate from developing countries, such as from China's local fashion designers, namely, Ne-Tiger, Longio, and Shang Xia (ATKearney, 2021). Meanwhile, the luxury fashion brands such as Farah Khan and Jimmy Choo originated in Malaysia, and Kraton, Peggy Hartanto, and Nataoka came from Indonesia (Dettmann, 2020;Khoo, 2019).

Relationship Between Coo's Image, Attitude, and Intention to Purchase
Previous studies indicated that the brand name and country image are extrinsic indications consumers apply during their product-evaluation processes. The influence of the country of origin's image significantly influences the purchase of luxury goods compared to other product categories (Shukla, 2011;Vijaranakorn and Shannon, 2017). This is supported by Nayir and Durmusoglu (2008), who indicate that the country of origin's image is the best external indicator for a product's prestige, mostly when the customer lacks knowledge and information about international brands (Hamzaoui and Merunka, 2006;Moradi and Zarei, 2012). Esmaeilpour and Abdolvand (2016) indicate that, when consumers are unfamiliar (lack of information) with the brand's or product's country, there is a greater tendency for the brand to be excluded from their evaluation list. Thus, the country of origin's image plays a significant role in affecting consumers' beliefs and attitudes toward luxury brands. A study among 419 Iranian Gen-Y members indicates that only partial COO image dimensions (i.e., technology advance) significantly affect the attitude toward luxury goods (Esmaeilpour and Abdolvand, 2016). The authors further suggest that the relationship should be tested in other countries, considering the different cultural settings. Besides, an interview among 165 consumers from eight countries emphasizes the importance of the brand, the COO, and the youngsters that make up the luxury brand's market (Aiello et al., 2009). A literature review also found that the country of origin's image can positively and negatively affect attitudes and purchase decisions. More importantly, Chao and Rajendran (1993) opined that the effect of the COO on attitude formation might also be reduced when assessed with multiple cues, this subsequently may affect the balance theory of attitude. The study by Abdolvand, Forough and Tavakkoli Targhi (2017) among 150 Iranian students further asserted that the COO's image influences consumers' perceptions and significantly affects the intention to purchase luxury branded Swiss watches. However, in the event of a multiple cues assessment, primarily using the extrinsic cues such as the price and brand, the effect of the COO is less critical for evaluating luxury items (Montanari et al., 2018). This is in line with Prendergast, Tsang and Chan's (2010) finding that extends the concept of the COO into the country of brand (COB). The scholars reveal that the COB has a significant influence on the intention to purchase, especially for consumers in a high involvement group. Based on the preceding discussion, it is hypothesized that: H1: Country of origin has a positive relationship with the attitude toward luxury brands.
H2: Country of origin has a positive relationship with the intention to purchase luxury fashion brands.

Relationship Between e-WOM, Attitude, and Intention to Purchase
The balance theory also suggests that attitude formation depends on what millennials perceive about certain things and on how others perceive the same things. Millennials typically depend a great deal on peer communication. The most common type of peer communication appears in the form of wordof-mouth (WOM). WOM is known to have a significant influence on consumer behavior. WOM refers to "any informal, person-to-person communication process between a perceived non-commercial communicator and the receiver regarding a service, organization, brand or product" (Walker, 2001:pp.63). Even though the communicator does not formally represent the organization and/or the brand, their shared comments, opinions, and suggestions could significantly impact the overall organization's and/or the brand's performance. Essentially, WOM is free advertising. Given the powerful impact of social media today, WOM via electronic media, or eWOM, is rather significant in determining people's attitudes and behavior toward a particular brand. According to Dinh and Mai (2016), eWOM refers to "any positive or negative statement made by potential, actual or former customers about a product or company." Nowadays, the consumer buying process is made more accessible. Consumers always refer to product information online, primarily based on testimonials and recommendations, mainly to reduce uncertainty and product risk (Abubakar et al., 2016), especially considering the premium price of luxury brands. Negative expressions via social media, or online forums such as Facebook and Instagram, would tarnish the overall brand image, affecting potential consumers' attitudes, behavior, expectations, and percep-tions (Trusov, Bucklin, and Pauwels, 2009). Though one would think WOM and eWOM are similar, Park, Lee, and Han (2007) stated that eWOM is more influential in affecting consumers' attitudes and behavior. Kudeshia and Kumar (2017) proposed that positive social eWOM, such as Facebook, will significantly impact brand attitude and purchase intention. A study among 311 Indian consumers revealed that positive reviews from a social network such as Facebook could enhance a message's credibility. It has been found that eWOM statistically influences both people's attitudes toward a brand and their purchase intentions. However, the context of that study was based on smartphones. Hence, it is worth studying such impacts in the luxury fashion context, especially among Malaysian millennials. Besides, Prendergast, Ko, and Yuen's (2010) study among 150 Hong Kong consumers found that their attitude toward online WOM directly and indirectly relates to their purchase intention. Farzin and Fattahi (2018) studied 369 university students in Iran and concluded that eWOM directly and indirectly affects purchase intentions. Hence, based on the preceding discussion, it is hypothesized that: H3: eWOM has a positive relationship with the attitude toward luxury brands.
H4: eWOM has a positive relationship with the intention to purchase a luxury fashion brand.

