Social media influencers and green buying behavior in emerging market FMCG sectors: Mediating role of brand equity and moderating effects of age and culture among university students

Social media has significantly changed how people interact with one another and how modern economies do marketing, and key among the factors influencing this transformation is the rise of SMIs (Vrontis et al., 2021; Syed et al., 2023). SMIs are individuals who have developed substantial followings on platforms like TikTok, Instagram, and YouTube and possess the power to affect their followers’ attitudes and behaviors (Lou & Yuan, 2019; Vrontis et al., 2021). Their persuasive capacity lies in their perceived credibility, authenticity, and relatability (Belanche et al., 2021). In green marketing, SMIs serve as digital opinion leaders who promote environmentally responsible behavior by endorsing sustainable brands or products (Kapitan et al., 2019). As a result, a study by Horrich et al. (2025) is consistent with other recent research, showing that the significance of SMIs in marketing is increasing and that a customized, influencer-centered approach is replacing traditional promotional techniques (Vrontis et al., 2021; Jin et al., 2021). These studies consistently demonstrate that SMIs enhance brand recognition and credibility and have a favorable impact on consumers’ purchasing decisions. (Nafees et al., 2021; Ashraf et al., 2023). The growing impact of SMIs is evident in emerging markets, where digital platforms often bypass traditional marketing channels and reach tech-savvy youth, including university students (Jin et al., 2020). However, most existing studies focus on Western or high-income contexts, with limited empirical exploration in African FMCG markets (Horrich et al., 2025).

GBB has developed into a strategy that considers social and ethical concerns in addition to environmental ones. A key component of this strategy is acknowledging the importance of incorporating people who are underrepresented in marketing strategies but have a sizable market potential (Sharma, 2021). As it is now necessary to consider young consumers when creating sustainable market practices, new viewpoints have altered our perceptions of university students, moving beyond the categorization of them as a construct (Prieto-Sandoval et al., 2022; Dziubaniuk et al., 2023; Leiva-Brondo et al., 2022). GBB refers to consumers’ tendency to purchase environmentally friendly, ethically sourced products produced through sustainable processes (Peattie, 2010; Joshi & Rahman, 2015). Such behaviors are influenced by environmental concerns, perceived consumer effectiveness, and social norms (Nekmahmud & Fekete-Farkas, 2020). Among younger consumers, green buying is often driven by exposure to social media campaigns and influencer endorsements that frame sustainability as an attractive lifestyle choice (Guiao & Lacap, 2022). However, the extent to which this behavior leads to actual purchasing decisions, especially in low-income contexts where price sensitivity is high, remains underexplored.

PDs are the consumer’s final action in selecting and acquiring a product based on individual preferences, perceived value, social influence, and brand familiarity (Yadav & Pathak, 2017). In sustainable consumption, GBB is often a precursor or predictor of PDs (Li et al., 2022; Song et al., 2019). While previous studies have highlighted the positive correlation between green attitudes and sustainable purchasing (Nekmahmud & Fekete-Farkas, 2020), less is known about how this relationship manifests in FMCG contexts within developing economies, where affordability and accessibility are key factors in decision-making.

BE, defined as the value of a brand in the mind of a consumer, significantly influences purchase behavior (Aaker, 1991; Kim & Hyun, 2011). Influencer-driven marketing can enhance perceived BE by increasing exposure, fostering emotional engagement, and providing social proof (Kapitan et al., 2019). In the FMCG sector, BE serves as a critical mediator between influencer content and purchasing outcomes; however, this linkage remains under-theorized in the sustainability literature (Ghorbani & Westermann, 2025; Gloria, 2025). Understanding how SMIs influence brand perceptions in green product categories presents a promising direction for bridging this gap.

