Brand misconduct is an increasingly common phenomenon. On January 30th, 2024, Akio Toyoda, Chairman of Toyota Motor Group, publicly apologized to consumers and investors in a press conference following an investigation of life-threatening car safety irregularities in three of the Group’s subsidiaries (CNBC, 2024). In 2017, United Airlines provoked outrage across social media platforms for violently removing customers from an overbooked flight (Hyken, 2017). The Volkswagen (VW) CO2 emissions scandal resulted in $30 billion in legal costs and a decline of approximately 50% in the firm’s share price (Ewing, 2019). Nike suffered reputational damage that lingered for years after the “sweatshops” crisis in Southeast Asia (Nisen, 2013). Brand transgressions represent a major threat for brands and a top priority for marketers and communication practitioners.
Brand transgressions refer to actions that violate expectations of appropriate brand behavior and harm consumer-brand relationships (Aaker et al., 2004). While prior research has investigated a variety of antecedents, processes, outcomes, and responses related to brand misbehavior (see for review Khamitov et al., 2020), our knowledge on brand transgressions presents four important gaps (see Table S1, Web Appendix). First, managerial concerns and research interests have focused on well-known multinational corporations and brand scandals that made “global news” (e.g., the VW emissions scandal, the BP oil spill). In all these cases, brand origin is ignored or assumed to be of little or no relevance in explaining cross-national differences in consumer reactions and post-transgression brand performance. This is surprising considering evidence pointing to the contrary. For instance, the Nestlé Baby Killer scandal caused a disproportionately higher consumer outrage in the brand’s regions of origin than in foreign markets that suffered more by the misconduct (Sasson, 2016).
Advertisement
Moreover, prior research has shown that brand origin triggers various types of negative spillovers following a brand transgression. Transgressions do not only hurt the image of the brand that committed them but also the image of the brand’s country of origin (country image spillovers). For example, the VW scandal hurt the “Made in Germany” label internationally (Magnusson et al., 2014). Domestic brand transgressions cause perverse halo spillovers that tarnish consumers’ attitudes toward other (non-transgressing) brands from that same country (Borah & Tellis, 2016). Moreover, domestic brand transgressions create media spillovers as they attract significantly higher media coverage and garner disproportionate attention and negative publicity than foreign brand transgressions (Stäbler & Fischer, 2020). Finally, domestic transgressions lead to local economy spillovers, as they undermine the brand’s competitive position and hurt domestic industries (Giannetti & Srinivasan, 2021).
Second, prior research has focused on causal attributions, justice perceptions, morality judgments and consumer-brand relationship harm to explain consumers’ responses to brand transgressions. However, limited attention has been given to the social identity implications of brand misbehavior. This omission is important, considering that consumers perceive domestic brands as fellow in-group members whose actions impact consumers’ national pride and sense of identity (Pandya et al., 2023). The same applies for foreign brands who are often appraised as threats to consumers’ local economies and home cultures (Davvetas et al., 2023).
Third, prior research has predominantly focused on the short-term effects of brand transgressions (cf., Aaker et al., 2004; Grégoire et al., 2009). Knowledge on the determinants of the long-term impact of transgressions on brand performance and post-transgression recovery patterns remains limited. More importantly, marketing research has so far remained silent on whether the impact of transgressions varies cross-nationally and how differently post-transgression recovery trajectories unfold in the brand’s domestic and foreign markets.
Fourth, despite extensive guidelines offered by the crisis management literature (Bundy et al., 2017; Coombs, 2007), the effectiveness of “blanket” crisis response strategies (e.g., apologies, communication framing, compensations) have not been tested across domestic and foreign brand crises. This void leaves practitioners with little guidance regarding how to best manage their brands’ post-transgression recovery across markets and optimally tailor their crisis interventions to minimize the impact of brand misconduct at home and abroad.
Advertisement
Against this backdrop, this research investigates how a brand’s origin affects consumer reactions and brand performance following a transgression and how post-transgression brand recovery should be managed at home and abroad. Drawing on social identity threats theory (Branscombe & Wann, 1994) and literature on group threats (Greenaway & Cruwys, 2019), we theorize the origin-backfire effect. We argue that domestic brand transgressions are seen as punishable acts of in-group betrayal because they threaten consumers’ social identity. We also argue that consumer ethnocentrism attenuates this effect. We then offer two novel crises response strategies (social identity conflict framing and collective compensation) that limit consumer backlash after transgressions in foreign and domestic markets, respectively.
We test our predictions using a multimethod approach that combines (1) emotionality analysis of 229,695 German and American consumer tweets related to two renowned real-world transgressions, (2) secondary-longitudinal data drawn from the YouGov BrandIndex dataset on VW’s reputation and value following the emissions scandal, and (3) experimental data that test the underlying processes and boundary conditions of the effect and the crisis interventions designed to address it in domestic and foreign markets. Across six studies, we find that brand transgressions cause significantly higher and more persistent negative sentiment among domestic market consumers than among foreign ones (Study 1); lead to reputational and brand value damage, which is larger in magnitude, follows a different recovery trajectory, and takes longer to bounce back to pre-transgression levels in domestic than in foreign markets (Study 2); and are forgiven less because they represent stronger threats to consumers’ social identity (Studies 3 A and 3B). We also find that ethnocentric consumers react less negatively to domestic brand transgressions (Study 4 A); and managers can alleviate negative consumer reactions to transgressions through framing their brand’s misconduct as a response to intergroup threats in foreign markets (Study 4 A) and through collective compensation strategies in domestic markets (Study 4B).
Our research contributes to the brand transgression literature by (1) establishing brand origin as a key driver of consumer responses to brand misconduct, (2) introducing a novel psychological mechanism explaining consumer reactions to transgressions in a globalized marketplace, and (3) assessing the long-term impact of transgressions on brand performance across domestic and foreign markets. The findings also contribute to the literature on home-country bias by identifying the conditions under which it may backfire. From a managerial standpoint, our findings offer insights with regard to (1) the adaptation of transgression monitoring, prevention and risk assessment systems across markets, (2) market prioritization, resource allocation, and long-term commitment to transgression recovery strategies, and (3) tailored crisis response strategies following brand misconduct at home and abroad.
Conceptual framework and hypotheses
Domestic versus foreign brands and in-group country bias
Country-of-origin (COO) research highlights the influence of a brand’s origin country on consumer decision-making (Hulland et al., 1996). The conceptualization of brand origin is a contentious issue that has attracted significant research interest in the international marketing literature (Magnusson & Westjohn, 2011). Following the surge in globalized value chains, there is considerable ambiguity regarding whether a brand’s COO is determined by where it is manufactured, designed, or sourced. In response, recent developments suggest that a brand’s COO should be defined as the “country which a consumer associates with a product or brand as being its source, regardless of where the product is actually produced” (Jaffe & Nebenzahl, 2006, p. 29). Samiee (2011) suggests that brand origin reflects the country of the brand’s headquarters and is what consumers typically associate with the brand (e.g., Nike is seen as a U.S. brand even though most of production occurs overseas). Thus, we define domestic (foreign) brands as brands that consumers perceive as originating in their own (a foreign) country, regardless of where they are produced or sourced.
Consumers are more favorably predisposed toward domestic than foreign brands, a phenomenon referred to as “domestic country bias” (Balabanis & Diamantopoulos, 2004). Socially, consumers see domestic brands as members of their in-group and support them to promote national interests. Economically, consumers support domestic products because they protect local economic structures, decrease unemployment, and create positive externalities for local economies (Van Ittersum & Wong, 2010). Culturally, domestic brands represent the national culture, unite local communities, protect local traditions, and elicit national pride through their achievements (Özsomer, 2012). Politically, domestic products benefit from consumers’ growing antiglobal sentiment and anticorporatist attitudes (Davvetas et al., 2023; Thompson & Arsel, 2004). Although in some countries consumers show xenocentric tendencies that favor foreign products (Balabanis & Diamantopoulos, 2016), the literature collectively suggests that domestic brands benefit from “in-group bias” while foreign brands suffer from the “liability of foreignness” (Table S2, Web Appendix).
Brand transgressions and consumer reactions
Transgression is the violation of the implicit or explicit rules that define a relationship (Aaker et al., 2004). Brand transgressions endanger the brand’s ability to provide its promised benefits to consumers (Dutta & Pullig, 2011), damage brand equity by causing revenue and market share losses (Ahluwalia et al., 2000), lead to negative corporate associations (Dawar & Lei, 2009) and affect the brand’s relationship with consumers by eliciting negative emotions, attitudes, and intentions (Dawar & Pillutla, 2000).
