Is brand activism an emotional affair? The role of moral emotions in consumer responses to brand activism

Abstract

As brand activism continues to move up the corporate agenda, a more comprehensive understanding of its effects is needed. This paper contributes to the evolving research by investigating moral emotions (i.e. other-praising and other-condemning emotions) as mediating factors and consumer-brand identification (CBI) as a moderator that shape consumer reactions to brand activism. Three scenario-based experiments on two divisive topics show that activist messages elicit moral emotions that determine how individuals respond to them, depending on whether or not they agree with the brand’s stance. Moreover, this effect of (dis)agreeing with the brand’s stance on brand attitude is moderated by CBI. In case of a strong identification, an activist message does not affect brand attitude as CBI attenuates the activation of moral emotions – both in the positive case of agreement and in the negative case of disagreement. Finally, brand activism may counter the brand’s social goals, as it disproportionally motivates opponents of the brand’s stand to advocate their own contrary views on the contentious issue. In sum, these findings underscore both the emotional nature of consumer reactions to brand activism and the high level of social responsibility of companies that position themselves as political actors.

Introduction

In today’s increasingly polarised world, companies and their brands have become political actors (Korschun et al. 2020). As a consequence, a growing number of businesses are willing to speak out on morally charged issues, such as abortion rights and discrimination (Marketing Week 2022). This development seems in line with consumer demands, as consumers who feel strongly about a controversial issue expect brands to take a clear stance on it and, ideally, even lead the way in social and political change (Mukherjee and Althuizen 2020; Swaminathan et al. 2020; Vredenburg et al. 2020). Therefore, it is tempting for brands that want to stay relevant to take a stand on divisive socio-political issues (Koch 2020). With engaging in a polarised public debate, however, a brand expresses its values that are only shared by certain consumers (Hydock et al. 2020; Wannow and Haupt 2022). Accordingly, companies who engage in brand activism might be praised by consumers supporting the brand’s stance while receiving backlash from others.

Research has already begun to investigate the positive and negative effects of brand activism on consumer responses (e.g. Mukherjee and Althuizen 2020; Schmidt, Ind, Guzmán, and Kennedy 2021) and has found it to be a risky strategy as opponents may exhibit disproportionately negative reactions (Jungblut and Johnen 2021; Mukherjee and Althuizen 2020). However, the psychological mechanisms mediating and moderating the effects of brand activism on consumer responses, such as brand attitudes, have been identified as a research gap (Mukherjee and Althuizen 2020; Swaminathan et al. 2020). Surprisingly, emotions have only been considered marginally so far, although they are a key variable in consumer behaviour (Bagozzi et al. 1999; Bruno et al. 2022) and strong emotional responses to brand activism are a reality (Vredenburg et al. 2020).

This research gap is addressed by our first and main research objective. Based on the observation that more and more companies are making political statements and thus contributing to morally charged public debates, we seek to provide insights into emotional reactions triggered by brand activism. We propose and show that moral emotions are crucial mediators between brand activism and consumer responses. Moral emotions typically occur whenever (relevant) moral principles are violated or upheld (Haidt 2003), as is usually the case when brands make activist statements (Sibai et al. 2021). Previous studies have shown that the moral values promoted by a brand can directly trigger moral emotions, such as contempt, anger, and disgust (Haidt 2007; Tangney et al. 2007; Xie et al. 2015). However, the role of both positive and negative moral emotions in consumer responses to activist statements by brands has not yet been investigated.

The second and related objective results from the consensus that brand activism should not be motivated (only) by selfish goals, such as bolstering brand attitude, but by a higher purpose, i.e. creating awareness for social issues and, ultimately, social change (Hajdas and Kłeczek 2021). Despite the widely acknowledged purpose-driven nature of brand activism (Koch 2020; Vredenburg et al. 2020), research to date has focused on brand- or company-related outcomes. To narrow this gap, the model considers an individual's willingness to advocate for the themed issue as a social outcome of brand activism. Consumers’ support for a particular issue can be linked to moral emotions since they compel individuals to act (Haidt 2003). Moral emotions are thus directly related to an individual's behaviour (Frijda et al. 1989; Lazarus 1991; Xie et al. 2015). Therefore, in addition to our focal consumer response, brand attitude, it will be assessed whether brand activism and moral emotions lead to issue advocacy.

Further contributing to important boundary conditions of brand activism's impact, the third research objective is to examine the extent to which consumer-brand identification (CBI) moderates the effects of brand activism on consumers’ emotional and attitudinal reactions. This area of research is also considered to be incomplete to date (Mukherjee and Althuizen 2020). In the context of brand transgressions, service failures, and the like, the moderating effects of existing consumer-brand ties have already been confirmed (Ahluwalia et al. 2000; Antonetti and Anesa 2017; Khamitov et al. 2020). In the realm of activist brands, however, it has not yet been investigated whether existing bonds between consumers and brands strengthen or weaken the effects of brand activism on brand-related responses.

Based on three experiments, this paper makes several important contributions. First, positive and negative moral emotions are introduced as mediators of the effect of brand activism on brand attitude. Second, it is examined whether CBI moderates the effects of brand activism. Finally, we investigate the extent to which brand activism leads to a multiplier effect by encouraging individuals to engage with the controversial issue.

Our findings provide managers with insight into the potential consequences of brand activism for both the brand and society by revealing the underlying mediating mechanism of moral emotions. Brand activism, which by definition relates to morally charged issues, triggers both other-praising and other-condemning emotions that lead to specific consumer reactions. For example, if brands honestly advocate a position shared by the majority of their target group, mainly positive moral emotions such as gratitude are expected to strengthen the brand and increase proactive advocacy behaviour by consumers. In contrast, when most (potential) customers disapprove the brand’s stance, the brand will be target of negative moral emotions, i.e. anger, contempt, and disgust, with detrimental effects on attitude towards the brand. Next to this, CBI has been found to buffer the negative consequences of disagreements between consumers and brands. Thus, a strongly identified customer base can provide a safety cushion. In summary, these findings guide managers in their challenging decision of whether and how to engage in socio-political controversies.

Conceptual framework and hypotheses development

Brand activism

On a macro-level, the rise of brand activism can be understood as a consequence of increasingly polarised political debates worldwide. In the United States, a deep division over socio-political issues and a polarised discourse has been diagnosed (Grubbs et al. 2020; Ketron et al. 2022). In Europe, mainstream parties are on the decline with populist parties gaining traction in several countries, e.g. Italy and Hungary, indicating political polarization and conflicting socio-cultural values (Vachudova 2021). Political polarization shapes diverse consumption-related behaviours (e.g. Fernandes 2020; Jung and Mittal 2020) including attitudes and actions towards activist brands (Ketron et al. 2022; Matos et al. 2017; Mukherjee and Althuizen 2020). On a micro-level, more and more companies and their brands are becoming political actors by actively engaging in controversial socio-political debates in an attempt to influence consumers (Weber et al. 2021).

