A brand built on sand: Is acquiring a local brand in an emerging market an ill-advised strategy for foreign companies?

Abstract

The literature that focuses on acquisitions from the consumer perspective has generally neglected the brand strategy of cross-border acquisitions in an emerging market by a developed country brand. However, research in this field appears necessary, considering the high failure rate of M&As, the common practice of Western/global companies of augmenting their brand portfolio through local acquisitions, and the sensitivity of emerging market consumers to foreign brands. The present study is an initial attempt to understand the loyalty of consumers toward the acquired brands. Moreover, we investigate how such an acquisition affects the relationship between quality and loyalty, as well as between price and loyalty. For fast-moving consumer goods brands in China, the findings indicate that from a customer’s perspective acquiring a local brand is not an advisable strategy for foreign brand conglomerates, because such an international takeover may decrease consumer loyalty. Additionally, consumers tend to expect higher quality after the takeover but may not want to pay more for the quality increase.

Introduction

When the German consumer goods group Beiersdorf purchased 85% of the Chinese hair care company C-BONS and its brands for €317 million in 2007, the management was euphoric. Already successful in the Chinese skin care market with its own brands of Nivea and Eucerin, Beiersdorf figured it could take a shortcut to a strong position in the Chinese hair care market by acquiring several locally well-established brands. However, the once promising acquisition soured, due to heavy losses (Wilson 2013) that led to impairments for the newly bought China hair care business and even to the discontinuation of one of the local acquired brands (Beiersdorf 2012). The setback of Beiersdorf in China can be traced back to numerous company-specific problems (e.g., managerial, organizational), but as the 2011 Q3 interim report admits, the ultimate reason for this failed acquisition adventure was that sales simply did not reach the expected numbers (Beiersdorf 2011). This confession indicates that Beiersdorf misjudged consumer reactions toward the acquisition.

The failure and success of acquisitions have attracted considerable attention but have yielded contradictory findings in the literature. While some authors claim a common failure rate of 70-90% of acquisitions (Christensen et al. 2011) and a meta-analysis posits a negative abnormal return for the acquirer in the long run (King et al. 2004), other studies argue that there are positive abnormal returns (Datta et al. 1992; Moeller et al. 2004), especially in specific areas like cross-border acquisitions (Morck and Yeung 1992), for companies outside the US, UK, and Canada (Alexandridis et al. 2010), or brand acquisitions (Wiles et al. 2012). However, most studies dealing with acquisitions draw one common conclusion: variance remains unexplained, because important variables are unidentified in existing research (King et al. 2004). One variable that may shed more light on the prospects of acquisitions is customer reaction (Homburg and Bucerius 2005). This may be particularly influential in cross-border acquisitions in an emerging market by a developed country brand, because consumers there are especially sensitive to foreign stimuli (Sharma 2011). International takeovers in emerging markets are sometimes seen as “national brands falling into enemy hands” (Yu 2009, p. 44). Moreover, emerging markets are currently witnessing rising nationalism (Sheth 2011), which may make foreign acquisitions of local brands an even more delicate matter in the future.

The existing literature offers only limited general advice on cross-border acquisitions in emerging markets by a developed country brand and provides no specific advice on consumer reactions (Homburg and Bucerius 2005; Shimizu et al. 2004). Homburg and Bucerius wrote in 2005 that the “lack of attention given to marketing issues in the context of M&A is in sharp contrast with many statements that highlight the importance of marketing-related issues for M&A performance” (p. 95). This statement is still true today; while the merger and acquisitions (M&A) related marketing literature is slowly expanding, many open questions remain, and thus Wiles et al. (2012) justifiably lament that little is known on whether and how firms benefit from the acquisition of brands. Although Homburg and Bucerius (2005) show that market share and customer loyalty have a stronger impact on financial performance than cost savings, the acquisition literature dealing with the consumer remains limited (Öberg 2013). In fact, the only literature review, to the best of the authors’ knowledge, that deals with cross-border acquisitions (Shimizu et al. 2004) does not even contain the word “consumer.” Also, recent M&A studies considering consumer predispositions to foreign influence are rare. One recent study (Lee et al. 2011), for example, covers consumer reactions to cross-border M&As in emerging markets, but it overlooks the potential impact of reactance, nationalism, or ethnocentrism. However, research in this field seems necessary, considering both the sensitivity of developing country consumers to foreign brands (Tian and Dong 2011) and the significant investments of Western/global companies in enriching their brand portfolios through local acquisitions.