An important factor that shapes the relationship between a brand and the consumer after a transgression is consumers’ willingness to forgive the transgressive brand. Forgiveness is defined as “a deliberate attempt to overcome unhappy feelings” that “usually requires the individual’s perception of the wrongdoer to change, from adopting a negative view to adopting a more sympathetic view” (Maltby et al., 2001, p. 882). Forgiveness entails emotional, cognitive, and behavioral involvement. Forgiving requires a person to behave constructively toward someone who behaved destructively, by forgetting, excusing, or justifying the offender (McCullough et al., 1997). Forgiveness depends on the severity of the brand’s misbehavior as well as brand-related factors such as brand attachment (Ahluwalia et al., 2000), familiarity (Dawar & Lei, 2009), expectations (Dawar & Pillutla, 2000), personality (Aaker et al., 2004), and the brand’s response to the crisis (Coombs, 2007).
Forgiveness depends on the type of transgression. Performance transgressions describe failures to deliver the promised functional benefits to consumers (Roehm & Brady, 2007), that range from minor to life-threatening (Klein & Dawar, 2004). Value transgressions describe social or ethical crises in which the brand fails to deliver the promised symbolic and psychological benefits (Dutta & Pullig, 2011). Value transgressions lead to stronger moral outrage among consumers (Kähr et al., 2016; Kübler et al., 2020), because they are associated with negative intentionality and immorality (Ingram et al., 2005). Performance transgressions lead to weaker reactions as they speak more to a brand’s competence (Cleeren et al., 2017) and are more receptive to crisis responses (e.g., product recalls) (Dutta & Pullig, 2011).
The origin-backfire effect
Transgressions can be initiated both by in-group and out-group brands. Literature on domestic country bias would predict that people are more forgiving toward transgressing in-group members to ensure in-group cohesion, survival, and protection from out-groups (Brewer, 1979). According to this perspective, consumers should treat in-group brand transgressors as prodigal sons who deserve forgiveness and reinstatement within the in-group to minimize damage to the group’s power relative to out-groups. Drawing from literature on group conflicts (Greenaway & Cruwys, 2019) and the moral assessment of misconduct (Haidt, 2007), we offer a counter-prediction. We hold that consumers treat in-group transgressors as home traitors who deserve punishment and expulsion from the in-group. Doing so, in-group members use brand transgressions to exemplify unacceptable corporate behavior, signal forgiveness boundaries to fellow in-group brands, maintain group norms and protect the group’s shared value system (Mendoza et al., 2014). Arguing in favor of this perspective, we propose the existence of an “origin-backfire” effect, manifested in harsher consumer reactions toward domestic (compared to foreign) brand transgressions (Fig. 1).
Fig. 1
Conceptual framework and overview of the studies
×
![]()
Victims and observers are more punitive toward transgressive in-group members to protect the in-group from offenders who belong to the same identity group (Abrams et al., 2000; Shinada et al., 2004). Social psychologists have labeled this phenomenon the “black sheep effect” (Marques & Paez, 1994). Transgressions by in-group members threaten the whole group (Tajfel & Turner, 1979) by putting at risk the group’s power and shared value system (Okimoto & Wenzel, 2008). Because violations of shared group values and the derogation of a group’s status lead to the enhanced power of rival groups (Sherif, 1996), forgiving a transgressing in-group member becomes more difficult and less probable. Individuals who feel victimized by an illegitimate offense to their status are more likely to seek revenge to reestablish their position (Vidmar, 2000). Thus, people feel more resentful toward wrongdoing that comes from within the in-group due to the ethical overlap that occurs when they identify with the group (Marques & Paez, 1994). Thus, ostracizing in-group wrongdoers helps dissociate oneself from misconduct (Greenaway & Cruwys, 2019).
This prediction is also supported by moral foundations theory (MFT; Haidt, 2007). MFT explains moral reactions to negative events in both interpersonal (Graham et al., 2018) and marketing domains (Kübler et al., 2020). According to MFT, individuals assess the morality of actions based on five moral foundations. One of them is in-group loyalty/betrayal, which captures whether the action protects/threatens the unity, solidarity, and shared identity of the in-group (Graham et al., 2018).1 Actions that are morally assessed as betraying the in-group elicit stronger moral outrage and motivate stronger punishment (i.e., are forgiven less).
Consumers categorize brands as in-group or out-group members based on national origin. Consumers view a brand as in-group (domestic) to the extent that they associate it with production within home-country borders, operations in the local territory, and/or symbolism associated with their national culture (Sichtmann et al., 2019). Upon such categorization, transgressions by domestic brands are interpreted as acts of in-group betrayal that trigger punitive reactions (i.e., lower forgiveness, more negative sentiment, lower willingness to repurchase). Because foreign brands are not in-group members, the corresponding punitive tendencies will be weaker following foreign brand transgressions.
H1
Consumers react more negatively to transgressions by domestic (vs. foreign) brands.
The mediating role of social identity threat
Social identity threats arise when people are exposed to negative information about their in-group (Dalton & Huang, 2014). Okimoto and Wenzel (2008) distinguish between two types of social identity threats: threats to a group’s power/status and threats to a group’s value system. Domestic brand transgressions are expected to trigger both power/status threats (as they undermine the nation’s image) and shared value threats (as they erode national values).
Individuals are focused on protecting their in-group status and reputation. Faced with social identity threats, individuals disengage from activities and events in which the in-group identity is stigmatized or stereotyped (Steele et al., 2002) and enhance their identification with the in-group to unwelcome or discriminate against other groups (Branscombe & Wann, 1994). Social identity threats also impair self-control and ability to achieve long-term goals (Inzlicht et al., 2006). When individuals are in self-protection mode, their tendency to forgive wrongdoers decreases (McCullough et al., 1997). Similarly, when individuals perceive a transgression as a threat to their self-esteem, they enter a defensive mode, distance themselves from the transgressive act, respond negatively (or even aggressively) to the source of the threat (Baumeister et al., 1996), and become less willing to forgive the wrongdoer (Strelan & Zdaniuk, 2015). Denying forgiveness can offer a sense of control, power (Aquino et al., 2006), status (Shnabel & Nadler, 2008), and identity (Okimoto & Wenzel, 2011).
In the marketing context, when consumers feel a social identity threat, they tend to avoid products associated with that identity (i.e., identity-reinforcing products) as a mechanism of self-protection (White & Argo, 2009). Recent research shows that following internal or external institutional threats to their nation, consumers adjust their preference for domestic products consistent with changes in national pride (Pandya et al., 2023). As transgressions by domestic brands threaten consumers’ social identity, domestic transgressors should suffer from stronger revenge and a lower forgiveness in consumers’ effort to restore the damage caused by the in-group transgression on the national status, honor, and integrity.
H2
Social identity threats mediate the origin-backfire effect.
The moderating role of consumer ethnocentrism
Consumer ethnocentrism is defined as consumers’ perceived inappropriateness or immorality of purchasing foreign-made products for the purpose of defending their national economy (Shimp & Sharma, 1987). Consumer ethnocentrism has a broad motivational foundation as it is associated with both approach goals (i.e., pride/support for the home country) and avoidance goals (i.e., prejudice/discrimination against foreign countries) (Fischer et al., 2022). Advocating an in-group/out-group divide in the moral assessment of marketplace actions, consumer ethnocentrism is highly relevant in the context of domestic brand transgressions. We propose that consumer ethnocentrism attenuates the origin-backfire effect by limiting the social identity threats caused by domestic brand transgressions.
According to the source model of group threat (Greenaway & Cruwys, 2019), when people experience threats from in-group members, they have three options: (a) expel the transgressor from the group, (b) deconstrue the transgression as non-threatening or (c) exit the group altogether. If exiting their national in-group is not possible (i.e., one cannot plausibly renounce their own nationality due to a brand transgression), consumers are left with the option to either ostracize the transgressing brand or reinterpret its transgression as less threatening. While non-ethnocentric consumers should be more likely to ostracize the transgressive brand, ethnocentric consumers should be more likely to construe the domestic brand transgression as less threatening. This differentiation is psychologically facilitated through ethnocentric consumers’ motivated reasoning processes (Kunda, 1990).