However, it has to be noted, that there is no consensus in management and marketing science on the fundamental question of whether it is the task of companies to engage in the field of corporate social responsibility (CSR) or even brand activism (Ferrell 2022). Already CSR is challenged by calling for a separation of government and business (Ahluwalia 2022) and questioning whether business leaders are qualified to influence public policy decisions (Cronin and Kerr 2022). If a company acts contrary to the aspirations of society, laws and public pressure from the government should make the company comply (Gaski 2022). In contrast, proponents of respective corporate action argue that responsibility for societal issues should be a shared responsibility rather than one to be strictly divided between agents (Demuijnck and Murphy 2022). This joint responsibility should serve as a bridge between the company and external stakeholders (Martin and Burpee 2022). Apart from this theoretical debate, brand activism has been obviously on the rise in the last years (Marketing Week 2022; Moorman 2020).

Brand activism is “a purpose- and values-driven strategy in which a brand adopts a non-neutral stance on institutionally contested socio-political issues, to create social change and marketing success” (Vredenburg et al. 2020, p. 446). Academics consider it an evolution of CSR (Sarkar and Kotler 2018). While CSR activities are widely supported in society, scholars accentuate the contentious, often polarising nature of the issues addressed through brand activism as a differentiator. Importantly, authentic brand activism is rooted in both corporate values and practices; otherwise consumers might accuse the brand of speaking up solely for egoistic motives and, thus, of woke washing (Mirzaei et al. 2022; Vredenburg et al. 2020). Still, typical topics underlying brand activism are often not linked to the core offering of the company (Vredenburg et al. 2020).

Due to its controversial nature, brand activism is considered a risky strategy, as it can potentially alienate stakeholders (Bhagwat et al. 2020). With more consumers demanding brands to engage in socio-political debates, companies must weigh the risk of either taking the 'wrong' side on a controversial issue versus not taking a position at all (Bhagwat et al. 2020; Korschun et al. 2016; Swaminathan et al. 2020). Furthermore, engaging in a polarised public debate could limit the brand’s strategic flexibility. Public opinion is frequently quite dynamic, but a brand risks being locked into a particular stance to maintain its authenticity and credibility (Sibai et al. 2021). Apart from the potential downsides for the brand, it is unclear whether or under which circumstances brand activism contributes to a further polarization of society. There is surely a certain risk – at least for individuals with extreme political ideologies – that merely addressing a contested political issue can reinforce polarization processes regardless of the stance taken due to biased information processing (Bliuc, Bouguettaya and Felise 2021; Wojcieszak et al. 2018; Wojcieszak and Warner 2020).

Although still at an early stage, research on brand activism already offers some valuable insights, especially on brand-related outcomes and their antecedents (see Table 1 for an overview). Several studies confirm the alignment between the consumer’s and the brand’s stance as a key predictor of responses to brand activism, such as purchase intention (Dodd and Supa 2014; Jungblut and Johnen 2021; Mukherjee and Althuizen 2020). However, studies have identified a negative net effect of brand activism on brand attitude and purchase intention (Mukherjee and Althuizen 2020), market share (Hydock et al. 2020), and firm value (Bhagwat et al. 2020). Social impact, ultimately one of the key goals of brand activism, on the other hand, has not been the focus of empirical studies. Knowledge about the role of mediating factors also seems to be patchy. Exceptions to this are the constructs identification (Hydock et al. 2020; Mukherjee and Althuizen 2020) and perceived hypocrisy (Korschun et al. 2016).

Table 1 Overview of prior research regarding brand activism (BA)

Research suggests that the overall impact of brand activism depends on various constraints. For instance, brand activism has been found to be riskier for larger companies as they are likely to lose more customers than they gain (Hydock et al. 2020). Furthermore, the company’s value orientation matters. Individuals expect value-oriented (vs. results-oriented) companies in particular to take a stand; otherwise, they would be perceived as hypocritical (Korschun et al. 2016). Accordingly, the relevance of brand identity congruence, authenticity, and credibility in driving brand equity and social change has been emphasised (Bhagwat et al. 2020; Eilert and Nappier Cherup 2020; Mirzaei et al. 2022; Sibai et al. 2021; Vredenburg et al. 2020). As with alternative (especially societal) outcomes of brand activism and additional mediators, there is a research gap regarding potential moderators. In the following, we will first introduce moral emotions as relevant mediating factors and link them to consumer responses, before discussing CBI as a moderating factor.

Moral emotions and consumer behaviour

Moral judgment and decision-making have long been understood from a predominantly cognitive perspective relying on the assumption of reasoned, deliberate thought processes (Haidt 2001, 2003; Vélez García and Ostrosky‐Solís 2006; Xie et al. 2015). In the domain of brand activism research, Mukherjee and Althuizen (2020) explicitly assume a deliberate moral reasoning process to explain consumer reactions to brand activism. However, today, the idea that individuals make intuitive – often emotionally charged – moral judgements, is established as a complementary theoretical perspective (Brescoll et al. 2018; Cova et al. 2015; Fernandes 2020; Xie et al. 2015; Zollo 2021). The relevance of moral intuitions was initially proposed by Haidt (2001). According to his social intuitionist model “moral judgment is caused by quick moral intuitions and is followed (when needed) by slow, ex post facto moral reasoning” (Haidt 2001, p. 817). The (biased) reasoning why something is considered moral or immoral often only serves to justify and explain the automatic reaction (Haidt 2001; Mercier and Sperber 2011; Xie et al. 2015). These intuitive responses to morally significant stimuli often manifest themselves as moral emotions (Grappi et al. 2013b; Zollo 2021). Additionally, Zollo (2021) argues that moral emotions also play a crucial role in the more systematic and conscious moral reasoning process that can follow the first moral intuition.

Moral emotions emerge when a specific pattern of appraisal occurs, i.e. witnessing reprehensible or exemplary moral expressions or actions that affect others (Grappi et al. 2013a; Haidt 2003). Haidt (2003, p. 853) describes these emotions as ‘disinterested’, meaning that they are “linked to the interests or welfare either of society as a whole or at least of persons other than the judge or agent”. Thus, whenever moral or ethical principles and stakes of third parties are involved, moral emotions arise (e.g. Romani et al. 2013a; Xie et al. 2019; Xu et al. 2021; Zollo 2021). The relevance of morality concerns in triggering emotions is also included in the cognitive appraisal theory. In this dominant theory of emotions, normative or moral compatibility has been identified as a relevant appraisal dimension (Frijda 1986; Johnson and Stewart 2005; Scherer 2001).

Further, the experience of moral emotions can motivate individuals to act morally (Tangney et al. 2007). Compared to other (basic) emotions, such as joy, sadness, or fear, they are more likely to trigger prosocial behaviour (Haidt 2003; Zollo 2021). This applies to both positive and negative moral emotions. Thus, even emotions with a negative valence can elicit positive (i.e. prosocial) responses. For example, moral anger stimulates people to punish the ‘offender’, or to redress a perceived injustice (Haidt 2003; van Doorn et al. 2014).

In the most common conceptualisation of moral emotions, Haidt (2003) introduces four categories, namely (1) other-condemning, (2) other-praising, (3) self-conscious, and (4) other-suffering emotions. These four categories differ in terms of their emotional valence and their focus, i.e. is the emotion directed at the transgressor (self vs. other) or a victim. On the one hand, when someone other than the self has violated or upheld a moral value, (1) other-condemning or (2) other-praising emotions arise. On the other hand, when the moral act was performed by oneself, individuals experience (3) self-conscious emotions like guilt or pride. Lastly, when there is a clearly identifiable victim of a moral transgression, (4) other-suffering emotions like compassion are typically aroused (Greenbaum et al. 2020; Haidt 2003; Zollo 2021).