Relationship Between Attitude Toward Luxury Brand and Intention to Purchase
Consumers' evaluations of their attitudes toward the brands often uses evaluative scales from favorable to unfavorable or lik-ing to disliking (Chang and Thorson, 2004;Escalas, 2004;Miniard et al., 1991;O'Cass and Choy, 2008). Hence, if a person believes that certain behavior would lead to favorable outcomes, then his or her attitude toward conducting the behavior is presumed to be positive. A vast body of knowledge indicates that consumers' attitudes significantly influence consumers' buying decision processes for luxury brands. For instance, Zhang and Kim (2013) highlighted that attitude is influenced by brand consciousness, social comparison, and fashion innovativeness. Another study found that consumer attitude has a significant relationship and effect toward the intention to purchase luxury fashion apparel made in sweatshops, such as in the survey conducted by Phau, Teah, and Chuah (2015). Rolling and Sadachar (2017) attempted to examine the attitude toward sustainable luxury fashion brands on the purchase intention of sustainable luxury brands among millennials. The scholars found the attitude toward sustainable luxury brands significantly influenced millennials' purchase intentions of luxury fashion brands. Thus, it is hypothesized that: H5: Attitude toward luxury fashion has a positive relationship with the intention to purchase luxury fashion brands.

Mediation Effect of Attitude Toward Luxury Brands
Bilge (2015) opined that consumers would have positive and negative attitudes toward luxury goods depending on the functionality, experimentality, and symbolism of the goods. Dubois, Czellar, and Laurent (2005) emphasized that consumers could be divided into three separate groups based on their attitudes toward luxury. The first group has the "elitism attitude," which is related to the belief that luxury is only accessible to certain people. Secondly, the "democratic attitude" believes that luxury should be available to everyone, and finally, the "distant attitude" where people believed that luxury is a different world to which they do not belong. The latter group may perceive luxury as unfavorable and consider it unnecessary and over-priced. Hence, the more negative a consumer's attitude toward luxury brands is, the lower the intention to purchase luxury brands will be. A literature review also indicates the indirect effect of the COO's image and eWOM on the intention to purchase a luxury brand. For instance, Abdolvand et al., (2017) in a study involving luxury Swiss watches emphasizes how the COO's image formed consumers' perceptions and how those perceptions influence the intention to purchase luxury brands. Besides, Elseidi and El-Baz (2016) and Kudeshia and Kumar (2017) also recorded the interaction of attitude on the relationship between eWOM and the intention to buy a certain brand of smartphone. Sheeraz et al., (2016) also highlighted the mediation role of attitude in explaining consumer purchase intentions. Hence, it is proposed that the attitude toward luxury fashion brands would mediate the relationship between the COO's image and eWOM on the intention to purchase luxury fashion brands. Based on the above discussion, the following hypotheses were developed: H6: Attitude toward luxury brands mediates the relationship between the country of origin's image and the intention to purchase luxury fashion brands.
H7: Attitude toward luxury brands mediates the relationship between eWOM and the intention to purchase luxury fashion brands.
Based on the preceding discussion, the following Figure 1 summarizes the proposed research framework of this study:

Methods
This study employed a quantitative approach based on a cross-sectional study. This correlational study aimed to link the relationship between the country of origin's image, eWOM, people's attitudes, and the intention to purchase luxury fashion brands by millennial consumers. For the purpose of the study, postgraduate students from one of the top-ranked management-based universities in Malaysia were approached, as this group of students could satisfy millennial categorization requirements. According to Ordun (2015) and Yazıcı (2016), a person born between 1981 to 1996 is considered to be a millennial or Generation Y. The total population of the study was 4,389. Based on Krejcie and Morgan's (1970) sample size determination table, a sample of 351 respondents would be considered adequate to represent the population under study. As a result, 400 questionnaires were distributed. As the sampling frame was not accessible due to privacy and confidentiality requirements for data collection purposes, this study em-ployed systematic sampling through "faculty interception" (Hornik and Ellis, 1988). Every 5th graduate entering the postgraduate building was approached with the screening questions about their age, to see if they were millennials, if they were familiar with any social media networks, and were familiar with any luxury fashion brands. Thus, based on the criteria set, the sampling method of this study was based on purposive sampling. A total of 400 questionnaires were obtained after two months. However, during the data's cleaning and screening process, 67 questionnaires were excluded due to incomplete answers or technical errors, resulting in 333 remaining items with usable data.
For the data's collection, a research instrument, namely a questionnaire, was developed by this study. Most measures were adapted from previous studies. For instance, the measurement for the attitude toward luxury fashion brands was adapted from Casidy et al., (2015) with five items. Based on the luxury brand literature, attitude could be based on thee cognitive (what we think and believe) and affective (what we feel and experience) components (Bian and Forsythe,  1. The luxury fashion brands originating from developed countries are more sophisticatedly made, compared to those from developing countries. 2. The luxury fashion brands originating from developed countries last for a long time, compared to those from developing countries. 3. The luxury fashion brands originating from developed countries are excellent compared to those from developing countries. 4. Developed countries have a reputation for making useful products, reflected in the luxury fashion products of their brands, compared to those from developing countries. 5. Developed countries have expertise in the merchandise they offer, reflected in the luxury fashion products of their brands, compared to those from developing countries. Attitude toward luxury brand 1. I look my best when wearing a luxury fashion brand name. 2. I consider the luxury fashion brand name when purchasing clothing for myself. 3. I like to wear luxury fashion brand names. 4. I will pay a higher price for luxury fashion brands that a popular designer or manufacturer makes. 5. I usually shop in fashionable luxury stores. Casidy et al., (2015) Intention to purchase luxury brands 1. I intend to buy luxury fashion brands constantly. 2. I intend to purchase luxury fashion brands within the next year. 3. The probability that I would buy luxury fashion brands within the next 12 months is high. 4. There is a strong possibility that I will buy luxury fashion brand products. 5. Whenever I need to buy goods, it is very likely that I will purchase a luxury fashion brand product. 6. I will buy luxury fashion brand products in the near future. Qian et al., 2017 according to Sweeney and Soutar (2001), the attitude toward luxury brands could be derived from the feelings generated by the luxury brands themselves, which could be related to the affective component. The measure for the country of origin originated from Vijaranakorn and Shannon (2017) with five items, and eWOM's measurement was adapted from Abubakar et al., (2016) with five items. Lastly, the intention to purchase luxury fashion brands was adapted from Soh, Rezaei and Gu, (2017) with six items. Table 1 summarizes the measurement items. Most items in this study mainly used a 5-point Likert scale ranging from "1-strongly disagree" to "5-strongly agree." The data analysis was performed by using SPSS Version 24.0 and SmartPLS 3.0. Table 2 presents the demographic profile of the respondents. Based on Table 2, most of the respondents were femalee and aged between 26 to 29 years old. About 53% of the respondents were Malays. The majority of the respondents were pursuing a Master's program (89.8%). Regarding their monthly income, about 32.1% earned less than RM1,000 and 28.8% between RM1,000 to RM3,000 per month. The remaining earned more than RM3,001 per month.