Consumers’ ethnocentric tendencies include affective, cognitive, and behavioral elements (Sharma, 2015). Siamagka and Balabanis (2015) propose that consumer ethnocentrism comprises five dimensions, three of which are relevant in the context of transgressions: prosociality, cognitive distortion, and reflexiveness. Prosociality captures ethnocentric consumers’ caring for the welfare of in-group members and an altruistic willingness to assist those members when help is needed. As ethnocentric consumers are morally conditioned to support their in-group members, they should feel a lower social identity threat from domestic brand transgressions and be more inclined to forgive home brands’ misconduct. Ethnocentric cognition is the tendency to distort objective information in ways that reinforce consumers’ pre-existing positive beliefs about in-group members. This dimension explains why ethnocentrists perceive domestic products as functionally superior to imported ones, even when objective quality measures suggest otherwise (Shimp & Sharma, 1987). Such cognitive distortion should make ethnocentric consumers disregard transgressive actions of domestic brands or at least downplay their significance and the corresponding threat. Finally, ethnocentric reflexiveness captures the automatic and unconscious activation of in-group protection mechanisms. Such mechanisms promote the forgiveness of domestic brand transgressions due to low cognitive processing of their negative consequences paired with an automatic retrieval of positive domestic brand associations.
H3
Ethnocentric (vs. non-ethnocentric) consumers react less negatively to domestic brand transgressions because they perceive them as weaker social identity threats.
Managing transgression responses in foreign markets: Social identity conflict framing
After a transgression, brands commonly use post-transgression communications to minimize consumer backlash (Coombs, 2007). Communication research shows that the effectiveness of these responses depends on framing—that is, the effort to “select some aspects of a perceived reality and make them more salient in a communicating text, in such a way as to promote a particular problem definition, causal interpretation, moral evaluation, and/or treatment recommendation” (Entman, 1993, p. 51). Among multiple framing options, communication strategists commonly use conflict framing, which highlights the competition between involved parties, and responsibility framing, which attributes the responsibility for an event to a third party (Valkenburg et al., 1999) or factors outside the brand’s control (Trope et al., 1988) to defensively justify the misconduct. We propose a new type of communication framing, combining conflict and responsibility elements, that marketers can use to alleviate consumers’ social identity threats following transgressions in foreign markets. We call this social identity conflict framing and define it as the brand’s effort to construe the transgression as an outcome of intergroup (vs. intragroup) conflict with the purpose of shifting the responsibility for the transgression to an out-group (vs. in-group) competitive threat.
Social identity threat theory identifies two types of group threats. Intergroup threats occur when the members of a group feel at risk from the actions of the members of another group (e.g., wars). Intragroup threats occur when members of the group feel at risk from the actions of other members of their in-group (e.g., homegrown terrorism) (Greenaway & Cruwys, 2019). Intergroup threats are commonly appraised as derogation of a group’s power while intragroup threats are interpreted as erosion of the group’s values (Okimoto & Wenzel, 2008, 2011). Consequently, intergroup threats motivate in-group identification while intragroup threats limit group identity self-enhancement (Pandya et al., 2023).
Consumers exhibit rising anti-global sentiment (Davvetas et al., 2023). This sentiment triggers reactionary tendencies to support domestic brands especially when they are imperiled by foreign rivals (Thompson & Arsel, 2004; Varman & Belk, 2009). Consumers increasingly construe the competition between domestic and foreign brands as a zero-sum social identity war whereby brands compete as “identity-charged” representatives of the national in-group or its rival out-groups. In such a context, consumers judge brand actions not only on their inherent morality but also on their ability to protect the in-group from out-group threats.
When consumers perceive a transgression by a foreign brand as a response to a threat from a home brand (i.e., intergroup threat), social identity threats should subside. Thus, using social identity conflict framing in crisis communications following a transgression in the brand’s foreign markets should decrease local consumers’ feelings of threat to their nation’s power/status because the transgression represents the inability of a foreign brand to beat domestic rivals without engaging in unethical practices. In parallel, social identity conflict framing also downplays threats to national values because the foreign brand is not a member of the national in-group, and thus its transgression cannot be interpreted as national ethical decay. Moreover, a foreign brand transgression framed in intergroup terms may signal an admission of the moral superiority of domestic brands over foreign ones, which further increases in-group pride. In extreme cases, intergroup threats are even assessed as less punishable because in-group members victimize the threatening out-group by attributing its transgression to unfair or excessive pressure from the in-group (Walter et al., 2016).
H4a
Social identity conflict framing in post-transgression communications alleviates negative consumer responses by decreasing social identity threats in foreign markets. 2
Managing transgression responses in domestic markets: Collective compensation
Firms often respond to transgressions by offering compensation to affected parties (Coombs, 2007). Compensation strategies include monetary rewards, gifts, and product replacements (Coombs & Holladay, 2008). Unlike post-crisis communications that defensively justify and remove responsibility for the misconduct, compensations are more accommodative and aim at repairing the damages to affected individuals (Bundy et al., 2017). Compensation is an effective crisis response as it aids forgiveness (Harrison-Walker, 2019), repairs trust (Dirks et al., 2011), and signals commitment to avoid future misconduct (Bottom et al., 2002).
We argue that not all compensation strategies are effective following domestic brand transgressions. Crisis response literature suggests that the match between the crisis event and the firm’s crisis response is critical in determining the evaluations of the transgressor (Bundy & Pfarrer, 2015). For instance, in the context of CSR crises, Kim and Choi (2018) report that only crisis responses that are congruent with crisis issue limit negative attitudes toward the transgressing firm. As consumer backlash following domestic transgressions is motivated by the need to minimize social identity threats, compensation strategies that only repair damages to affected individuals cannot attenuate negative consumer reactions due to their inherent inability to restore the social harm caused to the whole in-group. Instead, we propose that only collective compensation, defined as the firm’s effort to address collective in-group grievances through financial investment, can recover social identity losses following domestic brand transgressions. Unlike individual compensation, collective compensation signals respect for national values, caring for all in-group members, genuine repentance for the harm caused to the national identity and willingness to reinstate the brand in the in-group by repairing the disturbance of the group’s relational systems (Kahn et al., 2013).
H4b
Collective compensation alleviates negative consumer responses to brand transgressions by decreasing social identity threats in domestic markets.
Study 1: Emotional responses to domestic vs. foreign brand transgressions
Study 1 tests whether consumers’ emotional responses to transgressions (1) are stronger for domestic (vs. foreign) brands and (2) persist over time or fade soon after the transgression. We use unprompted consumer responses to two well-known, real-world transgressions. We web-scraped tweets from U.S. and German consumers reacting to the VW emissions scandal and the United Airlines (UA) passenger incident. We collected Twitter posts on the two brands (UA [U.S. brand] and VW [German brand]) at three time points: three months before the corresponding scandal’s press release (“before”), three months after the release (“short-after”), and three months after the scandal’s one-year anniversary (“long-after”). We avoided the weeks immediately before and after the scandal’s release to avoid emotional inflation. Tweets were collected using the handles @united, @vw, and @volkswagen_de. We used tweets from both U.S. and German users to ensure that the same transgression was seen as domestic by some (e.g., VW for Germans) and foreign by the rest (e.g., VW for Americans). Overall, we collected 229,695 tweets (200,000 for United Airlines and 29,695 for VW). Due to some missing data, we ended up with approximately 200,000 observations. We collected tweets through Pulsar (pulsarplatform.com). Pulsar has been repeatedly used in academic research for similar data collection procedures (e.g., Gerrath et al., 2023).
Variable operationalization
Consumer sentiment
We obtained sentiment classification using GPT-4, an artificial intelligence tool trained to detect, classify, and quantify emotions in vocabulary datasets, even in informal conversations with emoticons or slang. 22.72% of the tweets were categorized as negative, 37.62% as neutral, and 39.66% as positive. We coded negative sentiment as -1, neutral sentiment as 0, and positive sentiment as 1 (M = 0.17, SD = 0.77).
Domestic brand transgression
We created a binary variable indicating if the user’s country was the same as that of the brand. The variable took the value 1 for U.S. users’ tweets related to UA’s transgression and for German users’ tweets related to VW’s transgression. The variable took the value 0 for U.S. users’ tweets related to VW’s transgression and for German users’ tweets related to UA’s transgression.