This paper focuses on other-directed moral emotions, i.e. (1) other-condemning and (2) other-praising emotions. The brand as the sender of the activist message is the transgressor or advocate of a moral value and will, therefore, be the focus of the triggered moral emotions. As the message might hurt or support one’s moral principles, we include both positive and negative moral emotions. Both categories are well-established in psychological research (Hutcherson and Gross 2011; Landmann and Hess 2017; Thomson and Siegel 2017) and have been widely addressed in consumer research, particularly in the context of (un)ethical corporate behaviour, such as CSR (Kim and Park 2020; Romani et al. 2013a), cause-related marketing (de Vries and Duque 2018), and corporate transgressions (Antonetti and Maklan 2016; Grappi et al. 2013b; Romani et al. 2013b; Xie et al. 2015). Other-condemning emotions can be defined as “negative feelings towards others because they have violated moral standards” (Greenbaum et al. 2020, p. 96). They are typically represented by contempt, anger, and disgust (CAD) that can be collectively considered a hostility or resentment factor (Grappi et al. 2013a; Greenbaum et al. 2020; Xie et al. 2015). In contrast, other-praising emotions are triggered by exemplary moral behaviour and, therefore, represent the “brighter side to the moral emotions” (Haidt 2003, p. 862). Gratitude and elevation are the most prototypical representatives of this category (Greenbaum et al. 2020; Haidt 2003). People are not only moved by morally laudable actions, but also feel encouraged to act prosocial. That prosocial tendency may be focused on a particular benefactor, e.g. on a charitable brand. In addition, other-praising emotions are able to motivate people to become better persons and to positively contribute to society in general (Algoe and Haidt 2009; Greenbaum et al. 2020; Tangney et al. 2007).

The role of moral emotions as a mediator between brand activism and consumer responses

In an activist context, other-condemning emotions can be expected to occur when the brand’s stance clashes with consumers’ moral standards, i.e. the brand’s stand is judged as morally questionable or even repulsive. In this regard, brand activism that contradicts the values of a consumer can be categorised as a values-related brand crisis (as opposed to a performance-related brand crisis) – at least from the opponents’ point of view. This type of crisis “involves social or ethical issues surrounding the values espoused by the brand” (Dutta and Pullig 2011, p. 1282) and is not directly linked to the product or service offering (Dutta and Pullig 2011; Liu et al. 2018; Pullig et al. 2006). For example, when a brand speaks out in favour of abortion rights, this might evoke anger, contempt, and even disgust among pro-life supporters. The arousal of negative (moral) emotions by negative values-related brand crises or ethical transgressions has been confirmed in several studies (Baghi and Gabrielli 2019; e.g. Grappi et al. 2013a; Xie et al. 2015).

Effects on brand-related responses

Other-condemning emotions have been associated with various negative brand-related consumer responses, such as negative WOM, boycotting and protest behaviour (Antonetti and Maklan 2016; Grappi et al. 2013a; Xie et al. 2015, 2019). We argue that individuals in this emotional state should develop a negative brand attitude besides these behavioural reactions. It has been established that emotions in general have spill-over effects and influence consumers’ attitudes and judgments (Eggert et al. 2019; Palmatier et al. 2009; Romani et al. 2013a), in particular when the emotions are directly related to the attitude object and are not just incidental (Pham 2007). This is the case with brand activism which represents a purposeful brand action. Accordingly, other-condemning (and other-praising) emotions are clearly targeted at the brand (as opposed to joy or anxiety, for example). Grappi et al. (2013b) confirmed the mediating effect of anger (and gratitude) on the relationship between a company’s offshoring decision, which is perceived as a moral issue by the public, and attitudes and behaviours toward the company. Leak et al. (2015) found that consumer’s ideologically incongruent beliefs can cause him or her to experience anger-induced attitude change after a manager took a stance on same-sex marriage. They argue that the manager is directly tied to the brand or company; therefore, the experienced anger functions as new information to update attitudes towards the brand(s). The mediating role of contempt in the context of value-related negative brand publicity and corresponding brand evaluations has been confirmed by Liu et al. (2018). Thus, it can be hypothesised that consumer-brand disagreement on a particular topic will trigger other-condemning moral emotions, and these will subsequently negatively affect brand attitude:

H1a

Consumer-brand disagreement has a negative effect on brand attitude, mediated by other-condemning emotions.

Conversely, individuals who approve of a brand’s activism, i.e. perceive the brand’s stance as consistent with their own moral standards, will experience other-praising emotions (Mukherjee and Althuizen 2020). Consumers might feel supported, touched, or inspired by the brand advocating a controversial issue and appreciate the risk the brand takes (Algoe and Haidt 2009). In this regard, if there is an agreement between the consumer and the brand, feelings of gratitude or elevation are likely to occur as the brand openly stands up for the consumer's moral values. As already argued above, emotions elicited by brand activism can be attributed to a clear source and have, in turn, a clear target, namely the brand. In consequence, it is likely that these positive moral emotions caused by purposeful brand actions will shift consumer-brand attitudes in a positive direction (Pham 2007; Wannow and Haupt 2022). The relationship between other-praising emotions and brand-related consumer responses, mainly behavioural outcomes, has been confirmed in several studies (Eggert et al. 2019; Kim and Johnson 2013; Kim and Park 2020; Palmatier et al. 2009; Romani et al. 2013a; Thomson and Siegel 2017; Xie et al. 2015). Integrating an attitudinal outcome, Grappi et al. (2013b) show that gratitude mediates the effect of perceived corporate offshoring strategies on attitudes toward the company and word-of-mouth. More precisely, consumers felt grateful for the company’s decision against offshoring which in turn led to more positive attitudes and behaviours. Based on the theoretical and empirical evidence presented, we hypothesise:

H1b

Consumer-brand agreement has a positive effect on brand attitude, mediated by other-praising emotions.

Effects on issue advocacy

Campaigning on socio-political issues may not only elicit brand-related consumer responses (Bridger and Wood 2017; de Vries and Duque 2018), but also prosocial behaviour, including issue advocacy (Romani et al. 2013a; Romani and Grappi 2014). Issue advocacy captures a wide range of individual or collective contributions to a socio-political debate. These include participation in demonstrations or public debates (Matthes et al. 2010), as well as social media engagement, such as following or complaining to an opinion leader online (Ferrucci et al. 2020; Xie et al. 2015).

There is empirical evidence that CSR-messages or -activities positively influence issue advocacy (i.e. prosocial behaviour) (Mantovani et al. 2017). It has been shown that CSR increases consumers’ donation intentions and actual behaviour via the other-praising emotion elevation (Romani and Grappi 2014). On the other hand, negative emotions such as anger might motivate consumers to publicly defend their own stance, i.e. to advocate against the brand’s position. Anger as part of the other-condemning emotions is often linked to a high level of activation (Rucker and Petty 2004), leading people to blame the source of their anger and seek revenge or retaliation (Abdelwahab et al. 2022; Romani et al. 2015). It represents a reaction to an “intentional goal blockage” (van Doorn et al. 2014, p. 261) which triggers a strong tendency to act, i.e. to remove this barrier and accomplish the pre-established goal. For instance, anger, and to a lesser extent enthusiasm, strengthens the tendency of individuals to sign a petition, participate in a demonstration, or engage in volunteer work (Feldman and Hart 2016; Iyer et al. 2007; Valentino et al. 2011). Previous findings also suggest that anger increases individuals’ active resistance to brands (Feldman and Hart 2016; Romani et al. 2015). All of this reflects expressive behaviour in the sense of issue advocacy.

