Shifting Standards in Consumer Evaluations of Global and Local Brands After Product-Harm Crises

The shifting standards theory of social psychology suggests that when people judge individual members of a group, they compare them to within-category judgment standards (Biernat and Fuegen 2001). We tend to evaluate other people's attributes using different standards that reflect our expectations about specific individuals (Higgins and Stangor 1988). Such shifting standards are also used to evaluate behavior and ultimately shape our responses to these behaviors (Biernat and Manis 1994). The perceived desirability and social acceptability of an action might change depending on the actor. A two-year-old talking very loudly in a restaurant may be adorable, whereas an adult is frowned upon for the same behavior. We can easily forgive someone who does not reply to an invitation if they are merely an acquaintance, but a nonresponsive significant other might face social punishment.

Recent research suggests that perceptions of brand globalness (e.g., worldwide availability) and localness (e.g., local embeddedness) trigger categorization of products under the superordinate mental categories of global/local brands, which carry distinct stereotypical associations (Davvetas and Halkias 2019). Hence, we conceptualize that when consumers evaluate a global or a local brand, they compare it to within-category judgment standards. Adapting the shifting standards theory to the international marketing context, we propose that the mere designation of a brand as global or local constitutes an important categorization that might trigger differential associations and within-category judgment standards. We posit that once a brand is categorized as global or local, associations pertaining to the brand category are transferred to individual brands and serve as reference points or standards, with which consumers evaluate those brands. For example, global brands are associated with (higher) quality, credibility, prestige, and status (e.g., Kim, Moon, and Iacobucci 2019; Mandler, Bartsch, and Han 2021; Özsomer and Altaras 2008), leading consumers to expect higher levels of these attributes, thus shifting their standards, when evaluating a global (vs. local) brand.

Applying different judgment standards for global and local brands can have important consequences in cases of product failures and product-harm crises. If consumers have higher performance expectations from global brands, a product failure may lead to (greater) disconfirmation of expectations (e.g., Oliver 1980), affecting evaluations adversely. Extant literature suggests company response as a critical determinant of the effect of the failure or harm crisis on consumer brand evaluations (Aaker 1991; Hewett and Lemon 2019). Dawar and Pillutla (2000) demonstrate that objectively identical company responses may lead to different attitudes toward a brand, depending on consumers’ prior expectations. If global and local brands are compared to different within-category judgment standards, these differential prior expectations may influence how company responses to product failures and product-harm crises are evaluated. This research focuses on whether and how the perception of a brand as global or local influences consumer evaluations following a product failure or product-harm crisis. We aim to delineate when and why consumers might respond differentially toward similar behaviors of global versus local brands in such situations.

We investigate these relationships in the context of emerging markets. The distinction between consumer perceptions of and associations around global and local brands are stronger in emerging markets, whereas they are rather blurred in mature markets (Batra et al. 2000; Davvetas et al. 2022; Strizhakova and Coulter 2015). Most mature markets have their own global brands that are perceived as local brands in their domestic markets. For example, for a German consumer a BMW or Mercedes is a local brand, which happens to be global as well. However, emerging markets have fewer or no home-grown global brands (Davvetas et al. 2022), although they have been a battleground for major global brands. Furthermore, global and local players are more clearly differentiated in emerging markets (Batra et al. 2000; Strizhakova and Coulter 2015). Due to the historical scarcity of well-known global brands, consumers in emerging markets have more positive and aspirational associations with global brands (Davvetas et al. 2022; Holt, Quelch, and Taylor 2004). Mandler, Bartsch, and Han (2021) show that brand globalness boosts brand credibility in emerging markets that are still undergoing globalization, whereas localness serves as a credibility enhancer for consumers in developed markets that have already undergone globalization and have witnessed recent antiglobalization sentiments. Thus, the within-category judgment standards to assess global and local brands are more distinct in emerging (vs. mature) markets (Davvetas et al. 2022). Therefore, emerging markets offer a suitable context to investigate how shifting standards influence consumer evaluations of global and local brands, including situations involving product-harm crisis.

We demonstrate, through four main studies, that consumers from emerging markets react more negatively toward a global (vs. local) brand after a product-harm crisis. We reveal higher initial expectations for global brands as the underlying cause of this more pronounced consumer response, and empirically establish which expectations underlie the shifting standards used in consumer evaluations. We also show, via two supplementary studies, that these effects disappear when consumer expectations from global and local brands are externally equalized but persist even for high-priced and highly prestigious global versus local brands. That is, when within-category judgment standards are equalized, expectations from global and local brands are also aligned, eliminating differences in postcrisis evaluations. However, the within-category standards around globalness and localness still exert differential effects on consumer expectations and evaluations for equally strong brands. Additionally, we identify product category as a boundary condition for the proposed effects, which are attenuated for categories closely associated with localness such as food (Davvetas and Diamantopoulos 2016; Gineikiene, Schlegelmilch, and Ruzeviciute 2016; Özsomer 2012). Finally, we demonstrate that emerging market consumers with high ethnocentrism appreciate it directionally more when a local brand provides compensation after a crisis, compared with a global brand.

This research contributes to international branding and product-harm literatures in multiple ways. First, while global versus local branding has generated extensive research interest, its consideration within the product-harm context has been limited. Barbarossa and Mandler (2021) have recently shown a company's country of origin to be influential on consumers’ emotional responses to crises with a consideration of country warmth. Our work is different from theirs, as we focus on global versus local brand categorizations (Davvetas and Halkias 2019), which carry distinct within-category associations and expectations. Second, our work adapts the theory of shifting standards from social psychology to the intersection of branding and product-harm contexts in delineating an underlying process for the observed effects on consumer responses through expectations. Previous work has built on schema theory to demonstrate differential consumer expectations for global versus local brands (e.g., Davvetas and Diamantopoulos 2016; Xie, Batra, and Peng 2015). We replicate their findings and provide a nuanced theoretical explanation by applying the shifting standards theory around globalness/localness associations and by studying their downstream consequences on consumer reactions to harm crises and company response.

Our work also adds to the relevant streams of literature contextually. Prior empirical research on product-harm crises has focused predominantly on mature markets (e.g., the United States and Western Europe; Cleeren, Dekimpe, and Van Heerde 2017). Our studies are conducted in multiple emerging markets (i.e., Mexico, Poland, Turkey), where distinctions between brand globalness and localness perceptions are clearer (e.g., Davvetas et al. 2022). As a boundary condition, we provide a comparison to a mature market (i.e., Germany). Our studies reveal consistent findings across multiple product categories (i.e., pharmaceuticals, electronics, food) with an additional consideration of the recently relevant COVID-19 context in our scenarios.

Our findings have important implications for brand managers and offer valuable insights into the consequences of positioning strategies focused on globalness versus localness associations. Global brand managers often strive to portray their brands as global entities (Alden, Steenkamp, and Batra 1999). Our finding that globalness leads to greater expectations, causing a higher negative consumer reaction following a product-harm crisis and marginally decreasing the relative effectiveness of firm compensatory efforts, challenges the conventional wisdom that strong prior associations for global brands will work as a shield, protecting them from negative consumer reactions (Davvetas, Diamantopoulos, and Liu 2020; Dawar and Pillutla 2000). Moreover, our work sheds light on the appropriate responses that companies should employ when a global or local brand faces a harm crisis. A brand may alleviate the negative consequences of a product-harm crisis by understanding consumers’ expectations and running a crisis management strategy aligned with these expectations. We discuss the specific implications of our insights for branding and crisis management strategy in today's globalized business environment.

Conceptual Development

We first outline the main tenets of the shifting standards theory, which explains how stereotypes and group perceptions affect judgments about individual group members. We then apply this theory to global versus local branding. We explore the effect of shifting standards for global and local brands within the context of product-harm crises, taking into consideration the impact of company response strategies and the moderating influences of product, market, and consumer characteristics. Figure 1 presents the main aspects of our conceptual development.

OPEN IN VIEWER

Figure 1.

Conceptual Model.