The failure of the M&A literature to consider consumer perceptions is imprudent, since consumers approve or disapprove of company actions by voting with their shopping baskets (Smith 1990). Such changes in customer loyalty after an acquisition determine the financial performance of companies (Homburg and Bucerius 2005). Additionally, companies need to convert quality improvements into price increases for the acquired brand (Clemente and Greenspan 1997). Therefore, as a first step to a better assessment of acquisitions of emerging market brands by companies originating from developed countries, this research offers insights into three research questions: First, what is the general level of loyalty of consumers to these acquired brands? Second, how does such an acquisition affect the relationship between perceived quality and loyalty? Third, how does such an acquisition affect the relationship between perceived price attractiveness and loyalty?

Our study makes two additional contributions. First, we introduce a new brand type, International Brand Portfolio Acquisitions (IBPA), which is distinct from brand types like local or foreign brands that are frequently discussed in the literature (e.g., Alden et al. 1999; Gao et al. 2006; Özsomer 2012). These IBPA brands are emerging market brands acquired by a developed country firm to enrich its brand portfolio. A brand portfolio in this context refers to all brands managed by a company (Aaker 2004). IBPA brands entail the following three premises: a foreign-dominated strategy imposed on an emerging market brand, a preserved heritage of the local brand, and the presence of the acquirer in the emerging market with purely foreign brands. The IBPA brands are particularly important because the number of M&As involving emerging market firms has increased; about one in four M&As already concerns emerging markets (BCG 2013). With this growing interest in emerging markets, many multinational consumer product companies employ M&As in order to “add popular local/regional products and brands to their portfolios” (Deloitte 2012, p. 3). These companies include well-known Western brands like Danone, L’Oreal, and Nestlé.

As a second contribution, we theoretically assess how consumers may react to the newly defined IBPA brand type. There are two conflicting theory streams in the scarce literature that deals with consumer responses toward brand acquisitions. On the one hand, the theory of psychological reactance (Brehm 1966; Thørbjornsen and Dahlén 2011) assumes that an acquired brand is not appreciated by consumers, since consumers might interpret the takeover as a threat to their freedom of choice and thus attempt to restore their freedom by devaluating the appeal of the forced alternative in their mind. On the other hand, signaling theory (Wernerfelt 1984, 1988; Swaminathan et al. 2008) proposes that consumers may view an acquisition of a local brand more favorably, since the international brand name of the acquirer acts as collateral for the quality promises of the acquired local brand. The foreign acquisition may then lead to spillover effects and raise, for example, the brand equity of the target (Lee et al. 2011). We address these conflicting views by using the country-of-origin paradigm to integrate them into one model, and by testing the reaction of consumers to international brand portfolio acquisitions.

In the following section, we provide a brief overview of the theories on which the present research is based. Building on these theories, we develop what constitutes an IBPA brand and then derive hypotheses relating to our research questions. We test the hypotheses with a hierarchical linear model that utilizes data from 36 real fast-moving consumer goods (FMCG) brands in China. After presenting the results of the empirical study, we discuss the findings and the implications.

Theoretical background

To develop the specific traits of IBPA brands, we compare them to other branded products. It is generally accepted in the literature that a branded product enjoys enhanced quality and value perceptions, as well as greater consumer loyalty, compared to an unbranded product (Dodds and Monroe 1985; Dodds et al. 1991; Grewal et al. 1998). Keller (1993) refers to this as the differential effect of brands. This effect varies across product and brand groups and exists beyond situational factors like short-term price promotions or recent advertising. It can be thought of as the arithmetic mean of all brands in a particular product category. Following Kamakura and Russell (1993), the general effect of branded products in the market is thus defined as the market-wide average brand effect for a particular product segment. This effect is a helpful benchmark to identify the idiosyncrasies of IBPA brands.

To provide as precise an analysis as possible, we limit our study to one product category only, FMCG brands. Fast-moving consumer goods can be defined as frequently purchased, low-involvement goods that are sold at relatively low cost (Nijssen 1999; Silayoi and Speece 2007; Cleeren et al. 2013), such as household products, food, alcoholic beverages, soft drinks, tobacco products, and personal care (Koschate-Fischer et al. 2014; Olsen et al. 2014).

The general effect of branded FMCG products in emerging markets is determined by three brand types: the IBPA brands, which are introduced in the present study, as well as local and foreign brands, which have long been established as major brand types in emerging markets in this product category (Batra et al. 2000). Given that local and foreign brands are major brand types and that the assessment of IBPA brands is likely to be influenced by the consumers’ assessment of local and foreign brands, we include them in our theoretical and empirical analysis.