Assessment of Measurement Model
This study employed the two-stage model approach suggested by Chin (2010) which included measurement and structural models using SmartPLS-SEM 3.0. Figure 2 summarizes the measurement model of this study. The measurement model was used to assess the relationship among the indicators and the latent variables. The structural model specifically assessed the relationship among the latent variables (Hair et al., 2014;Henseler and Fassott, 2010). The measurement model was assessed through composite reliability, convergent validity, and discriminant validity. Internal consistency reliability was assessed to see whether the measure consistently represented the same construct which was assessed via composite reliability (CR) and average variance extracted (AVE). The acceptable threshold level for CR was 0.60 to 0.70 for exploratory research (Hair et al., 2014). Table 3, the composite reliability for all the study's constructs ranged from 0.856 to 0.899, which assured the measurement model's reliability. To assess the convergent validity, i.e., the degree to which one particular measure of the construct related to the other measure of the same underlying construct, the value of AVE was examined. Henseler and Fassot (2010) suggested that the AVE should be greater than 0.5 to show the amount of variance the items accounted for in the latent variable. Based on Table 3, all the constructs gained convergent validity as all the AVE ranged from 0.516 to 0.641.

Based on
Discriminant validity refers to the degree to which indicators differentiate across constructs by checking the overlapping correlation between measures. This study employed cross-loading criteria as suggested by Hair et  Hair et al., (2014), the difference between loadings among latent variables must not be less than 0.1. Based on Table 4, the result suggested there was ade-quate discriminant validity. Besides, Fornell and Larker's criteria and HTMT also found no violation of the criteria.

Assessment of Structural Model
The five-step approach for assessing the structural model was followed as proposed by Henseler and Fassot (2010). Firstly, it was essential to ensure that there was no collin-earity issue in the inner model of the study.  Secondly, to answer all the hypotheses in this study, a bootstrapping procedure was employed. Figure 2 presents the structural model of this study. The details are shown in Table 6. Table 6 shows the path coefficient assessment of the study. Based on Table 6, the country of origin and eWOM were statistically proven to have a significant effect on the attitude toward luxury fashion brands (β = 0.156, p<0.01) and (β = 0.435, p<0.01) respectively. Besides, the country of origin (β = 0.364, p<0.01) and attitude toward luxury fashion brands (β = 0.462, p<0.01) had a significant positive relationship with the millennials' intention to purchase luxury fashion brands. Hence, H1, H2, H3, and H5 were supported. H4 was not supported due to the negative relationship. Based on the proposed model, the attitude toward luxury fashion brands was also suggested to mediate the relationship between the country of origin and eWOM on millennials' intention to purchase luxury fashion brands, thus supporting H6 and H7. The next step was to measure the model's predictive accuracy via its co-efficient determination (R2). R2 represents the amount of variance in the millennials' intention to purchase luxury fashion brands by their country of origin, eWOM, and attitude toward luxury fashion brands. In this study, 45.2% of the variance in th millennials' intention to purchase luxury fashion brands could be explained by the country of origin and the attitude toward luxury fashion brands. A 27% variance in attitude toward luxury fashion brands was explained by the country of origin and eWOM. According to Hair et al., (2014), R2 values for this study described a substantial level of predictive accuracy. The fourth step was to determine the effect size via f2 (Cohen, 1988). Effect size measures the relative impact of exogenous variables to the endogenous variable. According to (Cohen, 1988), f2 is assessed at 0.02 (small), 0.15 (medium), and 0.35 (large). Table 7 shows the effect size of the predictors.
The last step was to assess the predictive relevance (Q2) using the blindfolding proce-  (Hair et al., 2014). Q2 values for the intention to purchase luxury brands and attitude toward luxury brands were 0.209 and 0.16, respectively, suggesting an adequate predictive relevance. Table  8 summarizes the values for the coefficient of determination (R2), predictive relevance (Q2), and effect size (f2) for the structural model.