Shifting Standards of Evaluation

The implicit use of different judgment standards when evaluating people, products, or experiences is a general phenomenon. Standards are norms that define the requirements for the expected level of performance considered necessary to be judged eligible as a group member (Higgins and Stangor 1988). The shifting standards theory of social psychology suggests that group membership may activate stereotypes, which subsequently function as category-specific standards that ascertain the expected attributes and behaviors of individual group members (Biernat and Fuegen 2001). These expectations are used to evaluate the group members (Holder and Kessels 2017; Lynch and Finkelstein 2015). If one group's standard is at a lower level than another's, then members of the group with lower standards can more easily surpass the members of the group with higher standards, as the threshold for qualifying for a trait is lower (Biernat and Manis 1994). For example, people judge the height of a woman relative to the standard for women, which is shorter than the standard for men. Hence, the meaning of “tall” is not the same for women and men (Biernat and Manis 1994). Similarly, individuals use different thresholds in determining whether a behavior is diagnostic of stereotypical traits for categories to which a target belongs: the expected range of aggressiveness in men begins and ends at a higher level, compared with the range for women (Bauer 2020). In other words, people routinely shift their judgment standards for members of different social groups.

Various consumer behavior contexts also facilitate the potential activation of such shifting standards. For example, the Michelin star system is a well-known evaluation scheme in restaurant ratings. A single star denotes “high-quality cooking, worth a stop,” two stars mean “excellent cooking, worth a detour,” and three stars represent “exceptional cuisine, worth a special journey” (Michelin Guide 2024). Consumers exposed to this information about a restaurant will make inferences and evaluate its service and products based on their knowledge of the star system. The same food might be evaluated more positively in a no-star or one-star restaurant than in a two-star restaurant. Similarly, we believe that the established associations for global and local brands will set shifting standards in consumers’ minds and affect their expectations and evaluations differentially.

Shifting Standards for Global and Local Brands

Consumers’ perceptions of brand globalness and localness are key drivers of their evaluations and responses, as established by the long-standing international marketing literature. While globalness and localness perceptions are not fully mutually exclusive, Halkias, Davvetas, and Diamantopoulos (2016) argue that individuals tend to perceive either globalness or localness of a brand as more dominant than the other and categorize the brand accordingly. These superordinate mental categories of global versus local brands carry distinct associations (Davvetas and Halkias 2019), especially in emerging markets (Davvetas et al. 2022; Özsomer 2019), shaped through historical consumer experience with brands in the marketplace and firm communications (e.g., Batra et al. 2000; Steenkamp, Batra, and Alden 2003).

While both global and local brands provide various functional and symbolic values to their customers, their stereotypical associations that drive these values are different (e.g., Özsomer 2012; Steenkamp, Batra, and Alden 2003; Swoboda, Pennemann, and Taube 2012). Global brands, which are characterized by worldwide reach and availability, enjoy the advantages of a “global brand halo effect” (Davvetas, Diamantopoulos, and Liu 2020). They are perceived to possess higher levels of credibility and quality (Erdem and Swait 1998; Holt, Quelch, and Taylor 2004; Mandler, Bartsch, and Han 2021) and create a sense of achievement and prestige for consumers (Özsomer and Altaras 2008). In contrast, local brands are characterized by local embeddedness, and are perceived to reflect local cultural identity (Özsomer 2012) and alignment with local lifestyles, values, and preferences (De Vries and Fennis 2019; Kapferer 2002; Xie, Batra, and Peng 2015). Local brands also provide affordability and service availability advantages (Batra et al. 2000). However, particularly in emerging markets, they have not been associated with high product quality and performance. Steenkamp, Batra, and Alden (2003) demonstrate, in two countries (Korea and the United States) and across different product categories (cola drinks, facial cream, wristwatches, toothpaste, refrigerators, and TV sets), that perceived globalness of a brand is positively associated with its prestige and perceived quality. They further show that a brand's local icon value is associated with prestige but not with quality and performance.

Global brands typically allocate a higher budget to research and development due to their extensive geographical reach (Dimofte, Johansson, and Ronkainen 2008; Özsomer 2012). This enables them to continuously innovate and integrate newest technologies into their production processes. Thus, global brands are generally associated with newer and better technology and higher product performance (Özsomer 2012; Samiee 2019). In contrast, local brands are often perceived as lacking modernity and aspiration (Dimofte, Johansson, and Ronkainen 2008) and are seen as lagging behind in terms of technological edge and innovation (Balabanis and Diamantopoulos 2016). Therefore, while consumer expectations may vary across multiple dimensions for global and local brands in emerging markets (e.g., quality, trust; Batra et al. 2000; Davvetas et al. 2022), associations with technological leadership and high-standard products are likely to be the primary factors driving differential consumer evaluations.

H₁: Consumers from emerging markets have higher expectations of product performance for a global brand than for a local brand.

Consumer expectations of product performance are especially critical within the context of product failures and product-harm crises in relation to their effect on postincident consumer evaluations. An individual-level product failure is when perceived performance is subpar after a specific purchase transaction or when the product fails in time (e.g., breaks down, wears out). If a product failure occurs on a broader scale because of an inherent defect or problem that might even be dangerous for the customers and/or have negative externalities for the public or the environment, it is called a product-harm crisis (Cleeren, Dekimpe, and Van Heerde 2017). As expected, product failures and harm crises negatively affect consumers’ perceptions about a brand, tarnish its reputation, cause market share and revenue loss, and destroy brand equity (e.g., Siomkos and Kurzbard 1994; Van Heerde, Helsen, and Dekimpe 2007).

Applying shifting standards theory to global and local brands in a product-harm crisis context, we predict that consumers’ higher initial expectations for a global (vs. local) brand, due to differential within-category judgment standards in emerging markets, will create a higher performance threshold for the global brand. Accordingly, per the expectation disconfirmation theory, which focuses on a comparison of perceived performance and expectations (Oliver 1980), once the thresholds are set at higher levels, a product-harm crisis will be perceived as a larger disparity in performance for global brands in comparison to local brands. As the magnitude of the gap between expectations and perceived experience will be higher for global brands, consumers’ reactions following the same product-harm crisis will be stronger for global brands than for local brands. Hence, we propose:

H₂: Following a product-harm crisis, consumer evaluations of a global brand decline more than the evaluations of a local brand in emerging markets.

H₃: The effect of brand type (global vs. local) on the decline in consumer evaluations after a product-harm crisis is mediated by expectations of product performance.

Note that H₂ and H₃ question the commonly accepted belief that strong prior associations for global brands can act as a safeguard, shielding them from negative consumer responses in the face of adverse events (Davvetas, Diamantopoulos, and Liu 2020; Dawar and Pillutla 2000). Building on shifting standards theory, we posit that being global may be a disadvantage when the brand cannot meet higher consumer expectations. When perceived performance is subpar, being global may lead to a greater negative attitude change, compared with being local.

Moderating Role of Country Level of Development and Product Category

The country context plays a significant role in shaping consumers’ perceptions of and associations around global and local brands. In mature markets, there is a rich history and extensive experience with both global and local brands (Strizhakova and Coulter 2015). Some local brands, such as BMW and Beiersdorf, the owner of the Nivea brand from Germany, have successfully expanded globally, attaining a widespread presence (Guo 2013; Kim, Moon, and Iacobucci 2019). Consequently, expectations associated with global and local brands are relatively similar. Additionally, mature markets have experienced a negative shift in consumer attitudes toward global brands, partly due to trends such as “buy local” campaigns, global brand boycotts, and anticorporatist attitudes (Thompson and Arsel 2004).

However, in emerging markets, the exposure to global brands is more recent, mainly in the late 1980s and early 1990s. Examples of local brands achieving global recognition are not as common in emerging countries (Chattopadhyay, Batra, and Özsomer 2012), leading to a stronger differentiation between the concepts of globalness and localness (Davvetas et al. 2022). Due to the scarcity of home-grown global brands in emerging markets, global brands continue to hold greater prestige, status, and aspirational value, offering consumers the chance to be part of a desired global consumer culture (Alden, Steenkamp, and Batra 1999; Davvetas et al. 2022). They are seen as a passport to modernity and global citizenship (Strizhakova and Coulter 2015). In Holt, Quelch, and Taylor’s (2004) study, a Costa Rican expressed the aspirational effect: “Local brands show what we are; global brands show what we want to be” (p. 3).