Time
We created two dummies (short-term: 1 = up to three months post-transgression, 0 = up to three months pre-transgression; long-term: 1 = up to 15 months post-transgression, 0 = up to three months pre-transgression) to capture the impact of time.
Controls
We used number of followers (in thousands) (M = 11.60, SD = 199.43) and number of friends (in hundreds) (M = 20.50, SD = 131.31) as controls (Valsesia et al., 2020).
Analytic procedure and results
We tested the effect of brand origin on sentiment using a difference-in-differences (DID) analysis. In marketing research, DID has been used to assess changes in brand performance (e.g., Chaudhuri et al., 2019), using a set of treatment firms exposed to a change (e.g., brands introducing loyalty programs) and a set of control firms (comparable brands not introducing loyalty programs). We apply DID to assess short- and long-term differences in sentiment toward the transgressive brand for a treatment group (i.e., consumers who share the brand’s nationality) and a control group (i.e., consumers who do not share the brand’s nationality).
DID estimation
We performed an OLS regression with sentiment as the dependent variable and a domestic brand transgression dummy, a time dummy, their interaction, and the two controls as predictors. We estimated the model for both short-term (before vs. short after) and long-term (before vs. long after) effects. In the following model, β3 is the DID coefficient—that is, the time-related effect of the domestic brand transgression on sentiment. Specifically, we estimated the following model, expecting β3 to be significant and negative:
Sentiment = β0 + β1 Domestic Brand Transgression + β2 Time + β3 Domestic Brand Transgression × Time + β4 Number of Followers + β5 Number of Friends + ε.
The results show a significant interaction between brand origin (domestic vs. foreign) and time (before vs. pooled short- and long-after) (β3 = -0.08, p < .001) that persists upon inclusion of controls (Model 1, 2; Table 1). In support of H1, planned contrasts show a significant decrease in sentiment after (vs. before) the scandal for domestic consumers (contrast = -0.01, SE = 0.03, 95% CI = -0.02 to -0.002) but the opposite for foreign ones (contrast = 0.07, SE = 0.01, 95% CI = 0.05 to 0.08). We repeated the analysis focusing only on short-term effects. We find a significant interaction (β3 = -0.07, p < .001) between brand origin and time (before vs. three months after) that persists upon inclusion of controls (Model 3, 4). Planned contrasts show a significant decrease in sentiment for domestic consumers (contrast = -0.04, SE = 0.004, 95% CI = -0.05 to -0.03), but the opposite for foreign ones (contrast = 0.03, SE = 0.01, 95% CI = 0.01 to 0.04). Finally, we tested long-term effects. The results show a significant negative interaction (β3 = -0.08, p < .001) between brand origin and time (before vs. 15 months after) that persists upon inclusion of controls (Model 5, 6). While the planned contrasts show an improvement in sentiment for both domestic (contrast = 0.04, SE = 0.004, 95% CI = 0.04 to 0.05) and foreign consumers (contrast = 0.13, SE = 0.01, 95% CI = 0.11 to 0.15), the improvement was significantly greater in the foreign than in the domestic market (contrast = -0.09, SE = 0.01, 95% CI = -0.11 to -0.07, Fig. 2). As a robustness check, we analyzed sentiment scores using the sentimentr package (Moradi et al., 2023), IBM Watson, and LIWC. The results generally replicate (see Web Appendix).
Table 1
Short-term and long-term effects of domestic brand transgression on consumer sentiment (Study 1)
Overall Effect | Short-Term Effect | Long-Term Effect | ||||
|---|---|---|---|---|---|---|
Model 1 N = 229,695 | Model 2 N = 229,695 | Model 3 N = 179,989 | Model 4 N = 179,989 | Model 5 N = 123,648 | Model 6 N = 123,648 | |
Domestic brand transgression | –0.01 t = − 0.85 (0.01) | –0.01 t = − 0.85 (0.01) | –0.01 t = − 0.85 (0.01) | –0.01 t = − 0.85 (0.01) | –0.01 t = − 0.84 (0.01) | –0.01 t = − 0.84 (0.01) |
Before vs. after (short-term and long-term) | 0.07*** t = 9.05 (0.01) | 0.07*** t = 9.01 (0.01) | ||||
Before vs. short-term after | 0.03** t = 2.96 (0.01) | 0.03** t = 2.96 (0.01) | ||||
Before vs. long-term after | 0.13*** t = 13.85 (0.01) | 0.13*** t = 13.72 (0.01) | ||||
Domestic brand transgression × Time | -0.08*** t = -9.26 (0.01) | -0.08*** t = -9.25 (0.01) | -0.07*** t = -6.89 (0.01) | -0.07*** t = -6.91 (0.01) | -0.08*** t = -8.03 (0.01) | -0.08*** t = -7.92 (0.01) |
Control variables | ||||||
Number of followers (in thousands) | < 0.01 t = 0.92 (< 0.01) | < 0.01 t = 0.96 (< 0.01) | < 0.01+ t = 1.64 (< 0.01) | |||
Number of friends (in hundreds) | < 0.01*** t = 6.97 (< 0.01) | < 0.01*** t = 6.33 (< 0.01) | < 0.01*** t = 8.15 (< 0.01) | |||
Fig. 2
Sentiment and domestic brand transgressions over time
×
![]()
Discussion
Brand transgressions elicit stronger negative sentiment among domestic than among foreign consumers. Domestic consumers’ sentiment toward the brand decreases right after the transgression and improves only in the long term. In contrast, transgressions do not take an equivalent emotional toll on foreign consumers whose sentiment toward the brand is steadily improving over time. The results thus hint at a time-dependent divergence in emotional reactions to domestic and foreign brand transgressions. Due to (1) the inherent biases of OpenAI in sentiment categorization/calculation (Abramski et al., 2023), (2) the noise and correlational nature of Twitter data, and (3) the analysis of tweets only from users who included brand handles in their tweets, the results of Study 1 prevent definitive causal inferences. To address these issues, Study 2 uses data outside Twitter.
Study 2: The origin-backfire effect—impact and recovery patterns
In Study 2, we test the managerial impact of the origin-backfire effect using longitudinal secondary data on VW’s brand reputation and brand value following the CO2 emissions scandal. The VW scandal (1) is one of the most well-known and managerially impactful cases of brand misconduct, (2) has been used as a prototypical transgression case in academic marketing research (Che et al., 2023), (3) affected both VW’s domestic and multiple foreign markets, allowing cross-country comparisons that are central to our research question, and (4) is old enough to enable the estimation of longitudinal effects without being outdated. We compare the reputational and value trajectory of the treatment (i.e., transgressive) brand (VW) before and after the scandal with that of a control brand (i.e., a non-transgressive domestic brand with similar characteristics, performance, and size). We use BMW as the control brand because BMW shares the same origin with VW, yet it was not implicated in the emission scandal in public perception. A pretest survey of 100 German respondents (M = 30.5, SD = 9.8; 40% female) asking participants to indicate the German car brand they deemed most similar to VW supported the choice of BMW (35%) over Mercedes-Benz (20%), Audi (11%), and other less frequently mentioned brands. Perceptual maps constructed using multidimensional scaling of consumer data also indicate BMW as one of the most similar competitors of VW in consumers’ minds (Won et al., 2018). We purchased data from YouGov’s BrandIndex dataset regarding VW and BMW in the U.S. and Germany for a period ranging from one year before (15th September 2014) the scandal (15th September 2015) to three years after (15th September 2018). The data comprise 5,852 daily observations.
Measures
Brand reputation
BrandIndex provides daily ratings of brand reputation by asking respondents to imagine they were looking for a job (or advising a friend looking for a job) and then to report which of the following companies they would be proud/embarrassed to work for. We used net reputation as proxy for brand reputation.
Brand value
BrandIndex asks respondents daily to rate which brand (among a list of options) they think represents a good/poor value for money. BrandIndex provides net values for these two metrics, calculated by subtracting the percentage of negative responses from the percentage of positive responses. We used net value as a proxy for brand value.