CU shapes consumer attitudes, norms, and interpretations of environmental values (Hofstede, 2001). In collectivist societies like Ghana, social conformity, respect for authority figures, and community-based values can amplify the persuasive impact of SMIs, especially when sustainability is framed as a moral or social good (Phau & Teah, 2009; Gupta & Ogden, 2009). Despite these cultural nuances, most green consumption studies rely on Western individualist assumptions, neglecting the complex interplay between social media, cultural identity, and purchasing behavior in emerging markets (Vighnesh et al., 2023; Rahman & Luomala, 2021; Ur et al., 2023)

Age is critical in shaping how consumers engage with sustainability and digital platforms. University students, predominantly members of Gen Z or young millennials, are highly responsive to social media content, particularly that which aligns with their personal values or lifestyle aspirations (Yadav & Pathak, 2017; Vrontis et al., 2021). However, age rarely modifies existing models exploring green consumption in the influencer economy. Understanding how generational dynamics affect responsiveness to SMIs in emerging economies can enrich FMCG brands’ segmentation and targeting strategies.

While literature affirms that SMIs influence sustainable attitudes and purchase intentions (Lou & Yuan, 2019; Horrich et al., 2025), significant gaps persist:

• Limited studies explore how these influences operate in the FMCG industry in low-income or resource-constrained contexts.

• There is a lack of models integrating BE as a mediating variable and CU or age as moderators in green consumption frameworks.

• Cross-cultural validation of existing influencer marketing theories remains underdeveloped in African contexts, despite high youth engagement on social media platforms.

This study addresses these gaps by examining how SMIs influence GBB and PDs among university students in emerging markets, particularly in Ghana. It employs two complementary theoretical lenses:

• TPB (Ajzen, 1991) to understand intention-driven behavior.

• SIT (Kelman, 1958) to assess how identification with influencers shapes purchasing conduct.

By integrating psychological and sociocultural dimensions, this study advances the influencer marketing literature and contributes to the development of sustainable business practices in emerging markets.

This study draws on two foundational theories – the TPB (Ajzen, 1991) and SIT (Kelman, 1958) – to conceptualize how SMIs shape GBB and PDs among university students in the FMCG industry in emerging economies. The TPB asserts that behavior is determined by behavioral intentions, which are shaped by attitudes, subjective norms, and perceived behavioral control (Ajzen, 1991). In this study, GBB is conceptualized as a planned behavior resulting from environmental attitudes (shaped by influencer content), perceived social pressure (from peers or influencers), and students’ belief in their ability to make eco-friendly choices despite affordability or accessibility issues. Simultaneously, SIT explains how referent others influence individuals through compliance, identification, and internalization (Kelman, 1958). In the digital era, SMIs serve as reference figures whose values and behavior can influence followers’ own consumer choices. Identification with an influencer who promotes sustainable consumption can encourage followers to emulate such behavior, especially among younger consumers who are digitally embedded (Kapitan, et al., 2019).

The integration of TPB and SIT in the conceptual framework allows for a dual-perspective approach:

• TPB anchors the internal cognitive drivers of green purchasing.

• SIT captures the external social mechanisms, particularly the role of SMIs in shaping norms and values.

This theoretical integration justifies the inclusion of BE as a mediator (i.e., how the influence of SMIs is transferred through brand perceptions) and age and CU as moderators, capturing generational and sociocultural dimensions of planned and socially influenced behavior.

Based on the TPB (Ajzen, 1991) and SIT (Kelman, 1958), this study proposes that SMIs play a critical role in shaping GBB, which in turn influences PDs. Furthermore, BE is hypothesized to mediate this relationship, while age and CU are considered as moderators that may strengthen or weaken the influence pathways.

The FMCG industry in emerging economies, such as Ghana, is crucial due to the high frequency and low involvement of purchases, as well as the increasing visibility of sustainability narratives in marketing communications.

Below is the conceptual framework for the study (Figure 1).

Figure 1

Conceptual framework.