Local brands, in contrast, derive their meaning from local symbolism and closer alignment with the local culture, catering to specific needs and tastes (Ger 1999). They provide a unique positioning that sets them apart from the more standardized positioning of global brands. Thus, there is a greater contrast between global and local brands in emerging markets (Batra et al. 2000; Davvetas et al. 2022), triggering home-country affinity and ethnocentrism effects for local brands, and aspirational and global-connectedness effects for global brands. Accordingly, country level of development constitutes a boundary condition in our conceptualization, and our proposed effects are limited to emerging market contexts:

H₄: The effect of brand type (global vs. local) on consumer expectations of product performance (and the consequential decline in evaluations following a product-harm crisis) is attenuated for consumers from mature markets.

Product category is another important factor that affects consumers’ perceptions of and preferences for global versus local brands (e.g., Özsomer 2012). Davvetas and Diamantopoulos (2018) argue that little is known about how product category influences consumer choices between global and local brands and call for further research. They demonstrate that global brands are preferred more than local brands within product categories that are instrumental for social signaling and identity construction. The same authors, in another study (2016), suggest that when global brands are more dominant in a product category, the globalness attribute becomes important in superiority assessment of the brands in that category. However, if local brands are more dominant in a category, localness becomes a critical assessment criterion. Put differently, mental schemas that consumers have for a specific product category shape their evaluations of global and local brands. For example, Davvetas and Diamantopoulos (2016) show that global brands are perceived as superior to local brands in technology categories but not in food categories. This is in line with the work of Alden, Steenkamp, and Batra (1999), who demonstrate that food products are frequently used to signal local consumer culture whereas technology products signal cosmopolitanism and global consumer culture. Several other studies also suggest that food products may benefit if they are marketed as local (Schuh 2007). For example, Gineikiene, Schlegelmilch, and Ruzeviciute (2016) demonstrate that local (vs. global) food products are perceived as healthier. Accordingly, we expect that in product categories that are associated with localness (e.g., food), being global will not necessarily lead to shifting standards of evaluation.

H₅: The effect of brand type (global vs. local) on consumer expectations of product performance (and the consequential decline in evaluations following a product-harm crisis) is attenuated for product categories associated with localness.

Company Response Following a Product-Harm Crisis

We now shift our focus on company response following a product-harm crisis, which is a critical determinant of consumer reactions to the failure (e.g., Aaker 1991; Hewett and Lemon 2019). Prior literature on harm crises suggests that companies need to act quickly after a crisis (Khamitov, Grégoire, and Suri 2020), offer a reasonable explanation (Mattila 2006), treat the customers fairly (Maxham and Netemeyer 2002), and provide fair compensation (Smith, Bolton, and Wagner 1999) to mitigate the negative effects of the crisis on consumer evaluations. Possible company response strategies vary along a continuum from unambiguous support (acceptance of responsibility, apology, remedial measures) to unambiguous stonewalling (denial of responsibility and absence of remedial measures) (e.g., Dawar and Pillutla 2000; Rasoulian et al. 2017). Provision of compensation is especially a strategically important decision, as it has direct implications on the financial resources needed to manage a crisis. Day and Landon (1977) demonstrate that the availability of compensation and the ease and convenience of obtaining it improve consumers’ reactions to a harm crisis.

It is important to understand when a compensation response will help the firm more. Dawar and Pillutla (2000) demonstrate that objectively identical company responses may have substantially different impact on consumers’ postcrisis evaluations, depending on their prior expectations about the brand. More specifically, prior beliefs and associations about the brand (Darley and Gross 1983), its perceived credibility (Erdem and Swait 1998), and its reputation (Siomkos and Kurzbard 1994) are important factors in how consumers will interpret and react to a response strategy following a harm crisis. For example, according to Dawar and Pillutla (2000), high prior expectations from a brand may buffer the effect of a crisis on brand equity. However, if the brand denies responsibility and does not offer any compensation, consumers may react more negatively. Similarly, Germann et al. (2014) demonstrate that high brand commitment alleviates negative consumer responses after a crisis but only for low-severity events. Interestingly, their results show that consumers with higher commitment levels react more negatively following high-severity crises. Therefore, in line with the established associations of global and local brands discussed previously, whether a brand is global or local should not only affect expected performance and postcrisis evaluations but also influence how consumers evaluate company response following a crisis.

The proximity that local brands have with consumers’ lifestyles and values helps them create emotional ties with their consumers (Kapferer 2002; Schuiling and Kapferer 2004). De Vries and Fennis (2019) imply that these emotional ties will increase consumer trust, especially in times of turbulence. Such emotional proximity and trust associations are naturally stronger for consumers with high ethnocentrism as they define their identities more with their patriotic values (Sharma, Shimp, and Shin 1994). Shimp and Sharma (1987) define ethnocentrism as favoring domestic products and rejecting imported products for patriotic and economic reasons. Consumer ethnocentrism manifests as a behavioral preference for domestic products, leading to repetitive consumption and spreading of positive word of mouth about them (Sharma 2015; Shimp and Sharma 1987).

We suggest that the differential emotional proximity to global and local brands (especially for highly ethnocentric consumers) will have direct implications for how consumers will react to company responses following a product-harm crisis, as prior beliefs were shown to influence blame attributions associated with a crisis (Laufer, Gillespie, and Silvera 2009). Previous work on blame attribution builds on studies analyzing judicial processes showing that prior beliefs about a defendant's social group may affect judgments of culpability (e.g., Jones 1991). Similarly, prior beliefs about a specific type of brand (global vs. local with differential emotional proximity) would play a role in the assessment of culpability and blame in a harm crisis context. Previous research has found that the degree of blame attributed to a company for a failure increases complaints (Richins 1983) and lowers purchase intentions (Siomkos and Kurzbard 1994). Therefore, we expect that consumers high in ethnocentrism will appreciate any compensation provided by a local brand more, compared with the same response by a global brand, due to their emotional ties and associated blame attributions. Hence, building on previous findings on the overall effects of compensation as a response strategy, and the different associations around global and local brands, we propose:

H_6a: Providing compensation to consumers after a product-harm crisis attenuates the adverse effects of the crisis on consumer evaluations, compared with a no-response strategy.

H_6b: Providing compensation after a product-harm crisis helps a local brand more in attenuating negative evaluations for highly ethnocentric consumers, compared with a global brand.

Overview of Studies

We test our hypotheses in four main and two supplementary studies. Study 1 tests the basic tenets of our conceptualization (H₁–H₃) within an emerging country (Mexico). The study aims to delineate shifting expectation dimensions and test their mediating influence on consumer evaluations after a product-harm crisis. Study 2 investigates the same hypotheses, using the expectation dimensions that derive the mediation shown in Study 1, this time for a single-incident product failure within another emerging country (Turkey). Studies 3a and 3b replicate this demonstration for a product-harm crisis scenario within an emerging and a mature market (boundary condition, H₄) respectively (Poland and Germany). Studies 3a and 3b also explore the attenuating effect of product category as another boundary condition (H₅). Study 4, then, inquires how company response (compensation vs. no compensation following a crisis) and consumer ethnocentrism affect consumers’ evaluations of global and local brands differently (H_6a, H_6b) in Mexico. Supplementary Study 5 provides indirect support for our underlying process explanation through consumer expectations by showing that the observed differential effects of global versus local branding on consumer evaluations disappear when these expectations are externally controlled (i.e., when shifting standards are removed). Supplementary Study 6 replicates the results of Study 4, more directly controlling for price and prestige, and finds evidence for the differential effect of brand type on consumer evaluations for equally strong global and local brands.

We use scenario-based designs from previous literature in our studies, introducing product failure and product-harm crisis instances ¹ (and company response strategies) for global versus local brands across different product categories. We ran a pretest with 151 participants (M_age = 24.13 years; 46% female, 54% male) on an online panel (Prolific) to select the product categories to be used in our studies. We measured the globalness versus localness associations of a set of categories typically studied in previous research (home appliances, electronics, pharmaceuticals, self-care, food, and furniture; e.g., Davvetas and Diamantopoulos 2016). Participants were provided with product examples for each category to ensure consistency (see Web Appendix A for the scale items and product examples). A repeated-measures analysis of variance (ANOVA) test indicated a significant effect of the category factor (F(5, 745) = 94.06; p < .01), where the electronics category was perceived to be the most associated with globalness (M = 5.44), followed by pharmaceuticals (M = 4.76), small home appliances (M = 4.61), self-care (M = 3.89), furniture (M = 3.45), and finally food (M = 2.89). Follow-up contrast tests showed that the electronics category was perceived to be significantly more associated with globalness than all the other categories (ps < .01). Globalness associations of home appliances and pharmaceuticals were similar (p > .10) and significantly more than self-care, furniture, and food (ps < .01). Accordingly, we selected the electronics and pharmaceuticals categories to test our propositions, and the food category as a boundary condition.