When comparing positive and negative moral emotions, anger was found to be a stronger driver of prosocial actions than empathy or guilt, for instance in terms of advocacy for disadvantaged people (Montada and Schneider 1989; van Doorn et al. 2014) or active political participation (Valentino et al. 2011; van Doorn et al. 2014). Therefore, negative moral emotions are hypothesised to have a stronger influence on activism-related outcomes such as issue advocacy than positive moral emotions:

H2

Consumer-brand disagreement has a stronger effect on issue advocacy compared to agreement, mediated by other-condemning and other-praising moral emotions.

Consumer-brand identification as a moderator

The role of CBI in negative brand events

Responses to brand activism might depend on consumer-brand relationships. Previous research has focussed in particular on the question how existing attitudinal ties between consumers and brands influence the effect of negative – rather than positive – brand-related events, such as product or service failures and moral transgressions (e.g. Ahluwalia et al. 2000; Antonetti and Anesa 2017; Davvetas and Diamantopoulos 2017; Khamitov et al. 2020; Trump 2014; Zhang et al. 2020). As consumers who are in opposition to a brand’s stance might view such brand activism as a kind of transgression, it seems promising to extent this line of research to the phenomenon of brand activism.

In the present study, the pre-existing relationship between consumer and brand is captured by CBI (Millán and Diaz 2014; Stokburger-Sauer et al. 2012; Tuškej and Podnar 2018). This construct is based on the long-established notion that consumers seek out products and brands that help them define and express their identity (Belk 1988). CBI can be understood as "a consumer's perceived state of oneness with a brand” (Stokburger-Sauer et al. 2012, p. 407) which captures the degree of connectedness between one’s own and the brand’s identity (Davvetas and Diamantopoulos 2017). Accordingly, CBI increases when consumers feel that they share self-defining attributes with the brand. This ultimately leads to a sense of belonging (Lam et al. 2010; Tuškej et al. 2013).

Given the symbolic function of brands, consumers for whom the brand is a core element of their identity (high CBI) should find brand activism highly relevant to their sense of self (Escalas and Bettman 2005; Stokburger-Sauer et al. 2012). In contrast, individuals with low CBI do not feel deeply connected to the brand, so such brand actions are not seen as identity-relevant. Scholars have found that connected consumers process negative brand-related information in a biased way to protect their self-worth, i.e. by counter-arguing, justifying and rationalising the misstep, or decoupling the moral judgement from its source ('love is blind-effect') (Ahluwalia et al. 2000; Einwiller et al. 2006; Leak et al. 2015; Wang and Kim 2019). For example, motivated reasoning theory explains why highly identified consumers tend to engage in defensive processing of negative brand information to uphold favourable brand beliefs, while weakly identified consumers are more motivated to form accurate judgments (Einwiller et al. 2006).

For some types of transgressions, however, the opposite effect was observed ('love becomes hate-effect') (e.g. Antonetti and Anesa 2017; Einwiller et al. 2019; Grégoire and Fisher 2006). In addition to extremely negative information or information highly relevant to the self, moral transgressions also appear to trigger this reverse effect (Ma 2020; Trump 2014). Since consumers were found to perceive information about (im)moral behaviour highly diagnostic (Goodwin et al. 2014), a brand’s identity risks being severely damaged if it acts in a questionable or even reprehensible manner according to the individual’s (moral) standards (Antonetti and Anesa 2017; Einwiller et al. 2019; Trump 2014). Such transgression poses a significant threat to the self of consumers who strongly identify with the brand (Einwiller et al. 2019), as they use the brand as a source for shaping and portraying their selves (Johnson et al. 2011). In such a case, the negative information will be impossible to ignore (Einwiller et al. 2006). This is especially true because the perception of one's own morality is an essential component of the self-concept (Strohminger and Nichols 2014). Consequently, consumers who strongly (vs. weakly) identify with the brand are likely to react more extremely (Bhattacharya and Sen 2003) and to “reject more vehemently the company they no longer see as sharing their moral convictions or aspirations” (Einwiller et al. 2019, p. 4).

CBI and moral emotions

Transferring these considerations to consumer reactions in the context of brand activism, it can be assumed that a strong identification with the brand strengthens the emergence of other-directed moral emotions. Driven by a perceived threat to one’s identity, which is particularly severe at a high CBI, other-condemning emotions in particular should be more pronounced (Einwiller et al. 2019; Kim et al. 2020; Ma 2020; Zhang et al. 2020). This suggestion is consistent with findings by Ma (2020) that brand crises enhance perceived anger and disappointment when consumers feel that self- and brand-defining (vs. non-defining) attributes are being undermined. In other words: Individuals should react more strongly to moral violations with other-condemning emotions when the violations threaten to cast a bad light on oneself and one’s character, which is the case for highly identified consumers.

In addition to the motive of self-protection, the high expectations that connected consumers have of ‘their’ brands provide an additional explanation for the presumed amplification effect of CBI. Violations of these high standards lead consumers to experience pronounced negative emotions, such as betrayal and anger (Grégoire and Fisher 2008; Ma 2020; Zhang et al. 2020). Moreover, this reinforcing effect is supported by the cognitive appraisal theory, which posits the degree of personal or goal relevance as a key appraisal dimension (Nyer 1997; Scherer 1988; Watson and Spence 2007). In general, a relevant stimulus is more likely to trigger a strong emotional reaction than an irrelevant stimulus (Nyer 1997). Since a high CBI increases the personal relevance of brand activities, the stronger the person identifies with the brand, the stronger the emotions triggered by identity threats should be (Lam et al. 2010). Consequently, CBI can be expected to enhance the effect of disagreement with a brand’s stance on other-condemning emotions:

H3a

The effect of consumer-brand disagreement on other-condemning emotions is positively moderated by CBI.

Accordingly, pre-existing brand ties may amplify positive emotional responses to brand activism. Extending our line of reasoning to the positive case of consumer-brand agreement, we assume that highly identified consumers appreciate the pursuit of a socio-political or moral goal not only for its own sake, but also to verify their own moral self (Romani et al. 2013a). For these consumers (vs. consumers with low CBI), it provides an additional benefit if the brand is viewed as morally upright or ethical (Gao and Mattila 2016). For consumers who share a close relationship with the brand, it should also be easier to view the brand as a moral exemplar and, thus, experience feelings of elevation (Thomson and Siegel 2017). More generally, a high degree of perceived ‘oneness’ with the brand makes its commitment to shared values and purposes more self-relevant, which in turn evokes stronger positive emotions (Nyer 1997).

As empirical evidence on the effect of CBI in the context of positively perceived moral acts is scarce, we relate to adjacent contexts. For instance, in the field of donations, a positive link between identification and gratitude has been established (Kwak and Kwon 2016). Furthermore, a meta-analysis confirmed that identification has a positive impact on the effectiveness of cause-related marketing as well as the mediating role of positive (and negative) moral emotions (Fan et al. 2020), but without analysing a possible interaction. In total, the theoretical arguments lead to the following hypothesis:

H3b

The effect of consumer-brand agreement on other-praising emotions is positively moderated by CBI.