Analysis and results
An illustration of model-free evidence shows a visible difference in post-transgression reputation and value between the domestic and the foreign market (see Web Appendix). As a formal statistical test, we regressed brand reputation and brand value on the time period (before vs. after the transgression), the brand (transgressive vs. non-transgressive), the market (domestic vs. foreign), their three-way interaction, the brand×time interaction (using a DID estimate [Pullig et al., 2006]), the brand×market interaction (our main focus), and the market×time interaction. The results reveal a significant three-way interaction on brand reputation (β = − 9.36, SE = 0.58, p < .001). Compared to the control brand, the transgressive brand experienced a significantly steeper decrease in brand reputation after the transgression in the domestic (vs. the foreign) market. The estimates were significant and negative in year 1 (DID = − 9.36, p < .001), year 2 (DID = − 12.38, p < .001), and year 3 (DID = − 13.06, p < .001) after the transgression in the domestic market. The results replicate for years 1, 2, and 3 for brand value. We present complete results in Table 2.3
Table 2
Difference-in-difference (DID) results (Study 2)
Panel A: Effects of brand transgression on brand reputation | ||||||||
|---|---|---|---|---|---|---|---|---|
Pre-transgression (t – 1) | Post-transgression (t + 1) | DID | Post-transgression (t + 2) | DID | Post-transgression (t + 3) | DID | ||
Net Brand Reputation | Net Brand Reputation | Net Brand Reputation | Net Brand Reputation | |||||
Domestic market | Treatment firm | 33.26*** | 5.66*** | –9.36*** (0.527) | 6.44*** | –12.38*** (0.53) | 6.13*** | –13.06*** (0.528) |
Control firm | 35.65*** | 33.26*** | 30.67*** | 26.86*** | ||||
Foreign market | Treatment firm | 23.11*** | 4.85*** | 10.31*** | 12.99*** | |||
Control firm | 33.91*** | 31.52*** | 30.59*** | 29.08*** | ||||
Panel B: Effects of brand transgression on brand value | ||||||||
Pre-transgression (t – 1) | Post-transgression (t + 1) | DID | Post-transgression (t + 2) | DID | Post-transgression (t + 3) | DID | ||
Net Brand Value | Net Brand Value | Net Brand Value | Net Brand Value | |||||
Domestic market | Treatment firm | 15.33*** | –0.06 | –2.71*** (0.41) | 0.51*** | –6.06*** (0.41) | –0.26° | –6.25*** (0.41) |
Control firm | –2.30*** | –3.43*** | –3.28*** | –5.70*** | ||||
Foreign market | Treatment firm | 10.67*** | –1.68*** | 2.78*** | 5.10*** | |||
Control firm | –0.20 | –0.99*** | –0.32* | 0.17 | ||||
In light of a significant three-way interaction, we break down the market×time interaction for the transgressive brand to investigate the changes for VW in the domestic versus the foreign market across periods (before and 1, 2, and 3 years after the transgression). On average, the effect of the transgression on VW’s reputation in the domestic versus the foreign market was by 9.35 points more detrimental in the first year post-transgression, 14.04 points in the second year post-transgression, and 17.02 points in the third year post-transgression, compared with the pre-transgression period. The results support H1, showing that consumers react more negatively to brand transgressions in the domestic market than in the foreign market. Compared with the first year post-transgression, in the second year, VW’s reputation in the domestic market was 4.69 points worse than it was in the foreign market. Specifically, while there was an improvement of 5.47 points in reputation in the foreign market between post-transgression years 1 and 2, the corresponding improvement in the domestic market was only 0.78 points. Moreover, while there was an improvement of 2.67 points in reputation in the foreign market in the period between post-transgression years 1 and 2, there was no improvement in the domestic market. The results suggest that post-transgression recovery is quicker in the foreign than in the domestic market (Fig. 3).
Fig. 3
Difference-in-difference results. A: Brand reputation of the treatment versus control firms in the domestic versus foreign market. B: Net brand value of the treatment versus control firms in the domestic versus foreign market
×
![]()
We also conducted a regression discontinuity analysis (Imbens & Lemieux, 2008) around the data of the event, using the mean squared errors (MSEs) as indicators to choose the optimal transgression date. We test different transgression dates as the event date (from the date of the event [15th September 2015] to 20 days after the transgression in five-day increments). The results suggest that using the + 5 days estimation leads to the smallest MSE for brand reputation. Using the + 10 days estimation leads to the smallest MSE for brand value. The results fully replicate (Fig. 4). For details, see also Web Appendix.
Fig. 4
Changes in brand reputation and net brand value in the domestic and foreign market
×
![]()
Despite visible drops in brand reputation and value in both foreign and domestic markets, the drop is statistically larger in the domestic one. Observed values suggest that the transgression led to drops in brand reputation and value between 2.29% and 9.38% larger in the domestic market than in the foreign market. Moreover, the post-transgression recovery pace in brand reputation and value is flatter in the domestic market but steeper and faster in the foreign one. The pace of post-transgression brand metric recovery is between 38% and 72% faster in the foreign market than in the domestic market for all metrics considered (i.e., brand reputation: 46.1%; value: 38.3%; satisfaction: 51.5%; purchase intention: 72.3%).
Robustness checks
We conducted several robustness checks (see the Web Appendix for details). First, we successfully replicate our effects using additional YouGov BrandIndex metrics (brand satisfaction and purchase intention; Colicev et al., 2018). Second, we conduct a regression discontinuity analysis using stock prices, drawn from the Investors Relations Fact Sheet of the VW Group over the last 10 years in the domestic market (Xetra), the EU market (FTSE and Euro Stoxx), and the U.S. market (S&P 500 and Dow Jones). The post-transgression stock price drops were significantly steeper in the domestic market than in all other markets. Third, we included the number of VW and BMW Google searches in the news category in Germany and the U.S. in the relevant time periods to account for the domestic media coverage of the VW scandal as a potential explanation of the larger drops in domestic brand metrics. The results remain robust. Fourth, we control for prior brand reputation (Mafael et al., 2022) through adding lagged reputation variables. The results persist. Finally, we add a BrandIndex control measuring brand ownership of the transgressive brand to account for the possibility that stronger drops in the domestic market are attributed to higher market shares of VW in the German relative to U.S. market. The results again persist.
Discussion
Study 2 shows that brand transgressions have a more severe impact on brand performance in the domestic than in foreign markets. The negative consequences of brand transgressions in domestic markets are not only larger but also take longer to bounce back to pre-transgression levels compared with foreign markets. The results support the origin-backfire effect and reveal significant differences in a brand’s post-transgression recovery patterns between domestic and foreign markets. While the choice of BMW as a control brand was supported by both primary and secondary data, BMW was also investigated following the VW emissions scandal, potentially implicating the brand in the same transgression in consumers’ minds. In the following studies, we use experimental designs that eliminate any brand comparability concerns, inherent in the context of a natural experiment using real-world brands.
Study 3: Domestic brand transgressors—prodigal sons or home traitors?
Having established the impact of the origin-backfire effect on consumer reactions (Study 1) and brand outcomes (Study 2) using secondary longitudinal data, in Study 3, we follow an experimental approach to test the mediating role of social identity threat. In this study, we test two mechanisms that capture opposing predictions about the effect of brand origin on forgiveness (“prodigal son” versus “home traitor”) by juxtaposing willingness to protect the in-group brand transgressor with social identity threat experienced by its transgression.
Method
We recruited 183 participants of an online consumer panel in the United Kingdom (Prolific Academic; Mage = 42.1 years, SD = 12.4; 49.7% female) in exchange for monetary compensation to incentivize participation. All participants were U.K. residents and nationals. We excluded participants who failed attention checks and had missing responses.
The session started with a presentation of a brand in the cartridge production industry. Participants were told that the brand would be referred to with the fictitious name ABC for privacy reasons. We randomly allocated participants to a domestic or a foreign brand condition. We manipulated brand origin through the location of the firm’s headquarters (London, U.K. vs. Seoul, South Korea), its city of founding (Manchester, U.K. vs. Busan, South Korea) and the founder’s name (Henry Powell vs. Seo-Jun Gwak).4 We told participants that the brand was recently involved in a crisis. We used a value transgression (i.e., sexual harassment in the workplace) in this study (see the Web Appendix for details). Next, participants were asked to describe their thoughts about the crisis in an open-ended question. Then, they completed measures of the dependent variables (forgiveness, willingness to purchase), competing mediators (social identity threat, in-group protection), controls (transgression severity), and manipulation checks. The scales were drawn from previous literature, were psychometrically assessed using confirmatory factor analyses, and exhibited satisfactory validity and reliability (see Web Appendix for details). Finally, participants provided demographic information, were debriefed, and thanked for their participation.