To evaluate consumer reactions toward IPBA, local, and foreign brands, we adopt a threefold theoretical approach including the country-of-origin (COO) paradigm, the theory of psychological reactance, and signaling theory. The COO paradigm is useful for developing hypotheses concerning local and foreign brands and helps to explain why the two latter theoretical approaches shed light on consumer quality and price assessments, as well as on their loyalty intention toward IPBA brands.

The main ideas are as follows: the cognitive ratings of quality and price are key antecedents for loyalty (Oliver 1999). Since we are dealing with a cognitive process, we employ signaling theory to better understand the parameters of this evaluation. However, affective/normative processes may prevent the transfer of this cognitive evaluation to loyalty intentions (Verlegh and Steenkamp 1999). One important factor in a consumer’s assessment of an acquisition may be a perceived loss of freedom (Thørbjornsen and Dahlén 2011). We thus rely on the theory of psychological reactance to develop this affective or normative consumer disapproval of a foreign acquisition of their local brands.

The country-of-origin paradigm

The country of origin (COO) of a brand may have certain impacts on consumer loyalty intentions and the role of brand quality and price as drivers of loyalty. These impacts can best be analyzed by separating the country-of-origin effect into three main effects, cognitive (i.e., the COO cue leads to rational considerations), affective (i.e., the COO cue arouses an emotional reaction), and normative (i.e., the COO cue evokes moral reflections, built on social or self-imposed norms) (Johansson 1989; Obermiller and Spangenberg 1989; Verlegh and Steenkamp 1999). While these processes are interlinked, one of their key characteristics is their build-up toward purchase intention, which is rooted in the theory of reasoned action (Fishbein and Ajzen 1975). Cognitive processes primarily lead to changes in consumer beliefs, affective ones to changes in consumer attitudes, and normative ones to changes in behavioral intentions (Brijs et al. 2011). Chinese consumers, for example, could potentially evaluate the quality of Japanese cars positively but in fact dislike Japanese cars because of anti-Japanese sentiment, or they might simply not consider purchasing a Japanese car, because of personal or social norms.

For the application of this theoretical concept to price, quality, and loyalty, we follow established research (Dodds et al. 1991; Kirmani and Rao 2000; Luomala 2007) and treat consumers’ process of judging the price and quality of a product primarily as a cognitive evaluation. The influence of the COO cue on consumer beliefs concerning quality is the prime example of a cognitive COO cue in the literature (Verlegh and Steenkamp 1999). Moreover, the literature treats consumer considerations as to whether a low-involvement product has an attractive price, as a cognitive tradeoff between perceived quality and the monetary sacrifice (Dodds et al. 1991; Martins and Monroe 1994). We argue that COO-driven consumer expectations of the quality and price of brand types may influence the role of price and quality to induce loyalty. Additionally, the affective and normative COO processes of consumers are more directly related to loyalty intentions and do not lower the cognitive evaluations of price and quality (Obermiller and Spangenberg 1989; Wang et al. 2012).

In addition to the general COO paradigm, two theoretical concepts are important for understanding consumer reactions toward IBPA brands: signaling theory and the theory of psychological reactance. Signaling theory is based on cognitive considerations (Connelly et al. 2011) and thus helps to address price and quality expectations. Psychological reactance, on the other hand, is evoked by affective or normative reactions of consumers (Brehm 1966; Burnkrant and Cousineau 1975). Both theoretical approaches will be introduced successively below.

Signaling theory

With the prerequisite of a rational and risk-averse individual, signaling theory elucidates the idea that the receiver of a signal engages in the cognitive effort of interpreting information to alleviate uncertainty (Kirmani and Rao 2000). This uncertainty stems from information asymmetry concerning latent and unobservable quality, between, for example, a company (sender) and its customers (receivers) (Akerlof 1970). Both senders and receivers of signals have an interest in reducing information asymmetry, because it could lead to shirking and underinvestment from the company side and under-consumption from the customer side. The key to a credible signal is that it contains sunk costs for the sender, which cannot be recovered if the sender defaults on the signal’s promise. A brand, for example, may be seen as a signal for reducing consumer uncertainty concerning the quality of a product in a pre-purchase situation (Dawar and Parker 1994; Erdem et al. 2006). If the brand cannot live up to the quality promise, it loses its value. The firm is rewarded for the sunk costs of building the brand by a price premium it can charge for the brand (Shapiro 1983). Hence, the price of a product is positively related to its quality (Tellis and Wernerfelt 1987). Consumer willingness to pay a higher price for a product of unobservable quality can thus be interpreted as a feedback to the sender of a successful signal (Connelly et al. 2011).