Across our studies, we used different samples (students and online panel participants) from multiple countries (Mexico, Turkey, Poland, and Germany). Our country selections were partly motivated by participant access (student subject pool and Prolific panel population), and partly to differentiate between emerging versus mature market contexts. The Fiscal Monitor published by the International Monetary Fund (2021, pp. 74–75) lists Mexico, Turkey, and Poland as emerging markets and Germany as a mature market.

We used fictitious brands across the studies to avoid any brand-induced confounds such as prior brand attitudes, relationship strength (Barbarossa and Mandler 2021), and perceptions of brand strength and familiarity (Dimofte, Johansson, and Ronkainen 2008). We controlled perceived price in our scenarios (i.e., price is irrelevant, free or equal price) to avoid any confounding effects of possible prior beliefs respondents might have about affordability of global versus local brands and related quality associations (Davvetas, Sichtmann, and Diamantopoulos 2015). Our scenarios also directly control for brand strength and are focused within the scope of well-established (global and local) brands with prominent market position (i.e., “well-known brand with market leader position in total sales with over 20 years of operations”).

Study 1: Expectation Dimensions Underlying the Shifting Standards for Global and Local Brands

This study provides insights into the shifting standards construct by establishing which of the dimensions of consumer expectations for global versus local brands drive the mediation effect on consumer brand evaluations.

Method

Design

We conducted a one-way between-subjects experiment with brand type (global vs. local) as the independent variable. Our dependent variable was the decline in consumer evaluations following a product-harm crisis (Germann et al. 2014), and the mediating variable was consumer expectations of product performance.

Procedure and measures

One hundred ninety-four people from Mexico participated in this experiment via an online panel. Two participants answered the attention check question (“I am carefully reading the questions. This is an attention check; please mark strongly agree.”) incorrectly and were excluded from the study, leaving a final sample of 192 participants (M_age = 25 years; 59% female, 40% male, 1% nonbinary). Participants were randomly assigned to the brand type manipulation, in which they were informed about a recently launched computer by a global/local brand (see Web Appendix B for the scenarios). Both brand conditions ensured a given level of brand strength and revealed the same price to prevent any potential confounding effect of differential price/quality associations. Note that we did not provide specific brand origin (domestic vs. nondomestic) information across the global and local brand conditions. Given the historical rarity or lack of home-grown global brands in emerging markets (e.g., Chattopadhyay, Batra, and Özsomer 2012; Davvetas et al. 2022), respondents may refer to a domestic local brand and a nondomestic global brand in their evaluations. Accordingly, brand origin was not kept constant in this study, but this will be addressed in subsequent studies. After participants reviewed the scenarios, brand evaluations (Ng, Faraji-Rad, and Batra 2021) and dimensions of consumer expectations (Dawar and Pillutla 2000; Siomkos and Kurzbard 1994) were measured (see Web Appendix B for the scale items and reliability coefficients). Next, they viewed a short newspaper article that mentions the product-harm crisis (see Web Appendix B), indicated their revised evaluations, and completed the experiment by reporting their age and gender.

Decline in consumer evaluations was operationalized by subtracting postcrisis scores from baseline scores. As this was a ratio-scale measure, we checked whether the normality assumption was supported. A significant Shapiro–Wilk test (p < .01) indicated deviation from normal distribution. Accordingly, we applied a natural log transformation to mitigate potential outlier effects. Greater decline in magnitude meant a more negative consumer reaction to the product-harm crisis intervention. The operationalization of this measure as a decline and its calculation as a difference score are particularly relevant to our conceptualization and in alignment with our hypotheses since consumers are expected to adjust their evaluations more negatively for a global (vs. local) brand following a product-harm crisis. Previous research has established the psychometric properties of difference scores in social psychology and has justified their use when conceptually relevant (e.g., Gollwitzer, Christ, and Lemmer 2014; Tormala, Clarkson, and Petty 2006, Experiment 4; Zemborain and Johar 2007, Experiment 3).

Results and Discussion

ANOVA tests with each dimension of consumer expectations as the dependent variable and brand type as the independent variable (H₁) indicated that consumers have higher expectations of “being of high quality” (F(1, 190) = 5.63, p < .05), “produced with newest technology” (F(1, 190) = 5.50, p < .01), “being leader in its field” (F(1, 190) = 8.60, p < .01), and “being of high standards” (F(1, 190) = 5.41, p < .05) for global brands, compared with local brands, justified by the significant main effect of brand type (see Table 1 for the means). Consumers did not expect a significant difference between global and local brands on the trustworthiness dimension (F(1, 190) = 1.59, p > .2). Another ANOVA test with the decline in consumer evaluations as the dependent variable and brand type as the independent variable revealed a significant effect of brand type (M_global = −3.90, M_local = −3.52; F(1, 190) = 4.29, p < .05). (See Table 2 for the means before and after the product-harm crisis intervention.)

OPEN IN VIEWER

Table 1.

Study 1 Means, Mean Differences, and Mediating Influences on Global Versus Local Brand Evaluations.

Expectation Dimensions	Global Brand Mean	Local Brand Mean	Mean Difference	Mediating Influence on Evaluations (95% CI)
Of high quality	5.35	4.98	p < .05	No [−.0653, .0008]
Produced with newest technology	5.21	4.73	p < .01	Yes [−.0742, −.0038]
Leader in its field	4.88	4.26	p < .01	No [−.0500, .0103]
Of high standards	5.01	4.61	p < .05	Yes [−.0710, −.0019]
Trustworthy	5.27	5.08	p > .2	No [−.0636, .0112]

OPEN IN VIEWER

Table 2.

Means for Consumer Evaluations Before and After the Product Failure or Harm Crisis Interventions Across the Main Studies.

A: Study 1: Mexico

Precrisis

Postcrisis

Decline

Global

5.89

1.99

−3.90**

Local

5.75

2.23

−3.52**

B: Study 2: Turkey

Precrisis

Postcrisis

Decline

Global

5.06

2.97

−2.09**

Local

4.73

3.61

−1.11**

C: Study 3a: Poland

Product Category: Vaccine

Product Category: Food

Precrisis

Postcrisis

Decline

Precrisis

Postcrisis

Decline

Global

5.40

2.85

−2.55*

5.86

3.04

−2.82

Local

5.29

3.13

−2.16*

5.98

2.77

−3.21

D: Study 3b: Germany

Product Category: Vaccine

Product Category: Food

Precrisis

Postcrisis

Decline

Precrisis

Postcrisis

Decline

Global

4.97

3.19

−1.78

5.48

2.80

−2.68**

Local

5.26

3.45

−1.81

5.63

2.30

−3.33**

E: Study 4: Mexico

Consumer Ethnocentrism: Low

Consumer Ethnocentrism: High

Company Response: No Compensation

Company Response: Compensation

Company Response: No Compensation

Company Response: Compensation

Precrisis

Postcrisis

Decline

Precrisis

Postcrisis

Decline

Precrisis

Postcrisis

Decline

Precrisis

Postcrisis

Decline

Global

5.91

2.05

−3.86

5.97

3.33

−2.64

5.88

2.41

−3.47

6.00

2.79

−3.21*

Local

5.58

2.47

−3.11

5.33

2.50

−2.83

5.60

2.21

−3.39

5.78

3.18

−2.60*

*p < .1, **p < .05 (significance of mean differences within a column for a study).

Notes: The table shows nontransformed mean consumer evaluations before and after the product failure or harm crisis interventions and the resulting decline (key dependent variable of interest) on a scale of 1 to 7.

We conducted a series of mediation analyses (using the PROCESS computational macro, Model 4, for SPSS; Hayes 2017) to explore whether the effect of brand type (global vs. local) on the decline in consumer evaluations would be mediated by each dimension of expectations. Our results revealed that the effect of brand type (global vs. local) on the change in performance evaluations is mediated by consumer expectations (H₃) and is specifically carried through the “high standards” and “produced with newest technology” expectation dimensions (see Table 1). This is in line with the conceptual discussion and existing survey-based research detailed previously (e.g., Balabanis and Diamantopoulos 2016).