Analysis and results
Manipulation checks
Participants in the domestic brand condition scored higher on a scale of domestic brand origin than those in the foreign brand condition (Mdomestic = 6.21, SD = 1.08 vs. Mforeign = 1.98, SD = 1.69; t = 20.25, p < .001). Participants perceived the crisis scenario as equally credible across conditions (Mdomestic = 5.79, SD = 1.13 vs. Mforeign = 5.82, SD = 1.09; t = − 0.19, p = .852). The transgression was perceived as equally severe across conditions (Mdomestic = 5.80, SD = 1.13 vs. Mforeign = 5.69, SD = 1.22; t = 0.60, p = .548).
Test of competing mediators
We ran a parallel mediation model to test the competing mechanisms of social identity threat and in-group protection, (Model 4, 10,000 resamples [Hayes, 2018]). We specified brand origin (1 = domestic, 0 = foreign) as the independent variable, social identity threat and in-group protection as mediators, brand forgiveness as the dependent variable and transgression severity as the control. Domestic brand origin has a positive effect on social identity threat (β = 1.36, t = 7.05, p < .001) and social identity threat has a negative effect of brand forgiveness (β = − 0.20, t = − 5.24, p < .001), leading to a significant indirect negative effect, in support of H2. While in-group protection positively affects forgiveness (β = 0.62, t = 10.82, p < .001), domestic brand origin has no effect on protection (β = 0.11, t = 0.83, p = .407). Overall, we find a significant indirect effect of domestic brand origin on forgiveness mediated by social identity threat (ab = − 0.27, SE = 0.07, 95% CI = [–0.41, − 0.15]) but not by protection (ab = 0.07, SE = 0.08, 95% CI = [–0.08, 0.23]). The direct effect is not significant (β = 0.17, t = 1.55, p = .123).5
Discussion
Study 3 offers experimental evidence that supports the mediating role of social identity threat as the explanatory mechanism of the origin-backfire effect. Counter to the in-group country bias literature, our findings suggest that consumers treat domestic brand transgressors as home traitors deserving punishment and not as prodigal sons deserving protection because of in-group membership. In a separate study (Study 3B, reported in the Web Appendix), we test and rule out 10 alternative mediators. Social identity threat is the only significant mediator.
Study 4a: Consumer ethnocentrism and social identity conflict framing
Study 4a tests the moderating role of consumer ethnocentrism and the effectiveness of social identity conflict framing in post-transgression communications across markets.
Method
We recruited 408 participants of an online consumer panel in the U.K. (Prolific; Mage = 37.6 years, SD = 13.6; 70.8% female) in exchange for monetary compensation. All participants were U.K. residents and nationals. We excluded participants who failed attention checks and had missing responses. We randomly allocated participants to conditions using a 2 (transgressive brand origin: domestic vs. foreign) × 2 (response framing: intergroup threat vs. intragroup threat) between-subjects design. Participants read that a brand called ABC committed wage discrimination against its female employees. We manipulated ABC’s origin as in the previous study. Next, participants read that the company issued a press release apologizing for the event and justifying its actions as a response to a competitive threat from either domestic or foreign rivals (see the Web Appendix for details). Participants expressed their view on ABC’s response in an open question and completed measures of the dependent variables, social identity threat, consumer ethnocentrism, controls, and manipulation checks. Finally, participants provided demographic information, were debriefed, and thanked.
Analysis and results
Manipulation checks
Participants in the domestic condition perceived ABC as more domestic than those in the foreign condition (Mdomestic = 6.42, SD = 0.86 vs. Mforeign = 1.82, SD = 1.42; t = 39.67, p < .001). Participants judged wage discrimination as a value violation (Mvalue = 6.19, SD = 1.17) and found the scenario credible (Mcredibility = 4.22, SD = 1.94).
Effects on social identity threat
A factorial analysis of variance (ANOVA) on social identity threat with brand origin, transgression framing, consumer ethnocentrism, all 2-way and the 3-way interaction as predictors (gender and transgression importance as controls) reveals a significant main effect of brand origin (F = 24.73, p < .001). Participants in the domestic brand condition reported stronger social identity threats than those in the foreign brand condition (Mdomestic = 4.51, SE = 0.10 vs. Mforeign = 3.50, SE = 0.10). In support of H3, the origin×ethnocentrism interaction was significant (F = 6.17, p = .013). In support of H4, the origin×framing interaction was also significant (F = 4.552, p = .034). Intergroup threat framing decreased social identity threat for foreign brands (Mforeign−intergroup = 3.15, SE = 0.15 vs. Mforeign−intragroup = 3.85, SE = 0.14) but not for domestic ones (Mdomestic−intergroup = 4.60, SE = 0.14 vs. Mdomestic−intragroup = 4.42, SE = 0.14). The 3-way interaction was not significant (F = 1.061, p = .35).
Moderated mediation
We ran a double moderated mediation model (PROCESS Model 9, 10,000 resamples [Hayes, 2018]) with brand origin (0 = foreign, 1 = domestic) as the independent variable, social identity threat as a mediator, brand forgiveness as the dependent variable, transgression framing (0 = intragroup threat, 1 = intergroup threat) and consumer ethnocentrism as moderators (and controls). In support of H1 and H2, domestic brand origin has a positive effect on social identity threat (β = 1.42, t = 3.54, p < .001), which in turn negatively affects forgiveness (β = –0.231, t = − 6.86, p < .001), leading to a significant indirect negative effect. Regarding the moderating role of consumer ethnocentrism, the interaction term was negative and significant (β = –0.272, t = − 2.37, p = .018; moderated mediation index [MMI] = 0.063, 95% CI = [0.006, 0.133]). As consumer ethnocentrism (CET) increases, the negative effect of domestic origin on forgiveness through social identity threat decreases (low CET: β = –0.320, SE = 0.08, 95% CI = [–0.496, –0.173]; mean CET: β = –0.241, SE = 0.06, 95% CI = [–0.359, –0.140]; high CET: β = –0.161, SE = 0.05, 95% CI = [–0.273, –0.069]).6 Regarding the moderating role of framing, the interaction term was significant (β = 0.866, t = 3.07, p = .002; MMI = –0.200, 95% CI = [–0.348, –0.072]). The origin-backfire effect is stronger when the transgression is framed as a response to an intergroup (vs. an intragroup) threat (intergroup: β = –0.359, SE = 0.08, 95% CI = [–0.529, –0.212]; intragroup: β = –0.144, SE = 0.06, 95% CI = [–0.269, –0.040]). Detailed results on willingness to purchase are reported in the Web Appendix.
Discussion
Study 4a reveals that ethnocentric consumers forgive domestic brand transgressions more because they perceive them as weaker social identity threats. This study also shows that using post-transgression communications that construe a transgression as a response to an intergroup threat is an effective framing strategy in foreign markets, but not in domestic markets. Consumers feel socially threatened by domestic brand transgressions even when the home brand’s misconduct is seen as a reaction to threats from out-group rivals. In Study 4b, we test an additional intervention tailored to domestic markets.
Study 4b: Compensation schemes as crisis responses in domestic markets
Study 4b tests (1) whether compensation is an effective crisis-response strategy following brand transgressions in domestic markets and (2) what type of compensation schemes are more effective in alleviating social identity threats in domestic markets.
Method
We recruited 498 participants on Prolific Academic in exchange for monetary compensation. (Mage = 41.2 years, SD = 13.2; 67% female; all U.K. residents and nationals). We excluded participants that failed attention checks or provided incomplete data. We randomly allocated participants to conditions using a 3 (compensation: collective, individual, control) between-subjects design. We informed participants that a U.K. brand, called ABC, was involved in a sexual harassment crisis. Depending on their allocated condition, participants read that the company issued a press release apologizing for the event without offering compensation (control), offering individual compensation (£100k) to the victims of the sexual harassment cases in the company (individual compensation) or investing the same amount (£100k) to found an organization helping victims of sexual harassment in workplaces around the U.K. (collective compensation) (see the Web Appendix for details). Participants rated forgiveness, social identity threat, and checks, provided demographics, were debriefed, and thanked.
Analysis and results
Manipulation checks
Participants in the collective condition scored higher on a 7-point scale measuring whether the company’s response is focused on the ABC employees (1) versus all U.K. employees (7) (F = 148.3, p < .001; Mcollective = 4.64, SD = 1.47 vs. Mindividual = 1.82, SD = 1.09; t = 16.28, p < .001 and vs. Mcontrol = 2.42, SD = 1.47; t = 12.92, p < .001).