Empirical studies

Overview

Figure 1 provides an overview of the conceptual framework. We conducted three experimental studies to test the proposed hypotheses. Study 1 provides initial evidence for the mediating effect of moral emotions on the relationship between consumers’ brand activism agreement and their brand attitude (H1a, H1b). In addition, this first study analyses the moderating impact of CBI on the link between consumer-brand (dis)agreement and moral emotions (H3a, H3b). Whereas studies 1 and 3 are based on the controversial topic of illegal immigration, study 2 uses the heavily contested topic of abortion rights to replicate the effects from study 1. Study 3 extends the mediating role of moral emotions to a society-related outcome, namely issue advocacy (H2).

Fig. 1

Conceptual framework

Pre-studies

The aim of the pre-studies was to identify a socio-political issue suitable for brand activism and brands that can take an equally credible stand for both sides. To this end, a multi-method approach was applied. Secondary statistics were used to identify, on the one hand, topics that polarise the population (Appels et al. 2020) and, on the other hand, brands whose offerings target as large a proportion of the population as possible. The topic search also included recommendations in existing studies (e.g. Moorman 2020; Mukherjee and Althuizen 2020; Vredenburg et al. 2020). The attribution of brand activism to a real-life brand was intended to ensure that respondents perceived the scenarios as vividly as possible. This resulted in a longlist of 8 issues and 27 brands.

In a second step, a shortlist for each of these was created based on two qualitative interviews with experts in the field. This led to the two issues of illegal immigrants and abortion rights and narrowed the list of brands down to 10. To prioritise them, 80 Amazon MTurk panellists were surveyed. Participants were randomly assigned to one of four scenarios (pro-immigrants, anti-immigrants, pro-life, pro-choice (see Appendix 1 for details)) and asked to what extent they considered it realistic for the respective brand to hold the opinion portrayed in the scenario.

Five brands did not meet the requirement that both alternative stances on an issue must be equally realistic for the brand. Of the remaining brands, a hypermarket chain and a home improvement retailer were selected. For these two companies – compared to the three alternative brands – the respondents considered it most realistic for the brand to take the position outlined in the scenario.

Study 1

Objective and study design

Study 1 pursued two main objectives. First, to examine whether other-condemning and other-praising emotions mediate the effect of brand activism on brand attitude. Second, to investigate whether CBI moderates this mediation. Based on the pre-study, illegal immigrants were selected as the controversial issue and the hypermarket chain was chosen as the source of the brand activist statement.

Participants were randomly assigned to one of three scenarios (between-subjects design, see Appendix 1). In these, the brand either took a ‘pro-immigrants’ (i.e. illegal immigrants may stay) or a ‘anti-immigrants’ position (i.e. illegal immigrants have to leave). The third scenario (control group) included a non-activist PR-statement about the general culture of the company. In order to verify that the other-directed emotional responses to brand activism were not the result of a general, possibly diffuse emotional state, ‘negative issue-related emotions’ were also included in the model. We controlled for pre-brand attitude and respondent’s involvement with the controversial issue.

The framework introduced by Hill et al. (2021) was used to evaluate possible causes of an endogeneity problem (see Appendix 2). As it is important to address the causes of endogeneity both theoretically (problem-avoiding) and empirically (problem-identifying) (Antonakis et al. 2010; Ketokivi and McIntosh 2017), the techniques used in the studies focussed on making better choices in study design and analysis. Overall, the risk of endogeneity remaining despite all precautions is assessed as low.

Sample characteristics

Empirical analyses were based on data from 179 Amazon MTurk panellists (36.3% female; M_age = 36.93, SD_age = 9.78) who completed the online questionnaire for a small monetary compensation. To qualify, respondents had to have a human intelligence task (HIT) approval of 95% or higher and pass an attention check. On a continuum from 1 = ‘very liberal’ to 7 = ‘very conservative’, 39.1% identified themselves as (rather) liberal and 47.5% as (rather) conservative, with the remainder indicating they were politically neutral (M = 4.15; SD = 2.07).

Measures

A three-way split of the sample served as the categorical independent variable: individuals who agreed with the brand’s stance, those who disagreed, and the control group. To identify the first two groups, participants were asked to indicate the extent to which they approved the statement "all illegal immigrants should be asked to leave the USA irrespective of how long they have been here" (1 = ‘fully disagree’, 7 = ‘fully agree’). In the case of the anti-immigrants scenario, individuals who scored 3 or less were classified as disagreeing, while participants who scored 5 or more were classified as agreeing. 13 respondents with a score of 4 (neutral) were excluded from the analyses. In the case of the pro-immigrant scenario, the classification was mirrored.

Established scales were used to measure the constructs other-condemning emotions (CAD) (Xie et al. 2015), gratitude (Xie et al. 2015) and elevation (Xie et al. 2019) representing other-praising emotions (Xie et al. 2019), sadness and fear representing negative issue-related emotions (Izard 1977), pre- (covariate) and post-experimental (outcome) brand attitude (Mukherjee and Althuizen 2020), and CBI (Stokburger-Sauer et al. 2012). Brand attitude was measured with a seven-point semantic differential, while the other variables were measured with a seven-point Likert scale (1 = ‘fully disagree’, 7 = ‘fully agree’). Individual’s involvement with the issue was captured with a scale from Boninger et al. (1995). Each construct exhibited sufficient internal consistency as well as discriminant and convergent validity (Hulland et al. 2018) (see Table 2).

Table 2 Scale items and statistics

Results

Model 8 from PROCESS (Hayes 2021) was applied to test the moderated mediation model (see Fig. 2) including pre-brand attitude and issue involvement as covariates. Indirect effects were tested for statistical significance using 95% confidence intervals (based on 5000 bootstrap samples).

Fig. 2

Results from study 1

As hypothesised (H1a), consumer-brand disagreement (vs. control) increased other-condemning emotions (b = 4.50, t(170) = 5.18, p < 0.001), whereas (H1b) consumer-brand agreement increased other-praising emotions (b = 2.52, t(170) = 3.87, p < 0.001). At the same time, neither disagreement nor agreement triggered negative issue-related emotions (each p > 0.1). The moral emotions, in turn, influenced brand attitude, in each case in the expected direction: other-condemning emotions had a negative (b =  − 0.16, t(170) =  − 2.63, p < 0.01), other-praising emotions a strong and positive effect on the dependent variable (b = 0.52, t(170) = 7.69, p < 0.001). Brand attitude was directly influenced by consumer-brand disagreement (b = − 1.52, t(170) =  − 2.55, p < 0.05), but not by agreement (p = 0.54). Thus, H1a and H1b are supported (see data details in Table 3).

Table 3 Study 1: Conditional process model for moral emotions as parallel mediators, CBI as moderator, and brand attitude as outcome

Regarding H3a, CBI moderated the effect of consumer-brand disagreement on brand attitude via other-condemning emotions (index of moderated mediation = 0.10, 95%CI [0.02, 0.20]). For low and medium CBI levels [i.e. M–1SD = 2.88 and M = 4.68], the indirect effects were significant, whereas for a high CBI level [i.e. M + 1SD = 6.47], the effect became insignificant (CBI_low: b =  − 0.44, 95%CI [ − 0.81;  − 0.12], CBI_medium: b =  − 0.26, 95%CI [ − 0.50;  − 0.07], CBI_high: b =  − 0.09, 95%CI [ − 0.30; 0.07]).