Study 2: Shifting Standards for Global and Local Product Failures

This study focuses on the context of single-incident product failures such as a product break-down to test our underlying contention that consumers from emerging markets use shifting standards when evaluating global versus local brands. The study tests the effect of global versus local branding on the decline in consumer evaluations following a product failure (H₂), and the mediating effect of expectations (H₁, H₃).

Method

Design

We conducted a one-way between-subjects design with brand type (global vs. local) as the independent variable. Our dependent variable was the decline in consumer evaluations for the brand following a failure (Germann et al. 2014), and the mediating variable was consumer expectations of product performance.

Procedure and measures

Two hundred thirty-two undergraduate students (M_age = 21 years; 56% female, 44% male) from a university in Turkey participated in this laboratory experiment. Participants first read a scenario where they were asked to assume they were planning to purchase a small electronics product and had decided on a “well-known global [local] brand.” This served as our brand type manipulation, qualified by the “well-known” prefix to ensure a given level of brand strength. We controlled for potential confounding effects of differential price/quality associations by noting price not to be a major concern for this purchase as it would not affect their budget. After reviewing the scenario, participants responded to questions measuring their baseline evaluations and expectations regarding the product and viewed the product failure intervention (see Web Appendix C for the scenarios). Participants then responded to another set of questions indicating their revised evaluations following the failure, and completed the experiment by responding to demographic questions for descriptive purposes.

As Study 2 focused on a single-incident product failure, we measured the decline in consumer evaluations as the change in perceived defect probability measured before and after the product failure. Consumers’ perception of defect probability has been used previously as a reflection of perceived product reliability (Chen, Kalra, and Sun 2009) and has been directly shown to be linked to consumers’ willingness to pay (Abbey et al. 2019). We subtracted the perceived defect probability after the product failure from baseline perceptions, as indicated by the participants on seven-point scales (1 = “extremely improbable,” and 7 = “extremely probable”) in response to the question “How would you rate the probability of the product to be defective?” A higher difference meant a greater decline in consumers’ performance evaluations for the brands. As in Study 1, a significant Shapiro–Wilk test (p < .01) indicated deviation from normal distribution for our dependent variable, and we performed a natural log transformation. We measured consumer expectations using the two items that Study 1 revealed as the drivers of the mediation, where participants rated their agreement with the following statements on a seven-point scale: (1) “produced with the newest technology,” and (2) “has high standard products” (r = .72, p < .01).

Results and Discussion

An ANOVA test on expectations revealed that consumers have higher expectations for global brands (M_global = 5.07) compared with local brands (M_local = 4.64), justified by the significant main effect of brand condition (F(1, 230) = 7.45, p < .01) (H₁). A second ANOVA test on decline in performance evaluations showed that the decline was significantly greater for global brands (M_global = −2.09) than for local brands (M_local = −1.12; F(1, 230) = 12.23, p < .01) supporting H₂. (See Table 2 for the means before and after the product failure intervention.) To test H₃, that the effect of brand type on the decline in performance evaluations after the product failure would be mediated by expectations of global and local brand performance, we used the PROCESS computational macro (Model 4) for SPSS from Hayes (2017). The total effect of brand type on the expectations was significant (β = −.2074, t(230) = −2.7303, p < .01) (H₁), as reported previously. Controlling for brand type, expectations had a significant and positive effect on the decline in consumer evaluations (β = .1820, t(229) = 13.0272, p < .01), indicating that the negative impact of the product failure increased with higher expectations, while brand type had a marginally significant effect on the decline (β = −.0310, t(229) = −1.8937, p = .06). The indirect path of the effect of brand type on the decline in evaluations through expectations was significant, with the 95% confidence interval excluding zero (effect size = −.0377, CI: [−.0717, −.0103]). Hence, we conclude that the differential effect of global versus local branding on the decline in consumer evaluations after a product failure is carried through consumer expectations, supporting H₃.

The following studies explore the proposed attenuating effects of product category and country context and incorporate key individual difference variables such as ethnocentrism and category knowledge. We also control for potential confounding effects of price more directly.

Study 3: Shifting Standards for Global and Local Product-Harm Crises

This study explores differential consumer reactions toward global and local brand product-harm crises through expectations (H₁, H₂, H₃) across emerging and mature market contexts (H₄) and for two product categories with different globalness versus localness associations (H₅). As random assignment was not possible across country samples, we ran two separate experiments in Poland (Study 3a) and Germany (Study 3b) with identical designs. Results were analyzed and interpreted separately. We expect to replicate the findings of Study 2 for harm crisis instances for the vaccine category (with globalness associations) in Poland (emerging market). The food category (with localness associations) and Germany (mature market) serve to demonstrate the boundary effects.

Method

Design

Studies 3a and 3b used 2 (brand type: global vs. local) × 2 (product category: vaccine vs. food) between-subjects designs. The dependent variable was the decline in consumer brand evaluations following a harm crisis, and the mediating variable was consumer expectations of product performance.

Procedure and measures

Our product category selections were motivated partly by their differential globalness/localness associations, established with our category pretest reported previously. Aiming to control for possible confounding effects of price more directly, we opted for free distribution of the products in the Study 3 scenarios. Free distributions of vaccines and food by institutions were highly prevalent in the recent COVID-19 environment. Accordingly, our chosen categories served as realistic and relevant input for our scenarios. Participants read a short scenario about a fictitious AVA Company, positioned as “a well-known global [local] company,” offering 1 million free vaccinations [ready-to-eat lunch menus] across the country in the next few months as support “during this time of the pandemic.” (See Web Appendix D for the scenarios.)

We conducted a pretest with participants from an online panel (Prolific) residing in Germany and Poland (N = 161; 31% female, 68% male, 1% nonbinary), to establish the effectiveness and credibility of our brand/category scenarios used in Study 3. The pretest revealed that the brand/category manipulation worked effectively and that both scenarios were perceived to be believable and relevant (see Web Appendix D for details). Additionally, we examined perceived warmth of Germany and Poland. Halkias, Davvetas, and Diamantopoulos (2016) propose that perceptions of higher warmth for a country may result in a general liking for that country, which, in turn, can affect attitudes toward its products. Our findings indicated that there was no significant difference in the perceived warmth of Germany and Poland.

After reviewing the pretested scenarios, participants in Studies 3a and 3b indicated their baseline evaluations and expectations. We measured evaluations the same way as we did in Study 1 (r = .68, .87, ps < .01, for Study 3a and Study 3b respectively) and expectations as in Study 2 (r = .51 and r = .47, ps < .01, respectively). Participants next read a product-harm crisis intervention scenario about the AVA Company (adapted from Germann et al. 2014; Whelan and Dawar 2016; see Web Appendix D) and responded to a second set of evaluation questions measuring their reactions to the crisis. As suggested by Germann et al. (2014), the responsibility for the product-harm crisis was kept ambiguous. We operationalized the decline in consumer evaluations as in previous studies. We performed a natural log transformation and reported untransformed means for ease of interpretation. Participants completed the study by responding to demographic questions for descriptive purposes and reviewing a debrief about the fictitious crisis scenario. Results from the Poland (Study 3a) and Germany (Study 3b) samples are presented next.

Study 3a Participants: Poland

Three hundred two people from an online platform (Prolific) in Poland completed the study. Two participants indicated they were not Polish, and nine participants failed the instructional attention check question (as in Study 1; Oppenheimer, Meyvis, and Davidenko 2009), leaving us with a final sample of 291 (M_age = 22.97 years; 35% female, 64% male, 1% nonbinary).

Study 3a Results and Discussion

An ANOVA test on consumer expectations with brand type (global vs. local) and product category (vaccine vs. food) as the independent variables revealed that participants had directionally higher (while not significant) expectations for global brands (M_Global = 5.11, M_Local = 4.88; F(1, 287) = 3.55, p = .06). The main effect of product category and the interaction were not significant (p > .3). Further exploring the results with follow-up contrast tests, we found a significant effect of brand type within the vaccine category in the expected direction (H₁; M_LocalVaccine = 4.84, M_{GlobalVaccine} = 5.21; F(1, 287) = 3.79, p = .05). The effect of the brand type on expectations was attenuated for the food category (high in localness associations) (M_LocalFood = 4.89, M_GlobalFood = 5.01; F(1, 287) = .50 p > .5).