Effects on forgiveness
A one-way ANOVA on consumer forgiveness using the crisis response as predictor reveals a significant main effect (F = 6.87, p = .001). Post-hoc tests show that participants in the collective compensation condition reported higher forgiveness than those in the control (Mcollective = 3.34, SD = 1.29 vs. Mcontrol = 2.85, SD = 1.21, p < .001) and the individual compensation condition (Mindividual = 3.01, SD = 1.16, p = .014). The difference between individual compensation and control was not significant (p = .258).
Mediation through social identity threat
We ran a mediation model (Model 4, 10,000 resamples, [Hayes, 2018]) with compensation (1 = collective, 0 = control, − 1 = individual) as the multi-categorical predictor, social identity threat as the mediator, and forgiveness as the dependent variable. In support of H4b, collective compensation (vs. control) decreases social identity threat (β = − 0.31, t = − 2.07, p = .04) which in turn has a negative effect on forgiveness (β = − 0.59, t = − 19.23, p < .001), leading to a positive indirect effect (ab = 0.18, SE = 0.09, 95% CI = [0.01, 0.36]). Individual compensation (vs. control) did not have a significant effect on social identity threat (β = − 0.17, t = − 1.16, p = .247) and no indirect effect on forgiveness (ab = 0.10, SE = 0.09, 95% CI = [–0.07, 0.27]).
Discussion
Study 4b tests a crisis response relevant to managers facing brand transgressions at home. We find that only compensation aiming at relieving nation-wide grievances caused by a transgression earns home consumers’ forgiveness while simple apologies or reparations to individual victims do not. Thus, even highly accommodative crisis response strategies might fail if they do not address the damage to social identity caused by domestic transgressions.
General discussion
This research reveals that consumers react more negatively to misconduct by domestic brands because they perceive home brand transgressors as in-group traitors that threaten the national identity. The damage of transgressions on brand equity is more severe and takes longer to restore in domestic than in foreign markets but weakens for consumers with high ethnocentric tendencies. Managers can mitigate the impact of transgressions through tailoring the framing of crisis communications and compensation strategies in foreign and domestic markets.
Theoretical contributions
Our research contributes to brand transgressions literature in three ways. First, we identify brand origin as a key determinant of consumer forgiveness and post-transgression brand performance. Previous studies have identified various factors that drive forgiveness, including brand personality (Aaker et al., 2004), consumer-brand relationship strength (Grégoire et al., 2009), brand reputation (Sohn & Lariscy, 2015) and crisis response strategies (Coombs, 2007). The origin-backfire effect extends this stream of literature by establishing the relevance of brand origin in understanding consumer forgiveness in both performance-based and value-based transgressions. Our findings reveal that consumers forgive brand transgressions by domestic brands less easily than transgressions by foreign ones. Thus, our findings suggest that the impact of brand transgressions and how consumers experience them are not uniform across markets; rather, they exhibit significant cross-national variation.
Second, prior research on transgressions has mostly drawn from theories of interpersonal relationships, attribution, justice, and moral foundations to explain consumer responses to brand misconduct (Khamitov et al., 2020). Our research contributes to this body of knowledge by unveiling social identity threat as a novel psychological mechanism underlying consumers’ reactions to transgressions committed by their home brands. Building on work around social group threats (Abrams et al., 2000; Greenaway & Cruwys, 2019), we propose that domestic brand transgressions are interpreted by consumers as events that threaten not only their personal but also their national identities, by putting at risk their nation’s status and value system. Such transgressions generate harsher consumer reactions that may take the form of brand ostracization and require nation-wide relational repair.
Third, we extend knowledge on the long-term effects of brand transgressions on consumers’ emotional responses and key brand performance metrics. Although there is consensus in the literature that brand misbehavior triggers negative consumer reactions (Khamitov et al., 2020), prior research has largely focused on the immediate effects of brand transgressions, with limited attention given to longitudinal effects (cf., Aaker et al., 2004). By investigating long-term effects of brand misconduct on consumers, we reveal that the trajectory of a brand’s post-transgression recovery depends on its national origin. Although we observe fading hostility over time toward brands that transgress in foreign markets, domestic transgressions appear to persist with consumers, who exhibit more negative sentiment toward their home-country brands even over longer periods of time. Thus, our findings extend our understanding of the time window during which transgressing brands risk losing goodwill and reputational capital. Consistent with previous literature, we find that transgressions damage brand value and reputation across all markets. However, the current investigation adds to this knowledge by showing that the harmful impact of brand misbehavior is more severe in domestic markets and that restoring brand health to pre-transgression levels is more difficult for domestic than for foreign brands.
We also contribute to situational crisis communication theory and crisis management literature (Bundy et al., 2017). While prior relevant work offers a wide range of defensive and accommodative crisis response strategies (Bundy & Pfarrer, 2015; Coombs, 2007), it has not tested their effectiveness across brand origins or national contexts. In response, we offer two specialized crisis responses that address these shortcomings. First, we show that using social identity conflict framing in post-transgression communications to construe the transgression as a response to intergroup domestic threats facilitates forgiveness for brands that transgress outside national borders. Such framing makes foreign brands appear less threatening, victimized, and worthy of forgiveness. Second, we show that only collective compensation strategies that go beyond reimbursing individual victims of misconduct and address the collective harm caused to the brand’s home nation can effectively promote forgiveness. These findings reinforce the view that crisis responses must be carefully tailored to the crisis event (Kim & Choi, 2018) and the threats it causes at both personal and social identity levels.
Finally, we contribute to international branding research by showing that a brand’s domestic origin, commonly accepted as a positive brand attribute, constitutes a threat in cases of transgressions. Prior research on in-group biases (e.g., Balabanis & Diamantopoulos, 2004) highlights that consumers feel the need to defend domestic brands against foreign rivals, implying that domestic brands may be immune from post-transgression punishment. Our research offers new counterintuitive evidence demonstrating that consumers are actually more punitive toward domestic brands that misbehave. These findings imply that there are limits to the beneficial effects of domestic country bias. When these limits are crossed, consumers’ social identities are threatened, and the bias backfires. Although consumer ethnocentrism attenuates the backfiring effect of brand origin, it does not eliminate it. Thus, brand transgressions take a toll even on domestic brands’ most loyal defenders. This finding also contributes to a revived discourse viewing local and foreign brands as adversaries in a zero-sum social identity clash in an era of rising antiglobal sentiment (Davvetas et al., 2023).
Managerial implications
Although not all transgressions are equally dangerous, they all represent negative events that require damage control. However, managers should expect transgressions to have different impacts and lead to distinct recovery trajectories across domestic and foreign markets. As Table 3 shows, the negative impact of transgressions on consumer metrics and brand performance is between 25% and 334% larger in domestic than in foreign markets. Also, post-transgression recovery is approximately half as strong in the domestic market as it is in foreign markets in the meso-term (3 years post-transgression), implying that restoring brand performance to pre-transgression levels takes significantly more time at home than abroad. Managers should thus abandon blanket post-transgression recovery strategies across markets.