Discussion
The findings revealed that the COO's image significantly affected the attitude toward luxury fashion brands and the intention to purchase luxury fashion brands. The findings concur with the previous studies of Montanari et al., (2018) and Hien et al., (2020). In this study, regarding the COO's image, the respondents were asked questions such as how they rate the image of luxury fashion brands from developed versus developing countries, in terms of their attributes such as sophistication, durability, performance, and reputation. The mean score for the COO's image was 3.77, indicating that the respondents preferred luxury fashion brands from developed countries more, based on the attributes that were mentioned. Using this millennials sample, the higher the COO's image was perceived by the consumers, the more favorable their attitude toward luxury fashion brands was, and the higher the propensity to purchase luxury fashion brands was. The result holds as this sample involved a university sample. About 80% of the students were Master's degree holders. They were mature and experienced consumers and capable of making good decisions. Besides, millennial consumers believe that developed countries have technologically advanced and highly competitive capabilities, thus producing better quality, more reliable, and more valuable luxury fashion brands (Kalicharan, 2014).
Surprisingly, eWOM was statistically insignificant in explaining the intention to purchase luxury fashion brands among millennial consumers. The result was inconsistent with the previous studies of Abubakar et al., (2016) and Farzin and Fattahi (2018). However, the finding was consistent with the remark made by Milaković, Mihić, and  Ivasečko (2017) that indicated the significant effect of WOM on the purchase intention but not for eWOM. Based on the findings, eWOM (reference to social media to support the decision being made, information seeking, review of others) only contributed to the attitude formation toward luxury fashion brands, but it had little influence on the consumers' intentions to purchase luxury fashion items. The plausible explanation is that reviews, recommendations, and information provided online are subject to trust and authenticity issues (Milaković et al., 2017), thus reducing the influence of eWOM on the purchase intention. Perhaps, eWOM such as that from social media, is shared by an unknown or distant acquaintance. The higher the distance between observer and subject, the lower the effect of the message/signal would be on other behavior (Warren and Mohr, 2017). Besides, Barnes (2015) indicated that the impact of eWOM on the purchase decision was product category-specific, especially for millennial consumers.
In this study, the attitude toward luxury fashion brands was proposed to mediate the relationship between the COO's image and eWOM on the intention to purchase luxury fashion brands. The findings were supported whereby the attitude toward luxury fashion brands mediated the relationship between the COO's image and eWOM on the intention to purchase luxury fashion brands. This study extends the literature on luxury fashion brands, using the millennials sample, indicating the importance of attitude formation toward luxury fashion brands. The COO's image had both direct and indirect effects on the intention to purchase a luxury fashion brand, whereby eWOM only had an indirect effect on the intention to purchase a luxury fashion brand.

Conclusion
In conclusion, this study contributes both theoretical and practical implications. The following subsections will further elaborate on the contributions.

Theoretical Implication
Based on the findings and the balance theory, both the COO's image and eWOM positively affect the attitude toward luxury fashion brands. Hence, millennial consumers would not solely depend on their perception of a particular object (COO's image) for the luxury fashion brand's context but may also depend on how others view it (eWOM). Based on the balance theory, the positive relationship between the elements would facilitate a favorable attitude formation. If one relationship between the elements is not positively related, it is not balanced, thus significantly affecting the attitude formation. This study extended the boundary of knowledge on the applicability of the balance theory on attitude formation toward luxury fashion brands and extended it into the intention to purchase. The findings add to the existing knowledge in the COO literature, especially for luxury fashion brands. As the past literature indicated that the effect of COO would be less when multiple cues were presented, this study revealed that other extrinsic cues, namely eWOM, failed to play a significant role in facilitating the purchase intention of luxury fashion brands. The effect size shows that the COO's image explains the more prominent effect on the intention to purchase, whereby eWOM explains a larger effect on the attitude toward luxury fashion brands.

Practical Implication
This study also helps the managers of luxury fashion brands to understand how the COO's image plays a significant role in forming millennials' attitudes toward luxury fashion brands and their purchase intentions. Millennial consumers perceive that developed countries of origin offer better product values than developing countries. In this case, luxury fashion brands from developed countries would quickly secure the millennials' market. Brands from developing countries should cautiously market themselves to the millennials' market. Marketers need to educate millennial consumers on the equality of the value and image of the luxury fashion brands from developing countries. A brand that goes viral through the eWOM medium could be the cheapest and easiest way to educate and communicate the value of that fashion brand. However, again, eWOM would only be influential in attitude formation but would not stimulate the intention to purchase luxury fashion brands. However, the luxury brands' manager should not neglect what consumers post on social media because eWOM still indirectly explains the intention to purchase.

Limitation
The main limitation of this study is the use of university students as the sample. Future research should consider general consumers as the sample. Besides, a comparative study among the consumer's generation (such as Gen X, Y and Z) also would add to the current understanding of the purchase intention and the behavior of luxury fashion brands. Future studies could also consider additional factors, or intrinsic or extrinsic cues, such as personal, interpersonal, and organizational factors.