To investigate and meet the goals of the research, Gen Y (1981–1996) and Gen Z (1997–2012) within the Universities in Ghana were selected (69 universities). The purpose of sampling Gen Z and Y was their high demand for fashion design products in the FMCG industry. To explore the developed hypothesis, a questionnaire was distributed to university students over four months (February–May 2025). The research questionnaires were distributed to the students in person and electronically according to the participants’ preference. We sample the top ten universities in Ghana, with six public and four private. The ten universities were selected based on the following criteria: (a) These universities have campuses across the regions. (b) These universities gave us approval to conduct the research. (c) These consist of almost 90% of the entire student population in Ghana. To ensure generalizability and inclusion of all participants, this research employed a stratified sampling approach to select university students from all 16 regions. This method presents a unique opportunity to bring together participants from diverse universities with varying characteristics. Additionally, the method’s strength lies in selecting different levels (first, second, third, and fourth year) of students to reduce sampling bias. We employed the G-power software to determine the minimum sample size required to investigate the research hypothesis, which presents the null and alternative hypotheses in a graphical output. The X and Y plots display the parameter alpha (0.05), effect size (0.3), and the power (0.95). From the graphical illustration, the minimum sample size is 88 (Figure 2). A total of 664 samples were submitted to participants in 69 tertiary universities. However, the final sample size necessary for analysis was 576 (86.7%) to unravel the relationship between the developed hypothesis and the SMIs, GBB, CU, BE, and PDs. Table 2 presents the detailed outcome of participants’ bio information.

Figure 2

G-power sample size.

To ensure the reliability of the questionnaires adopted and modified to meet the research objectives, the researchers invited a select group of university students to complete the questionnaire. This pilot test aimed to determine the duration for participants to complete one set of questionnaires. Additionally, the pilot test helped identify the gaps these participants pointed out for correction. The limitations and ambiguities were addressed prior to the commencement of the survey process.

In this inspection, all the variables of the survey questionnaires developed for the research objectives and conceptual model were adopted from mature literary works (Table 1 and Table A1 – Appendix 1). Responses to the constructs were on a five-point Likert scale, ranging from 1 (strongly agree) to 5 (strongly disagree). To capture the construct of SMIs four items were adapted (Lou & Yuan, 2019; Belanche et al., 2021), GBB consisted of four items (Peattie, 2010; Joshi & Rahman, 2015; Yadav & Pathak, 2017), BE utilized four operationalization (Aaker, 1991; Kim & Hyun, 2011; Chen & Chang, 2012), PDs four indicators (Yadav & Pathak, 2017; Paul et al., 2016), and CU was measured with four items (Hofstede, 2001; Sharma, 2014; Nguyen et al., 2017). The developed questionnaires were reviewed by three researchers in the Department of Management and Marketing to ensure clarity.

To strengthen the robustness of the analysis, control variables such as gender, income, and education are included, given their established influence on GBB. Prior studies have shown that women often display stronger pro-environmental attitudes than men, income levels influence the affordability of eco-friendly products, and higher education is linked to greater sustainability awareness. Controlling these factors helps reduce omitted variable bias. It ensures that the effects of SMIs, BE, and CU on GBB are not confounded by underlying demographic differences (Rodrigo & Mendis, 2023).

Data collected from the repondents as stated in Table 2 indicate that males 211 (36.6) females 365 (63.4) confirms females huge desire for fashion products. Regarding the age of the respondents, the majority fall within the 26–30 bracket (38%), and the lowest is 36–40 (14.4%). Additionally, the bio information on education revealed that participants with a bachelor’s or undergraduate degree and those with a master’s or graduate degree had the highest participation, at 263 and 139, respectively. Finally, the faculty with the most participants was Education and Business Administration.

The structural equation model of Partial Least Squares (SEM-PLS) was employed to attain and present a significant estimation of the research hypothesis developed. The analysis was conducted using the SmartPLS 4 software. Previous literary works have highlighted the material significance of SEM-PLS over assessment, as seen in the hierarchical regression model (Takyi et al., 2024). First, SEM-PLS has two forms: (a) the measurement model, which assesses the model’s reliability and validity. (b) The structural model assesses the hypothesis developed with a bootstrapping of 10,000 iterations. Hair et al. (2019) explained in their article that when complex models involve moderation and mediation constructs, it is suitable to adopt the SEM-PLS approach.