To test H₂, we conducted another ANOVA to explore the effect of brand type and product category on the decline in consumer evaluations. The analysis revealed a significant main effect of product category (M_Vaccine = −2.35, M_Food = −3.02; F(1, 287) = 9.62, p < .01) and a significant interaction effect (F(1, 287) = 4.83, p < .05), but not a significant main effect of brand type (p > .8) (see Table 2 for the means before and after the intervention). Follow-up contrast tests revealed that the decline in consumer evaluations was directionally higher (while not significant) for the global brand compared with the local brand within the vaccine category (with high globalness associations) (M_{GlobalVaccine} = −2.55, M_LocalVaccine = −2.16; F(1, 287) = 2.88, p = .09). Brand condition did not have a significant effect within the food category (with high localness associations; p > .15).

Finally, we tested whether consumer expectations mediated the effect of brand type on the decline in evaluations within the vaccine product category (H₅) with the PROCESS computational macro (Model 4) for SPSS. The effect of brand type on expectations was marginally significant (β = .6235, t(141) = 1.75, p = .08). Controlling for brand type, expectations had a significant and positive effect on the decline in evaluations (β = .0650, t(141) = 2.95, p < .05), indicating that the negative impact of the product-harm crisis increased with higher expectations, while the brand type no longer had a significant effect on the decline (p > .2). The indirect path of the effect of the brand type on the decline in consumer evaluations through expectations was significant within the vaccine category, with the 95% confidence interval excluding zero (effect size = .022, CI: [.0040, .0560]). Hence, we can conclude that the effect of brand type (global vs. local) on the decline in consumer evaluations following a product-harm crisis is carried through performance expectations in emerging market contexts, especially for product categories highly associated with globalness.

Next, we inquire how these effects might differ in a mature market context.

Study 3b Participants: Germany

Two hundred ninety-seven people from an online platform (Prolific) in Germany completed the study for a small compensation, with an identical procedure to Study 3a. Two participants failed our instructional attention check question (Oppenheimer, Meyvis, and Davidenko 2009), and 47 participants indicated that they were not German and were originally from emerging markets, leaving us with a final sample of 248 (M_age = 30.32 years; 38% female, 60% male, 2% nonbinary). Per the scope of our conceptual development, we do not expect to find an effect of brand type on consumer expectations and evaluations in a mature market context (H₄).

Study 3b Results and Discussion

An ANOVA test with expectations as the dependent variable and brand type and product category as the independent variables revealed a significant main effect of category on consumer expectations (M_Vaccine = 3.88, M_Food = 3.01; F(1, 244) = 3.21, p < .01) and a marginally significant interaction (F(1, 244) = 3.04, p = .08). Brand type did not have a significant effect on consumer expectations (p > .5). Follow-up contrast tests demonstrated a directional effect of brand type within the vaccine category in the opposite direction (M_LocalVaccine = 4.07, M_{GlobalVaccine} = 3.68; F(1, 244) = 3.21, p = .08). German consumers had directionally (but not significantly) higher expectations for a local (vs. global) vaccine. This finding is aligned with the successful efficacy rate of the Pfizer-BioNTech vaccines in the market. The effect of the brand type on expectations disappeared for the food category (M_LocalFood = 2.91, M_GlobalFood = 3.11; F(1, 244) = .76, p > .4).

Next, we conducted another ANOVA test on the decline in consumer evaluations following the crisis intervention. The results revealed a significant main effect of product category (M_Food = −3.00, M_Vaccine = −1.79; F(1, 244) = 35.89, p < .01) and a marginally significant effect of brand type (M_Global = −2.23, M_Local = −2.56; F(1, 244) = 2.93, p = .09), but not a significant interaction (p = .12; see Table 2 for means before and after the intervention). Further exploring the simple contrasts, we found that the decline in consumer evaluations did not differ between the global and the local brand conditions for the vaccine category in Germany (M_{GlobalVaccine} = −1.78, M_LocalVaccine = −1.81; p > .7). This is in line with the scope of our research to be limited to emerging market contexts (H₄). Our proposed differential effects for global versus local product-harm crises were not replicated in a mature market context.

Interestingly, we observed an opposite significant effect within the food category, where the decline in evaluations was higher for the local brand than for the global brand following a product-harm crisis (M_GlobalFood = −2.68, M_LocalFood = −3.33; F(1, 244) = 4.41, p < .05). Thus, we see that the local brand gets more strongly penalized when the product-harm crisis involves a category associated with high localness within a mature market context. Although we did not have a formal hypothesis about this effect, it is consistent with insights from the work of Davvetas and Diamantopoulos (2016, 2018) on category influences discussed previously: if local brands are more dominant in a product category, the localness attribute becomes the critical superiority assessment criterion and, in turn, hurts the brand more.

Study 4: Company Response Following a Product-Harm Crisis

The results of Study 3 results have established the scope of our propositions within emerging market contexts (H₄) and for categories associated with globalness (H₅). Therefore, in this study, we focus on another emerging market, Mexico, and another category with globalness associations, electronics (per the category pretest). We also shift our attention to the effect of global and local company response strategies (compensation vs. no compensation) on consumer reactions following a product-harm crisis (H_6a), with a consideration of the moderating influence of consumer ethnocentrism (H_6b).

Method

Design

We used a 2 (brand type: global vs. local) × 2 (company response: no compensation vs. compensation) × 2 (ethnocentrism: low vs. high) between-subjects design in this study. Both the brand and the company response variables were manipulated through a scenario, and ethnocentrism was included as a measured variable. We additionally controlled for category knowledge of individual participants as a covariate.

Pretest

First, we ran a pretest (N = 133) via an online panel (Prolific) in Mexico to examine the effectiveness of our brand condition manipulation and equivalence of the scenarios used in the main study. Three participants failed our instructional attention check question, leaving us with a final sample of 130 participants (M_age = 24.34 years; 45% female, 55% male). We measured perceived globalness versus localness of the brand with the same scale used in Study 3 (Cronbach's alpha = .93) as a manipulation check for the main experiment. We measured perceived severity of the crisis to control whether participants have different severity perceptions for a product-harm crisis by a global versus local brand. Finally, we measured scenario equivalence. Analyses of the pretest demonstrated that global and local brand manipulations worked effectively and that there was no significant difference in participants’ perception of product-harm crisis severity and scenario characteristics (see Web Appendix E for detailed pretest procedure and results).

Procedure and measures

The main experiment was conducted with 353 participants from an online panel (Prolific) from Mexico. Sixteen participants failed our instructional attention check question and 15 participants indicated that they were not Mexican, leaving us with a final sample of 322 participants (M_age = 25.05 years; 53% female, 45% male, 2% nonbinary). Participants read a short scenario about a fictitious AVA Company, positioned as “a well-known global [Mexican] electronics company.” They were further informed about the company’s recent launch of a TV, available at a discounted price (adapted from Dawar and Pillutla 2000). A specific price was indicated as part of the scenario, equal across the global versus local brand conditions, to control for potential price/quality associations (see Web Appendix E for the full scenario).

After reviewing this information, participants indicated their baseline evaluations of the AVA Company. Then, they were presented with a product-harm crisis intervention adapted from Barbarossa and Mandler (2021), in which a newspaper article informed them about instances of fires or electrical damage due to overheating of the electrical equipment of AVA TVs, including reports of injuries and potential causes of the incidents, creating ambiguity about AVA Company’s responsibility (see Web Appendix E). Following the crisis intervention, participants read about the company's response in the form of no compensation versus compensation in line with their randomly assigned conditions (see Web Appendix E). Then, they reported their revised evaluations and indicated their thoughts and feelings about the situation in open-ended protocols for further insight. They completed the study by responding to ethnocentrism, category knowledge, and demographic questions and by reviewing the debrief about the harm crisis scenario.