Table 3
Transgression impact and recovery across domestic and foreign markets
Domestic market | Foreign market | Differences | |
|---|---|---|---|
Transgression impact | Short-term (vs. pre-transgression) ♣ Brand reputation (–27.67) ♣ Brand value (–15.44) ♣ Brand satisfaction (–16.98) ♣ Purchase intention (–3.65) | Short-term (vs. pre-transgression) ♣ Brand reputation (–18.29) ♣ Brand value (–12.38) ♣ Brand satisfaction (–10.81) ♣ Purchase intention (–1.37) | Short-term (vs. pre-transgression) ♣ Brand reputation (51% worse domestically) ♣ Brand value (25% worse domestically) ♣ Brand satisfaction (57% worse domestically) ♣ Purchase intention (168% worse domestically) |
Long-term (vs. pre-transgression) ♣ Brand reputation (–27.21) ♣ Brand value (–15.64) ♣ Brand satisfaction (–18.45) ♣ Purchase intention (–5.30) | Long-term (vs. pre-transgression) ♣ Brand reputation (–10.61) ♣ Brand value (–7.92) ♣ Brand satisfaction (–5.38) ♣ Purchase intention (–1.22) | Long-term (vs. pre-transgression) ♣ Brand reputation (168% worse domestically) ♣ Brand value (179% worse domestically) ♣ Brand satisfaction (243% worse domestically) ♣ Purchase intention (334% worse domestically) | |
Post-transgression recovery pattern | Recovery magnitude (average between years) ♣ Brand reputation (0.47) ♣ Brand value (–0.20) ♣ Brand satisfaction (–1.46) ♣ Purchase intention (–1.65) | Recovery magnitude (average between years) ♣ Brand reputation (8.14) ♣ Brand value (6.78) ♣ Brand satisfaction (5.43) ♣ Purchase intention (0.14) | Recovery magnitude (average between years) ♣ Brand reputation (94% slower in domestic) ♣ Brand value (103% slower in domestic) ♣ Brand satisfaction (127% slower in domestic) ♣ Purchase intention (1243% slower in domestic) |
Company interventions | Post-transgression communications: ♣ Mere apology is unlikely to mitigate social identity threats and facilitate forgiveness. ♣ Social identity conflict framing not effective: avoid attributing the transgression to pressures from competitors (both foreign and domestic). Compensation strategies: ♣ Offer collective compensation: compensate in ways that restore damage in a collective level (e.g., nationwide initiatives) rather than in an individual level (e.g., reimbursing individual victims) | Post-transgression communications: ♣ Apologize for the transgression ♣ Use social identity conflict framing: attributing transgression to domestic competitor pressures attenuates social identity threats and facilitates forgiveness. | ♣ Prioritize domestic (vs. foreign) markets in transgression risk assessment systems. ♣ Commit more resources in post-transgression brand recovery in domestic (vs. foreign) markets. ♣ Long-term post-transgression brand health monitoring is more important in domestic (vs. foreign) markets. ♣ Brands targeting ethnocentric consumer segments face less post-transgression danger in domestic (vs. foreign) markets than brands targeting non-ethnocentric ones. |
Post-transgression recovery strategies should adapt across domestic and foreign markets on several fronts. First, companies should develop monitoring and transgression risk assessment systems to track their brand’s behavior and estimate the risk of misconduct. As brands are exposed to different risks depending on their origin and the country where their transgression took place, such systems should be nationally adapted. Because the reputational and brand value risks are higher at home than abroad, the domestic market should be prioritized in these efforts. This does not imply that managers can neglect, underestimate, or disengage from transgression repair in foreign markets. This would signal unethical market discrimination and allow the reputational backlash from their misconduct abroad to reach the domestic shore. Yet, the home market should take precedence in brand monitoring and risk assessment, especially for brands that rely heavily on domestic revenues to finance their foreign operations. Second, the reputational impact of a domestic transgression erodes pre-transgression home-court advantages. While post-transgression reputation and value are slowly, yet steadily, restored over time in foreign markets, transgressions lead to a long-term reputational plateau in domestic markets. This trajectory manifests in a slower pace of recovery to pre-transgression levels. Thus, from a resource allocation perspective, transgression management efforts should be directed not only at restoring a brand’s global reputation but also at committing resources to domestic reputation repair for a longer period.
Post-transgression communication and repair strategies should also account for brand origin and differentiate between domestic and foreign markets. While firms must apologize for transgressions across all affected markets, brands may benefit by employing social identity conflict framing in post-transgression communications that target foreign market consumers. As brand transgressions threaten consumers’ national identities, spinning transgressions as responses to domestic competition alleviates social identity threats and facilitates forgiveness. While social identity conflict framing helps in foreign markets, it is less effective in the home market. Thus, defensive appeals to in-group biases must be abandoned in favor of more accommodative interventions (e.g., financial compensation) in domestic markets. However, even highly accommodative repair strategies might be ineffective in domestic markets if they do not tackle the social identity implications of the misconduct. Thus, managers should develop compensation schemes that do not only reimburse individual transgression victims but instead focus on restoring the wellbeing of the whole in-group. Such efforts should be particularly considered following scandals that involve violations of ethical values that domestic consumers hold dear.
Fourth, from a segmentation and targeting perspective, our findings suggest that ethnocentric consumers are more lenient toward domestic brand transgressions. This finding has three implications. First, brands targeting chronically ethnocentric segments (e.g., elders, conservatives) face lower risks following brand misconduct in the home market. Second, domestic brand transgressions should be forgiven more easily in markets that go through phases of periodically heightened ethnocentrism (e.g., countries undergoing crises [Fritsche et al., 2011]). Third, post-transgression communications that situationally prime ethnocentric attitudes (e.g., ethnocentric spokespersons in apology advertising, ethnocentric tone in crisis statements) may alleviate consumer backlash in the aftermath of a domestic brand scandal.
Limitations and future research
Our findings should be interpreted considering certain limitations inherent in our research design choices. First, although we combine experimental, web-scraped, and secondary data using a multimethod approach, we do not test the effect of brand transgressions on purchase volumes or churn rates. Future studies could consider outcomes such as actual sales and consumer loyalty. Second, we capture social identity threats through psychometric scales. Future research should use subtler measures (e.g., Implicit Association Tests) to capture the concept’s subconscious nature. Third, although we test the origin-backfire effect in two real-world scandals and multiple manipulated transgressions, replication in different transgression contexts and alternative control brands is needed to establish its generalizability.
Our experimental manipulation of brand origin relied on the headquarters’ location and the founder’s origin. While this aligns with the established conceptualization of the construct (Samiee, 2011), we did not consider other elements that categorize a brand as domestic among consumers. Following consumers’ increasing appreciation for localness, brands position themselves around consumers’ local culture by highlighting their national heritage, local sourcing, and domestic manufacturing (Steenkamp, 2019). Although we expect all domestic brands to face higher risks from transgressions compared to foreign brands, the strength of a brand’s association with its national origin should be a moderator of the origin-backfire effect. While strongly tying a brand to its national culture should help it reap the benefits of localness, it may simultaneously raise the ethical standards by which home consumers evaluate the brand’s actions. Future research should thus investigate the dark side or the unintended consequences of positioning brands around their national identity.
Our investigation used data from the U.S.A., the U.K., and Germany. Although these countries are key battlegrounds for most global brands, transgressions may well happen in countries with different cultural profiles. In a pilot study (not reported herein), we obtained preliminary evidence that consumers punish more domestic brands that violate values that rank high on their home culture’s moral hierarchy. As culture determines which types of misconduct are deemed more morally condemnable by consumers (Graham et al., 2018), future research should investigate the cross-cultural boundaries of the origin-backfire effect by exploring the interplay between morality and culture in the context of transgressions. Relatedly, research on the interaction between a brand’s origin and its transgression’s location would also shed further light on the origin-backfire effect’s variation.
Our research reveals that social identity threat is the underlying mechanism of the origin-backfire effect. Therefore, traits relating to consumers’ social identity hierarchy might moderate the effect. Consumers have multiple location-based identities (e.g., citizens with European, national, and regional identities operating simultaneously [Micevski et al., 2019]). Thus, they may respond differently to domestic brand threats by shifting their focus from the threatened national identity (e.g., German) to superordinate (e.g., European) or subordinate (e.g., Bavarian) identities less affected by the transgression. Given the increasing number of consumers with complex identity structures, the impact of social identity hierarchy on reactions to transgressions is an intriguing opportunity for further research. Relatedly, consumers identify with brands for reasons beyond their in-group membership. In such cases, the buffering effect of consumer-brand identification following a transgression (Cheng et al., 2012) may operate in conflict with their motivation to eliminate social identity threats through brand punishment. Identifying the factors that determine which of the two opposing effects will dominate over the other is another interesting future research direction.
Finally, we proposed social identity conflict framing and collective compensation as interventions that alleviate post-transgression punishment across markets. However, the effectiveness of these interventions should largely depend on the interplay between the brand’s pre-crisis reputation and its post-crisis coverage in the media. As brand reputation creates expectations for more accommodative post-crisis responses (Sayin et al., 2023), researchers should investigate the double standards that consumers may have for domestic and foreign brands and the dynamics between a brand’s repair strategies and their coverage in the media. Disentangling the relative influences of these forces, understanding what determines which crisis narrative will be adopted by the media, and investigating the optimal timing of crisis responses (e.g., Chen et al., 2009) warrant further research.
Acknowledgements
The authors thank the JAMS Editor, the Associate Editor and the four anonymous reviewers for their constructive feedback throughout the review process.
Declarations
Conflict of interest
The authors have no conflict of interest.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.