Similarly, CBI moderated the effect of consumer-brand agreement on brand attitude via other-praising emotions (index of moderated mediation = − 0.24, 95%CI [ − 0.49, − 0.07]). More precisely, for a low level of CBI, the moderating effect was significant. Yet, for medium and high levels of CBI, the indirect effects became insignificant (CBI_low: b = 0.63, 95%CI [0.19; 1.20], CBI_medium: b = 0.20, 95%CI [ − 0.04; 0.43], CBI_high: b =  − 0.23, 95%CI [ − 0.69; 0.05]).

A high level of CBI seems to buffer the occurrence of moral emotions, whether condemning or praising, and to immunise the brand against brand activism. Conversely, the less individuals identify with a brand, the greater the leverage to influence their brand attitude through brand activism. As we found a buffering instead of the hypothesised reinforcing effect, H3a and H3b are not supported. When excluding the covariate issue involvement from the model, there are only minor differences with regard to the estimated parameter values, and the significance tests lead to the same result in each case. For reasons of parsimony, this covariate was therefore not included in studies 2 and 3.

Study 2

Objective and study design

Study 2 aimed to examine whether the findings of study 1 can be replicated in a different context. For this purpose, the controversial topic chosen this time was abortion rights, and the home-improvement retailer served as the activist brand. Participants were randomly assigned to one of three scenarios (between-subjects design). The brand took either a pro-life (i.e. against abortion) or a pro-choice stance (i.e. in favour of abortion rights) or, in a no stance-scenario, explicitly refused to take a stance. The first two scenarios were adopted from Mukherjee and Althuizen (2020) (see Appendix 1).

Sample characteristics

244 participants from MTurk (36.1% female; M_age = 37.37, SD_age = 9.85) completed the online questionnaire. As manipulation check, respondents had to correctly classify whether the brand was pro-choice or pro-life based on the scenario they had just read. Rated from 1 = ‘very liberal’ to 7 = ‘very conservative’, 44.7% identified themselves as (rather) liberal and 41.4% as (rather) conservative, with the remainder indicating they were politically neutral (M = 3.84; SD = 2.19).

Measures

The constructs consumer-brand agreement, other-condemning emotions, other-praising emotions, brand attitude, and CBI were measured with the same scales used in study 1. Since, according to the findings of study 1, gratitude and elevation are highly correlated, for reasons of parsimony, other-praising emotions were represented by gratitude in study 2. Each construct proved both valid and reliable (see Table 2).

Results

The moderated mediation model (see Fig. 3) proposed was measured using PROCESS model 8 (Hayes 2021), with pre-brand attitude serving as a covariate. Significance of the indirect effects were determined by 95% confidence intervals (based on 5000 bootstrap samples).

Fig. 3

Results from study 2

Overall, the results confirm that the effects discovered in study 1 are robust to an alternative issue, an alternative brand, and the inclusion of a deliberate no stance-statement (see data in Table 4). Consumer-brand disagreement (vs. no stance) increased other-condemning emotions (b = 2.86, t(227) = 3.93, p < 0.001). Likewise, consumer-brand agreement led to stronger other-praising emotions (b = 1.13, t(227) = 1.97, p < 0.05)). In turn, the two mediators influenced brand attitude (other-condemning emotions: b = − 0.08, t(227) = − 2.11, p < 0.05; other-praising emotions: b = 0.36, t(227) = 7.82, p < 0.001).

Table 4 Study 2: Conditional process model for moral emotions as parallel mediators, CBI as moderator, and brand attitude as outcome

Regarding H3a and H3b, CBI moderated the effect of consumer-brand disagreement on brand attitude via other-condemning emotions (index of moderated mediation = 0.03, 95%CI [0.001, 0.07]). For low and medium values of CBI [i.e. M – 1SD = 2.23 and M = 4.25], the indirect effect via other-condemning emotions was significant but became insignificant for high values of CBI [M + 1SD = 6.26] (CBI_low: b =  − 0.16, 95%CI [ − 0.32; − 0.01], CBI_medium: b =  − 0.09, 95%CI [ − 0.20;  − 0.01], CBI_high: b =  − 0.02, 95%CI [ − 0.12; 0.05]). In contrast, CBI did not moderate the effect of consumer-brand agreement via other-praising emotions (index of moderated mediation = − 0.01, 95%CI [ − 0.11, 0.10]). Consequently, the buffering effect of CBI found in study 1 can be replicated for the negative case of disagreement but not for the positive case of agreement.

Study 3

Objective and study design

Study 3 pursued the objective to examine the influence of brand activism on subjects' behavioural intention to advocate for the controversial issue. For this purpose, the same controversial topic and brand as in study 1 were chosen. Subjects were randomly assigned to either the pro-immigrants or the anti-immigrants scenario (between-subjects design, see Appendix 1) already used in study 1.

Sample characteristics

Data from 205 Amazon MTurk members (45.1% female; M_age = 38.86, SD_age = 11.21) were ultimately used for the empirical analysis. To qualify, they had to have a HIT approval of 95% or higher and pass both an attention and a manipulation check. On a scale of political orientation (1 = ‘very liberal’, 7 = ‘very conservative’), 33.2% described themselves as (rather) liberal and 55.6% as (rather) conservative, while the rest identified themselves as politically neutral (M = 4.29; SD = 1.97). Because the groups only narrowly passed the randomization check for age, it was included as a covariate in the model.

Measures

The constructs already introduced in the previous studies were measured with the same scales used in studies 1 and 2. In line with our hypothesis, we compared respondents who agreed to the brand’s position to those who disagreed, excluding people with a neutral opinion. In addition, participants rated their willingness to advocate for the issue addressed in the statement using seven items introduced by Lu and Yuan (2021). One item was removed from the issue advocacy scale due to insufficient factor loading. The constructs exhibited adequate reliability and validity measures (see Table 2).

Results

To test the hypothesised mediation, model 4 from PROCESS (Hayes 2021) was employed with pre-brand attitude and age as covariates. Indirect effects were tested for statistical significance by 95% confidence intervals (based on 5000 bootstrap samples).

For the dependent variable issue advocacy (H2), the results essentially confirm the mediating role of moral emotions (see data in Table 5). Consumer-brand agreement (vs. disagreement) reduced other-condemning emotions (b =  − 1.30, t(177) =  − 4.53, p < 0.001) and increased other-praising emotions (b = 1.08, t(177) = 4.04, p < 0.001). As expected, both mediators enhanced issue advocacy (other-condemning emotions: b = 0.17, t(177) = 3.64, p < 0.001; other-praising emotions: b = 0.25, t(177) = 4.75, p < 0.001). The indirect effect of agreement via other-condemning emotions was negative (b =  − 0.23, 95%CI [ − 0.40; − 0.07]), whereas the effect via other-praising emotions was positive (b = 0.26, 95%CI [0.11; 0.45]). This means that both positive and negative moral emotions motivate individuals to stand up for their opinions in society (see Fig. 4).

Table 5 Study 3: Mediation model for moral emotions as parallel mediators and brand attitude as outcome

Fig. 4

Results from study 3

Furthermore, issue advocacy was directly influenced by consumer-brand agreement (b =  − 0.63, t(177) =  − 3.22, p < 0.01), and the total effect was negative (b =  − 0.59, t(177) =  − 3.05, p < 0.01). This supports H2 stating that issue advocacy is influenced more strongly by disagreement than by agreement.