The initial assessment of SEM-PLS using SmartPLS involves a measurement that examines the reliability and validity of the constructs. The PLS algorithm, as illustrated in Table 3, indicated that for the constructs assessed, the loading was above the threshold value of 0.7 (Hair et al., 2022). The consistency and reliability were assessed using Cronbach’s alpha and composite reliability. Takyi et al. (2024) stated that a value greater than 0.7 is considered an acceptable reliability and consistency check for the model developed to carry out the structural assessment of the hypothesis. Furthermore, the validity was investigated using the average variance extracted, which was found to be within the required figures of > 0.50. The findings presented the R-square value of PDs (0.798), GBB (0.661), and BE (0.427).

The issue of multicollinearity is serious when collecting and analyzing self-administered questionnaires. Therefore, to mitigate these issues, we are employing the variance inflation factor (VIF) to check for common method bias (CMB). Hair et al. (2022) and Takyi et al. (2024) indicated that a value less than 5 is considered ideal. Table 3 presents the collinearity of all the constructs, with values ranging from 1.656 to 4.494. We concluded that there is no issue of CMB that will impact practical policy for the research.

Estimating SEM-PLS is a significant test for the validity of the constructs; as such, the two most common validity tests are the Fornell-Larcker and HTMT approaches. Tables 4 and 5 demonstrate the outcome of the two approaches. According to Hair et al. (2022) and Takyi et al. (2024), all constructs under the two estimations should be below 0.90. The findings indicate that all the constructs were within an acceptable range. Also, the diagonals of the Fornell-Larcker are the square root of the AVE, which is vilified.

The empirical findings presented in Table 6 and Figure 3 indicate that SMIs had a strong, favorable, and significant affiliation with GBB (H1: Coeff = 0.802, T = 24.583, prob = 0.000). Again, GBB illustrates a direct positive and significant nexus with PDs. H2: Coeff = 0.813, T = 31.783, prob = 0.000. Similarly, the outcome displayed a direct relationship between SMIs and BE (H3), with a Coeff = 0.653, T = 20.850, prob = 0.000. Surprisingly, BE exhibited a low, insignificant association with GBB among university students (H4: Coeff = 0.017, T = 0.477, p = 0.634).

Figure 3

Structural model output.

The research investigated the mediating effects of BE and GBB on both SMIs and PDs of Gens X and Y within Ghanaian universities. The outcome from the PLS-specific indirect model suggests that GBB serves as an intermediary in the affiliation between SMIa and PDs. Table 6 demonstrated SMIs → GBB → PDs Coeff = 0.652, T = 18.657, prob = 0.000. This implies that when SMIs affect the attitude of Gens Y and Z toward their PDs, they should channel through GBB. Contrary to these results is the insignificant mediating influence of BE. Table 6 illustrates a marginally positive mediation of SMIs → BE → GBB Coeff = 0.011, T = 0.467, prob = 0.634. The findings of the hypothesis suggest that the desire of SMIs to promote GBB among Gens Y and Z should not be directed through BE.

To identify the interaction nexus between age and CU on GBB and consumer PDs of Gens Y and Z, we performed a 10,000-iteration bootstrapping. The findings, as illustrated in Figure 5 and Table 6, demonstrate that the researchers tested the direct effect of the moderator on its endogenous variables. Age → GBB (Coeff = 0.017, T = 0.355) had a substantial positive direct effect on GBB in Gens Y and Z universities in Ghana. However, when it was moderated, Age*SMIs → GBB (Coeff = −0.001, T = 0.013, prob = 0.990), the results indicated a negative and insignificant moderation influence. This implies that age is not a material component that SMIs in Ghana can use to attract university students. Second, we tested for the direct effect of CU on PDs (Coeff = 0.138, T = 4.201), indicating that CU had a favorable and significant impact on consumer purchasing decisions. When CU was introduced as a moderator, the relationship changed to an inverse significant effect (CU*GBB → PDs; Coeff = −0.044, T = 2.399, prob = 0.016). Figure 4 presents the slope analysis of the moderating effect.