Consumer evaluations were measured (r = .71, p < .01) and operationalized as in previous studies. We again conducted natural log transformation on this measure to support the normality assumption (Shapiro–Wilk test; p < .01); untransformed means are reported for ease of interpretation. Ethnocentrism was measured with a three-item, seven-point scale (adapted from Shimp and Sharma 1987; see Web Appendix E). As Study 4 focused on the moderating influence of ethnocentrism, we converted the continuous scale to high versus low ethnocentrism through a median split procedure. Category knowledge covariate was measured with a three-item, seven-point scale (adapted from Oliver and Bearden 1985; see Web Appendix E).

Results and Discussion

We conducted an ANOVA test with the decline in consumer evaluations as the dependent variable; brand type, company response, and ethnocentrism as the independent variables; and category knowledge as a covariate. Our results revealed a significant three-way interaction between the brand, company response, and ethnocentrism variables (F(1, 314) = 3.96, p < .05). Brand type had a significant effect on the decline in evaluations (M_Global = −3.30, M_Local = −2.98; F(1, 314) = 4.42, p < .05) (see Table 2 for the means before and after the intervention). The main effect of company response was also significant: providing compensation led to a smaller decline in evaluations following the harm crisis, supporting H_6a (M_Compensation = −2.82, M_{NoCompensation} = −3.45; F(1, 314) = 14.44, p < .01). The covariate, category knowledge, did not have a significant effect (p > .3).

Exploring the results with simple contrast tests, we found a directionally greater (while not significant) decline in consumer evaluations for the global (vs. local) brand (M_{LocalNoCompensation} = −3.24, M_{GlobalNoCompensation} = −3.67; F(1, 314) = 3.68, p = .056) within the no-compensation condition in line with our results from the previous studies. Furthermore, for participants with higher ethnocentrism scores, we observed that providing compensation helped the local brand directionally more than the global brand within the compensation condition (M_{LocalCompensationHighEthno} = −2.60, M_{GlobalCompensationHighEthno} = −3.21; F(1, 314) = 3.28, p = .07), providing directional support for H_6b.

Post Hoc Analyses

We did not have formal hypotheses on the processes underlying differential consumer reactions to compensation by global versus local brands since our full set of hypotheses already entailed multiple mediation routes and collecting too many discrete measures has the potential to disrupt the internal validity of experimental designs (Schwarz and Oyserman 2001, p. 150). Instead, we opted to analyze the open-ended protocols collected as part of Study 4 as additional insight on consumers’ cognitive and affective responses. Our conceptual development discussed the importance of blame attribution in consumers’ evaluations of harm crises and firm response strategies. Accordingly, two independent research assistants, unaware of the hypotheses and conditions, rated participants’ open-ended responses (where 1 = blame attributed to causes other than the company, 4 = no mention of blame, and 7 = blame attributed to the company). An ANOVA test on this measure, with brand type, company response, and ethnocentrism as the independent variables, showed a significant three-way interaction (F(1, 314) = 6.19, p < .05). Follow-up contrast tests showed that the blame attributed to a local brand by highly ethnocentric consumers was directionally lower (while not significant) than the blame attributed to a global brand, when the company provided compensation (M_{LocalCompensationHighEthno} = 4.75, M_{GlobalCompensationHighEthno} = 5.19; F(1, 314) = 3.37, p = .07). Therefore, we can indirectly conclude that blame attributions contributed to our previously reported observations per H_6b about the differential effect of global versus local company response strategies on consumer reactions to harm crises.

Supplementary Studies

The four main studies showed that consumers employ shifting standards when evaluating the performance of global and local brands and respond differently to individual and large-scale product failures of these brands as well as to company response strategies. The shifting standards are materialized through consumers’ differential expectations from global and local brands in emerging markets. To enhance our understanding of the process through expectations and eliminate alternative explanations related to price and prestige, we conducted two supplementary studies with a continued focus on company response strategy.

Supplementary Study 5

We used a 2 × 2 between-subjects design to test the interactive effect of brand type (local vs. global) and company response strategy (no compensation vs. compensation) on the decline in consumer evaluations following a product failure intervention. Importantly, we externally controlled the baseline performance expectations from global and local brands in this study.

We conducted a computer-based experiment with 283 participants (M_age = 25.22 years; 40% female, 60% male) from a consumer study panel. Participants initially reviewed scenarios providing information about a well-known local/global small home appliance brand in Turkey. The scenarios included results of a “Consumer Evaluations Survey,” which were identical for both global and local brand conditions, thus establishing positive and equitable baseline expectations for both brand types. By doing so, we aimed to control for shifting standards and predicted that when the expectations are set at a similar level, we would not observe a difference in consumers’ reactions toward global and local brands following a product failure. After reviewing the scenarios, participants reported their evaluations and expectations. Next, participants read about a product-failure intervention and a company response (adapted from Germann et al. 2014; Whelan and Dawar 2016; see Web Appendix F for the scenarios), and reported their revised brand evaluations as in the main studies (r = .78, .66, ps < .01). Finally, they indicated perceived localness/globalness of the brand (as a manipulation check; seven-point scale; 1 = “Local,” and 7 = “Global”), and their age and gender.

We conducted an ANOVA test on perceived localness/globalness of the brand and found a significant effect (M_Local = 3.44, M_Global = 4.05; F(1, 281) = 6.26, p < .05), establishing the effectiveness of the brand type manipulation. Next, we conducted an ANOVA test on consumer expectations and did not find a significant effect of brand type (M_Global = 5.45, M_Local = 5.71; F(1, 281) = 2.43, p = .12). This result suggested that sharing the results of the “Consumer Evaluations Survey” with the participants successfully set their expectations for global and local brands at similar levels. Another ANOVA test to examine the effect of brand type and company response on the decline in consumer evaluations found only a significant main effect of the company response condition (M_Compensation = −.65, M_{NoCompensation} = −1.48; F(1, 281) = 10, p < .05). As expected, we did not find a significant main effect of brand type (F(1, 281) = .01, p = .92) nor an interaction (F(1, 281) = .89, p = .35). As consumers had similar expectations from both the global and local brands, we no longer observed the greater negative effect of the higher expectations for the global brand on the decline in consumer evaluations, providing additional indirect support for our process explanation.

Supplementary Study 6

This study, designed as a modified replication of Study 4, focused on the compensation condition of company response and used a different electronics product, laptops. We used a 2 × 2 between-subjects design to test the effects of brand type (local vs. global) and consumer ethnocentrism (low vs. high) on expectations and decline in consumer evaluations following a product-harm crisis. Importantly, the product was positioned as a newly launched and prestigious computer in the scenario with a relatively high price for the category. Thereby, we intended to rule out the potential confounding effect of price and quality associations for global brands by focusing on a high-priced product to show that the mere designation of a brand as global or local matters.

We conducted a computer-based experiment with 188 participants (M_age = 28.73 years; 48% female, 50% male, 2% nonbinary) from an online panel (Prolific). Participants first read the scenarios that provided information about a global or local computer brand in Mexico (See Web Appendix G) and reported their evaluations and performance expectations. Subsequently, they were presented with a product-harm crisis intervention, which included information about the company's decision to compensate (adapted from Germann et al. 2014; Whelan and Dawar 2016; see Web Appendix G). Participants completed the study by reporting their revised evaluations, ethnocentrism, category knowledge, age, and gender.

We measured and operationalized evaluation decline, performance expectations, ethnocentrism, and category knowledge the same way as we did in the other studies (r = .70, .27, ps < .01, Cronbach's alpha = .67, .89). We conducted an ANOVA test on expectations and found a significant effect of brand type (M_Global = 6.03, M_Local = 5.54; F(1, 187) = 16.40, p < .01). We did not find a significant effect of ethnocentrism (p = .97), nor an interaction (p = .86). Category knowledge did not have a significant effect (p = .17). The results showed that performance expectations for the global brand were still significantly higher compared with the equally strong and prestigious local brand. Another ANOVA test on the decline in consumer evaluations revealed a marginally significant interaction effect (p = .08), but no main effects (ps > .3). Category knowledge did not have a significant effect (p = .41). Simple contrast tests revealed that the proposed shifting standards effect was replicated such that evaluations of highly ethnocentric consumers declined more for the global brand following the crisis, even with company compensation, compared with the local brand (M_{GlobalHighEthno} = −3.39, M_{LocalHighEthno} = −2.69; F(1, 187) = 4.60, p < .05).