Discussion

General discussion

The decision to engage in brand activism is one that a company must make in a distinct field of tension. The legitimate aim of being commercially successful and increasing value for shareholders regularly conflicts with the ambition of contributing to the welfare of society as a whole. This is all the truer as brand activism is, by definition, about taking a stance that supports a specific interest group but at the same time works against the goals of its opponents. To what extent this kind of partisan behaviour should be the task of a company and contributes to the consensual solution of a fundamental societal challenge remains largely open. Experience suggests that creating frontlines does not facilitate this endeavour (Bliuc et al. 2021; Iyengar and Westwood 2015; Levendusky 2018; Wojcieszak et al. 2018; Wojcieszak and Warner 2020).

Nevertheless, more and more companies are getting involved in brand activism. For them, the challenge is to optimise its impact for the brand and to achieve the goal of contributing to the solution of societal challenges. To help answer these questions, this research investigated how consumers react to brand activism, focusing on the mediating role of moral emotions and the moderating impact of CBI. Three scenario-based experiments identified the crucial role of other-praising and other-condemning emotions as mediators between consumer-brand (dis)agreement and brand attitude (studies 1 and 2) and issue advocacy (study 3), respectively. Furthermore, CBI was identified as a moderator (studies 1 and 2). Contrary to our hypothesis, consumers with a low (vs. high) level of CBI experienced stronger moral emotions in response to brand activism. No significant effect of consumer-brand (dis)agreement on brand attitude could be observed for consumers with a high level of CBI. Lastly, moral emotions have been shown to not only lead to brand-related outcomes, but also to trigger socially relevant behaviour (study 3). Interestingly, brand activism seems to motivate opponents of the brand’s stance to advocate for their position more strongly than advocates.

Theoretical contributions

This work contributes to research in several ways. First, moral emotions are introduced as important factors mediating between brand activism and its effects. Brand activism elicits both positive and negative moral emotions, depending on whether the consumer's stance is consistent with or contrary to that of the brand. Because negative issue-related emotions do not mediate the effect of brand activism on brand attitude, the observed effects do not result from general emotional excitement triggered by the issue but relate specifically to the two categories of moral emotions. Hence, they contribute significantly to explaining consumers' responses to brand activism.

Second, this study expands knowledge of the importance of pre-existing bonds between consumers and brands. Both study 1 and 2 identify CBI as a shield against negative consumer reactions towards the brand. A single activist statement does not seem to be such a serious transgression as to severely affect brand attitude. Our findings confirm those of Leak et al. (2015), but contrast with other studies that have claimed the opposite effect for ethical, self-relevant, and controllable events (Grégoire and Fisher 2006; Trump 2014). One possible explanation could be that, unlike an orchestrated campaign, a single statement is not considered as a fundamental preference for a particular stance. Moreover, consumers might recognise the subjective nature of the moral values that are at the heart of brand activism. For instance, promoting residence rights for illegal immigrants does not reflect or violate a universally shared norm, but is based on different value priorities (Fernandes 2020).

Third, the study contributes to the emerging field of research on businesses and individuals as political and social actors. Emotions triggered by brand activism motivate people to advocate for their socio-political opinions and values. This represents a desirable form of civic engagement and political participation. Yet, there is also a risk of a negative feedback effect (Korschun et al. 2020). The findings suggest that brand activism disproportionately activates the advocacy behaviour of individuals who want to achieve the opposite than the brand. Therefore, brand activism not only risks alienating certain consumer groups, but also that of going against its very purpose.

Managerial implications

The findings of the three studies provide marketers with important insights into whether and how they should or can effectively advocate on controversial issues. In terms of the ‘how’, the empirical findings suggest that activist messages should be designed to encourage the emergence of positive (and avoidance of negative) moral emotions. Gratitude, for instance, may be elicited if the campaign is judged to be purposeful and genuine, and if the company is credited with taking a risk in doing so (Bridger and Wood 2017). In particular, care must be taken to ensure that the target audience perceives the company’s involvement as sincere and authentic (de Vries and Duque 2018); otherwise, there is a risk that consumers will suspect the brand of woke washing and acting opportunistically (Mirzaei et al. 2022; Vredenburg et al. 2020; Yang and Mundel 2021). For instance, brands should be sensitive when engaging in brand activism, and consistent in their actions and statements (i.e. having ‘moral integration’) (Sibai et al. 2021). Furthermore, the type of commitment (i.e. financial or rhetorical) and message framing need to match the brand type in order to be perceived as authentic (Ahmad et al. 2022). Gratitude also plays a key role, as it serves as a motivator for (further) moral behaviour (Haidt 2003), i.e. it evokes social engagement and reciprocity. If consumers feel grateful towards a brand, this moral emotion in turn elicits the desire to return the favour received by the brand, for example, by increased purchase or advocating intentions for this brand (Algoe and Haidt 2009; de Vries and Duque 2018; Romani et al. 2013a; Xie et al. 2019). Particular caution should be taken in relation to the emotion of anger. Because activist campaigns are likely to arouse negative emotional reactions of those who disagree with the stance, brand managers must actively mitigate this risk. They should counteract the impression that the brand itself is unfair, disrespectful, or participating in a blame game in order to reduce the intensity of other-condemning emotions (Rozin et al. 1999). If this is not successful, outraged consumers are likely to seek revenge and engage in anti-brand activism (Romani et al. 2015). Therefore, a clear own stance should always be linked to a dialogue-oriented approach that reaches out to those with a different opinion. Last but not least, building a reservoir of positive moral emotions respectively of associations that position the brand as a caretaker in the minds of its customers could generate a neutralising effect.

Since a high level of CBI provides a shield in case the brand's stance contradicts the individual's position, the respective status of the target group should be considered when planning a brand activism campaign. While enthusiastic followers of a brand are usually not deterred by brand activism contrary to their own opinion, caution should be taken with low CBI-individuals. However, (continued) brand activism against a consumer segment’s moral values threatens to reduce brand identification over time (Mukherjee and Althuizen 2020), eroding the protective shield. Therefore, it is of importance to continuously strengthen it. This can be achieved, for example, by emphasizing consumer-brand similarities in other domains (e.g. shared quality standards and design preferences), through special relationship benefits (e.g. via loyalty programs or local community support), by highlighting brand prestige, or by creating memorable brand experiences (e.g. special events) (Stokburger-Sauer et al. 2012; Tuškej et al. 2013; Tuškej and Podnar 2018). The relevance of a protective shield provided by a high CBI also leads to the implication that companies operating in a product market where the identification with brands is structurally low (vs. high) should take a much more cautious stance on ‘whether’ to engage in brand activism.

More generally, brand managers must weigh up the strategic relevance and weight they assign to brand activism as part of their overall brand strategy. The respective activism campaigns need to be orchestrated with the content of the regular campaigns and corporate activities. Following Sibai et al. (2021), brands need to have moral integration, i.e. they must be able to follow their moral beliefs in all situations (or ‘practice what you preach’). In contrast, when consumers assume that a brand lacks moral uprightness, they devalue this brand and label it as inauthentic and opportunistic (Moorman 2020; Sibai et al. 2021; Vredenburg et al. 2020).