Figure 4

Slope analysis of CU*GBB → PDs.

Figure 5

Slope analysis of Age*SMIs → GBB.

This study advances theory on social media influence and green consumerism in several ways. First, it extends the TPB and social influence by integrating BE as a key mediator. While TPB highlights attitudes and norms, our results indicate that SMIs operate through brand-related cognition: high-quality influencer messages enhance brand trust and image, thereby driving green purchase intentions. This mechanism adds nuance to existing frameworks and aligns with meta-analytic evidence that influencer effectiveness hinges on credibility and attractiveness (Barari et al., 2025; Munaro et al., 2024). Second, by explicitly testing age and CU as moderators, we highlight boundary conditions of influencer impact. Although age effects are subtle, our work underscores the importance of demographic heterogeneity. CU, modeled by collectivism/individualism, emerges as a potent factor, echoing Leonidou et al. (2010) and Sharma (2010), which confirms that collectivist values amplify pro-environmental action. In doing so, we address calls for more cross-cultural green marketing research and illustrate how cultural psychology shapes the SMI → GBB → PDs pathway. Together, these contributions enrich the sparse literature on influencer-driven green behavior in emerging markets, demonstrating theoretically that influencer attributes, mediated by BE, impact sustainable outcomes under different demographic contexts (Munaro et al., 2024).

The implications for policymakers, managers, marketers, and SMIs are significant. First, brands should partner with SMIs who are perceived as knowledgeable and sincere about sustainability. The findings of this study imply that SMIs with strong environmental expertise and authenticity (rather than mere popularity) most effectively build BE and green engagement (Barari et al., 2025; Munaro et al., 2024). Marketing campaigns must therefore vet influencer credibility and environmental alignment to avoid skepticism over “greenwashing.” Second, since BE mediates the effect of SMIs on GBB, practitioners should design SMI content that explicitly reinforces brand values. Endorsements should highlight a brand’s eco-friendly features (quality, reputation, loyalty) to translate SMI buzz into tangible brand trust. This means coordinating influencer messages with broader brand strategy (emphasizing certifications, environmental initiatives) so that students internalize a positive brand image when making green purchases. Third, companies should tailor influencer campaigns to different age segments. Younger students (18–24) may respond strongly to TikTok or Instagram green influencers, whereas older students might require more detailed, informational content (blogs or podcasts) about product sustainability. Pilot testing suggests that the SMI effect is weaker in older cohorts, so firms might use alternative channels (expert endorsements, traditional media) alongside influencer posts for older segments. Fourth, marketers in collectivist emerging markets should frame green messages in communal terms. For example, campaigns can highlight how sustainable choices benefit family or community, leveraging collectivist values to boost green purchases. In contrast, in more individualistic contexts (or minority subcultures), messages might stress personal health or status benefits of eco-brands. A one-size-fits-all campaign risks missing the cultural nuance that we found significant: collectivist students translated green attitudes into action more readily.

Although the study offers significant insights, it does have some limitations. First, the study surveyed university students in one emerging-market setting (Ghana), which limits its generalizability. Future studies should test these relationships among consumer groups (working adults, rural communities) and other emerging markets (Asia, Latin America) to enhance generalizability. Second, the data collected are cross-sectional and self-reported, so causality cannot be established. Future studies should therefore consider experimental or longitudinal studies, which could better confirm the direction of effects. Third, cultural orientation was captured broadly (e.g., collectivism), but CU is multi-dimensional. Future studies can therefore measure individualism explicitly and consider other Hofstede dimensions (e.g., uncertainty avoidance), which might also moderate green behavior. Fourth, focusing on the FMCG sector restricts applicability to other sectors. Future studies should consider other sectors, such as fast fashion or tech products, which may produce different influencer effects. The research was biased in terms of participants, as 63.4% (365) were females, which may influence the generalizability of the results and their applicability to policy. Therefore, future investigation can include more male participants to compare with the current research findings.