General Discussion

In this research, we build on shifting standards theory from social psychology to explain consumers’ disparate reactions toward global versus local brands within the context of product failures and product-harm crises. We introduce brand globalness and localness as superordinate mental categories, highlighting how the mere classification of a brand as global or local can evoke distinct associations and within-category judgment standards for consumers in emerging markets. This categorization shapes consumers’ performance expectations and evaluations following product-harm crises, as well as their responses to company actions. Thus, this work brings together two important streams of research, global branding and product-harm crisis, in an integrated conceptual model and provides an empirical test for the hypothesized effects of brand type with a careful consideration of boundary conditions. We show that consumers in emerging markets have higher initial expectations from global brands, which establishes a heightened performance threshold. Consequently, when a global brand faces a product-harm crisis, the perceived gap between expected and actual performance intensifies, leading to more negative consumer reactions compared with local brands. Even though being global is associated with higher quality, prestige, and aspirational value (Batra et al. 2000; Davvetas et al. 2022; Strizhakova and Coulter 2015), our results imply that the associations with and the expectations for being global might put brands in a disadvantageous position following a product-harm crisis.

Through a series of four main and two supplementary studies, we find support for our predictions using different product failure and harm crisis scenarios, across different product categories, and for multiple emerging market contexts. First, our results provide insights on the composition of consumer expectations toward global and local brands. Study 1 reveals that consumer expectations are higher for global (vs. local) brands on many performance dimensions (Table 1). However, only the consumer expectations for global brands' utilization of newest technologies and for their products being of higher standards lead to more negative consumer reactions following a product-harm crisis. Next, we demonstrate that consumers in emerging markets have higher expectations for global (vs. local) brands, and that the higher expectations mediate the effect of the brand type (global vs. local) on negative consumer evaluations after a product failure (Study 2) or a product-harm crisis (Study 3a). We additionally show that these effects are attenuated for product categories associated more closely with localness, such as food, and for consumers in mature market contexts (Studies 3a/3b). Our results also reveal that consumers high in ethnocentrism appreciate any compensation provided by a local brand directionally more when compared with the same response by a global brand (Study 4). In two supplementary studies, we strengthen the empirical evidence for our conceptualization and rule out alternative explanations by showing that the observed differential effects of global versus local branding on consumer evaluations disappear when these expectations are externally controlled (Study 5), but persist for equally strong global and local brands when price and prestige are directly controlled (Study 6).

Theoretical Implications

Our systematic investigation of the effects of brand type (global vs. local) and different company crisis response strategies has important contributions from a theoretical standpoint. This research contributes to extant international branding and product-harm crisis literatures by focusing on the role of shifting standards for global versus local brands within the context of product-related failures and harm crises. We introduce brand type as a superordinate mental categorization variable, which affects consumers’ reactions to harm crises and company responses in emerging markets. We extend the previous work that utilizes schema theory to explain differential consumer expectations for global versus local brands (e.g., Davvetas and Diamantopoulos 2016; Xie, Batra, and Peng 2015), by replicating their findings and providing a novel theoretical explanation through shifting standards and expectation disconfirmation theories (Biernat and Manis 1994; Oliver 1980).

We also show what kinds of expectations differ and which of these shifts (being of high standards and produced with new technology) drive the differential consumer response to adverse events. Our work offers new insights to the product-harm crisis literature by examining the moderating impact of product category associations, country level of development, and level of ethnocentrism on consumer evaluations after a crisis. Furthermore, we contribute to the extant branding literature by providing new insights concerning the role of global versus local brands in a product-harm crisis context.

Managerial Implications

Since product failures and product-harm crises are common in the marketplace with important potential impact on consumer attitudes and behavior, managers need to understand the processes to handle them. Our results are of practical importance for managers since they clarify the consequences of brand positioning strategies along global versus local associations and the choice of company response tactics for global and local brands.

Our work highlights that while a global brand enjoys higher performance expectations than its local counterpart in emerging markets (Davvetas, Diamantopoulos, and Liu 2020; Mandler, Bartsch, and Han 2021), it is more vulnerable in case of a harm crisis. Are companies like P&G and Unilever, which are eliminating many of their local brands to provide greater support for their global brands (Schuiling and Kapferer 2004) to leverage positive customer associations with their global brands, in fact taking a risk? Could global brands alleviate negative effect of harm crises through localization of their brand strategies, such as by carefully tailoring product offerings and communications to local markets? These interesting questions are becoming relevant in the context of product failures and harm crises for global brands.

Further, our results directionally show that highly ethnocentric emerging-market consumers appreciate postcrisis compensation more when the brand is local rather than global. Brand managers may need to consider their consumer's ethnocentrism and their brand's positioning when deciding on whether to compensate or not, to mitigate the decline in evaluations. Results concerning ethnocentric consumers may be more relevant in the current volatile global context. National protectionism (Liu et al. 2021) is increasing in a de-globalizing world in which China–West tensions, Brexit, climate crises, and a pandemic could increase the size of ethnocentric consumer segments while simultaneously encouraging even nonethnocentric consumers to appreciate local brands more. In such a context, global brands may be even more vulnerable in case of a product-harm crisis.

Limitations and Future Research Directions

The current work can be extended empirically and theoretically in several ways. Empirically, we needed to carefully control for confounding factors with our experimental approach, where internal validity is critical. Accordingly, the studies could not incorporate the manipulation or measurement of the full list of moderating and mediating influences simultaneously. Field studies, in which more variables can be tested and measured concurrently, could complement our experimental studies.

Theoretically, shifting standards regarding global versus local brands can be materialized through other mediators, in addition to consumer expectations. For example, consumers may perceive strong local brands as icons and symbols of their local culture (Özsomer 2012; Steenkamp, Batra, and Alden 2003), and as embodying a patriotic aura. When such local brands respond to a product-harm crisis and provide compensation, consumer pride may serve as a mediator. This mediation effect might be particularly strong for ethnocentric consumers, who have closer connections to their in-group. Second, it would be interesting to study how shifting standards in global versus local brand evaluations can be offset by brand personality. Would consumers be more forgiving of a harm crisis involving a warm (vs. competent) global brand relative to a competent (vs. warm) local brand? Could the personality of the global (local) brand moderate expectations, and hence consumer evaluations? Such insights could be particularly useful for global brand managers as they craft the positioning of their brands in different markets around the world. Third, exploring the persistence of the effects of shifting standards on other downstream consequences could provide another area for future research. In our work, we focused on consumer evaluations as our dependent variable of interest. What about purchase intentions and actual purchases? Could global (local) brand associations such as lower (higher) perceived risk (Erdem and Swait 1998) and higher (lower) credibility (Mandler, Bartsch, and Han 2021) offset the downstream consequences of shifting standards?

One way to mitigate the vulnerability of global brands in the context of harm crises could be to partner with local stakeholders such as nongovernmental organizations, universities, and government agencies for social responsibility and brand activism (Sarkar and Kotler 2021) in emerging markets. Such partnerships and joint projects could build, for the global brand, connections to the local community, culture, and traditions, and could make the decline in evaluations less severe, particularly for ethnocentric consumers. Another way to manage this vulnerability could be to strategically mix the global and local branding associations with a glocal brand positioning. A glocal brand is a well-balanced combination of global and local elements (Godey and Lai 2011; Schmidt-Devlin, Özsomer, and Newmeyer 2022), and benefits from an integration of the positive associations of both (Steenkamp 2019). When the categorization is less clear, as in the case of a glocal brand, shifting standards theory would suggest that the decline in consumer evaluations would be less severe after a crisis incident since the expectations are also lower. Thus, pursuing a glocal branding strategy not only builds closer connections with local customers (Ger 1999; Özsomer 2012) but also could decrease the adverse effects of a failure incident. We hope our study stimulates future research on failure incidents when brands are glocal. We believe the shifting standards theory provides a solid theoretical lens for testing such interesting pathways in the context of global and local brands.

Supplemental Material

sj-pdf-1-jig-10.1177_1069031X231222865 - Supplemental material for Shifting Standards in Consumer Evaluations of Global and Local Brands After Product-Harm Crises

Supplemental material, sj-pdf-1-jig-10.1177_1069031X231222865 for Shifting Standards in Consumer Evaluations of Global and Local Brands After Product-Harm Crises by Eda Sayin, Nilüfer Z. Aydınoğlu, Ayşegül Özsomer and Zeynep Gürhan-Canlı in Journal of International Marketing