However, before a brand starts to look at how its brand activism should be designed in concrete terms to achieve its goals, it should first clarify whether it wants to be a brand activist at all. Although brand activism seems to be in vogue, this question requires a conscious decision based on a comprehensive assessment of the opportunities and risks. In other words, brand activism should not be seen as a fad to follow just to be trendy. Rather, this decision calls for a strategic response that balances two objectives: the value it creates for society and the value it creates for its shareholders. As brand activism aims to promote change in society, the brand must realise that it is turning into a political actor. This reinforces its responsibility towards society. Furthermore, the potential consequences of such initiatives need to be carefully assessed in advance. These include not only the risk of losing customers and the profit they generate, but also the fact that an activism campaign encourages in particular opponents of the brand’s stance to stand up for their own behalf. As a consequence, the supporters should be actively involved in the campaign and thus encouraged to promote the company’s stance. Obstacles in this regard must be removed in order to mobilise (otherwise) silent supporters. In short, brand activism should not be undertaken in a perfunctory way, it should either be undertaken consistently or not at all.

Limitations and future research

Limitations of this paper provide interesting avenues for future research. First, the study focuses on the two groups of moral emotions that the individual directs towards the brand as the other element in the relationship. Including further moral emotions such as pride, shame, or guilt (self-conscious emotions) into the model might prove insightful.

Second, the scenarios presented the company’s stance without providing a detailed justification of the position. Additional information or a request for understanding could be an opportunity to influence the target group’s emotional response. Examining effects of more inclusive communication and de-escalation strategies on people’s emotions could therefore be a task for future studies.

Third, only a single, isolated statement served as a stimulus. In reality, however, brand activism seems to be an essential part of the brand-DNA and as such a strategic decision rather than a tactical measure (Sarkar and Kotler 2018). Longitudinal studies lend themselves to capture this. In addition, constructs such as authenticity, credibility, and consistency could provide interesting insights in this context.

Forth, further studies could examine whether activism actually serves its goal of changing society for the better. According to the findings, activism particularly benefits the opposing side’s cause. Moreover, brand activism threatens to further polarise society, one of the top five challenges worldwide (World Economic Forum 2021). Brand activism thus faces a serious dilemma. This should be elucidated in a multi-disciplinary approach by analysing in detail the long-term societal impact of brand activism.

Finally, the interplay between cognitive and affective responses to brand activism deserves further attention. The importance of cognitions and emotions in moral consumer decision-making has already been confirmed. However, the detailed hierarchy of effects and potential alternative pathways have not been assessed in the context of brand activism.

Appendices

Appendix 1 Scenarios used in studies 1–3

Studies 1 and 3

During a press conference last week, in response to a question about immigrant workers, a spokesperson of [Company Name] said:

Case 1 (pro-immigrants): “As a company, we believe that all illegal immigrants cannot be asked to leave the country without considering how long they have already been living here. This is a matter of compassion. There should be some solution. This country has been benefited in many ways by the contributions of such illegal immigrants.”

Case 2 (anti-immigrants): “As a company, we believe that all illegal immigrants have to be asked to leave our country irrespective of how long they have been living here. They violated the law of the land. There should be no amnesty. The rightful citizens of our country have been badly affected for far too long by such illegal immigration.”

Case 3 (control): "As a company, we believe our success is largely caused by our people. Everyone here contributes a part—their time, motivation, joy and work—to our processes and, in turn, to the creation of something great. Our activity makes sense for all of us and also for our customers. Therefore, everyone in this company can find meaning in their daily work."

Study 2

Case 1 (pro-choice, adopted from Mukherjee and Althuizen 2020): According to reports, [Company Name] has been socially active and voiced their opinion regarding contemporary issues. For example, the company recently took a strong pro-choice stand in the reproductive rights debate. The company also financially contributes to the pro-abortion rights movement that actively supports Planned Parenthood.

Case 2 (pro-life, adopted from Mukherjee and Althuizen 2020): According to reports, [Company Name] has been socially active and voiced their opinion regarding contemporary issues. For example, the company recently took a strong pro-life stand in the reproductive rights debate. The company also financially contributes to the anti-abortion rights movement that actively opposes Planned Parenthood.

Case 3 (no stance): According to reports, [Company Name] has been socially active but impartial on contemporary issues. For example, the company refrained from taking a stand in the reproductive rights debate. The company also decided to not financially contribute to the socio-political movements, such as the pro-life or pro-choice movements.

Appendix 2 Diagnosis of possible causes of endogeneity and techniques applied to counter them

Endogeneity cause	Hypothesised effects
	Brand activism → Moral emotions*	Moral emotions → Brand attitude Moral emotions → Issue advocacy	Moderation by CBI
Omitted variable	Participants were randomly assigned to an experimental condition** (Hill et al. 2021; Shadish et al. 2002), sample sizes were appropriate and manipulations were effective (Krause and Howard 2003)	Pre-brand attitude (studies 1–3), issue involvement (study 1) and age (study 3) were considered as control variables (Hill et al. 2021) In all studies, key omitted variables (gender, age and political ideology) are evenly distributed across treatment groups (Hill et al. 2021)
Simultaneity	Participants were randomly assigned to an experimental condition** (Hill et al. 2021; Shadish et al. 2002), sample sizes were appropriate and manipulations were effective (Krause and Howard 2003)	It is considered proven that there is not a large gap (in time) between a stimulus and the emergence of moral emotions (Berkowitz and Harmon-Jones 2004; Lerner et al. 1998) Pre-brand attitude (studies 1–3), issue involvement (study 1) and age (study 3) were considered as control variables (Hill et al. 2021)	Consumer-brand identification was measured before the participants were exposed to a scenario (and thus before they were expected to elicit moral emotions and rate post-brand attitude) Pre-brand attitude was considered as control variable (Hill et al. 2021) and measured before post-brand attitude
Measurement error	Validated measurement models were used (Greco, O'Boyle, Cockburn, and Yuan 2018), whose high validity was confirmed in the studies (see Table 2) (Hill et al. 2021) In all studies, the tested models passed Harman’s single factor test (see Table 2) (Podsakoff, MacKenzie, Lee, and Podsakoff 2003; Podsakoff, MacKenzie and Podsakoff 2012; Siemsen et al. 2010)
	The scenarios were either taken from an established study (study 2) or subjected to a pre-test
Selection into sample	mTurk covers a broad, but not the entire spectrum of the target population (Hill et al. 2021) Participation was limited to panellists with an HIT approval of 95% or higher (Goodman and Paolacci 2017) Participants were randomly assigned to an experimental condition** (Hill et al. 2021; Shadish et al. 2002), sample sizes were appropriate and manipulations were effective (Krause and Howard 2003) The share of those who clicked on the survey link but did not fill in the questionnaire or did not fill it in completely does not differ between the groups in any of the studies
Selection of treatment	Participants were randomly assigned to an experimental condition** (Hill et al. 2021; Shadish et al. 2002), sample sizes were appropriate and manipulations were effective (Krause and Howard 2003)

1. * An analogous rationale applies to the (non-hypothesised) effects brand activism → brand attitude and brand activism → issue advocacy
2. ** Although the attitude of the respective subject to the polarising issue played a role in the assignment to one of the categories of brand activism (agreement vs. disagreement), the grouping is nevertheless random, since the scenario presented was randomly selected, the assignment followed a fixed rule, and the individual was not aware at the time he or she took a position that this was relevant for the allocation to a group