Findings in the literature indicate that transfers of signals are possible, for example, to a new product when using umbrella branding (Wernerfelt 1988), when a product is sold through a retailer with a credible reputation (Chu and Chu 1994), or to brands in a brand alliance (Rao et al. 1999). This can be thought of as a spillover of the signal. The established signal of the sender then acts as a bond and reduces uncertainty concerning a new signal. This is possible, since poor quality of the new product/brand would transfer back to the established signal by reducing its value. The spillover would, therefore, enhance a consumer’s quality expectations regarding the new product and, according to the positive price-quality correlation (Tellis and Wernerfelt 1987), would also lead him/her to consider a higher price than before as still attractive. For such a spillover, it is essential that consumers perceive coherence, or fit, between the original brand and the new one (Pina et al. 2013).

The theory of psychological reactance

Signaling theory explains how consumers assess brands cognitively, how uncertainty concerning unobservable quality can be diminished, why price and quality are essentially related, and why a signal from an established brand results in higher quality and price expectations toward a new brand. However, signaling theory cannot explain how consumers respond when their personal freedom is reduced by the signaling action itself. Such a phenomenon has been indicated in the marketing literature for product scarcity, and psychological reactance has been made out as one possible consumer response (Ge et al. 2009). From a more abstract perspective, psychological reactance may thus be interpreted as a form of feedback to the sender, when the receiver is touched affectively/normatively by an unwanted signal.

The theory of psychological reactance (Brehm 1966) considers how individuals respond to the elimination, reduction, or threatened reduction of personal freedom. The consequence is psychological reactance, which is a state of motivational arousal directed to regaining the lost or threatened freedom. This endeavor results in increased attractiveness of the eliminated or threatened alternative and in a devaluation of the imposed alternative.

Psychological reactance is not necessarily conscious; it may also occur subconsciously and unintentionally (Chartrand et al. 2007). As a consequence, psychological reactance can best be deduced from attempts of consumers to reassert their freedom (Clee and Wicklund 1980). A decrease in loyalty intentions is, for example, one established way in the literature for consumers to reassert their freedom (Algesheimer et al. 2005). Notably, psychological reactance may also be generated if the choice object that is threatened/eliminated is relatively low in terms of preference (Clee and Wicklund 1980). However, the magnitude of reactance depends on the importance of the freedom, so that the threatened freedom has to be relevant to important needs of the individual to induce reactance (Brehm 1966). Consumers have been shown to display reactance in a number of circumstances (Clee and Wicklund 1980), for example, in cases of product unavailability (Fitzsimons 2000), induced marketing promotions (Kivets 2005), and dominant M&As (Thørbjornsen and Dahlén 2011). Moreover, psychological reactance has also been suggested against two companies’ co-branding efforts (Hillyer and Tikoo 1995).

Summing up, product scarcity, acquisitions, and brand alliances have all been described from both a signaling and a psychological reactance perspective, but no link between both theory streams has been established. Using IBPA brands, we consider in the following hypotheses how an acquisition of an emerging market brand by a foreign company may trigger two opposing but connected processes. First, a foreign takeover may send a signal that raises the quality and price expectations of consumers. Second, the M&A may cause psychological reactance in the form of decreased loyalty, because the signaling action of the acquisition is at the same time connected to a reduction of consumers’ perceived freedom.

Hypothesis development

We first outline how IBPA brands are distinct from other brands. Subsequently, we depict a simple model that is applicable to brands in general. We then discuss in detail why the effects of local, foreign, and IBPA brands are different from this general brand effect.

As the initial Beiersdorf example indicates, it is frequently observed that foreign companies acquire local brands in emerging markets. Danone, L’Oreal, and Nestlé have undertaken just such endeavors. To ensure a meaningful analysis of consumer reactions toward IBPA brands, we examined these practical examples and then, building on signaling theory and psychological reactance, distilled commonalities among them. The commonalities of the IBPA brands relate to the key moderators of signaling theory and psychological reactance, which are signal credibility (Connelly et al. 2011) and a perceived threat to one’s freedom (Brehm 1966).

There are three important IBPA brand commonalities. First, the strategy of the emerging market brand has to be foreign-dominated to qualify for the IBPA category. While ownership structures after an acquisition may be diverse, it is important that the controlling stake of the brand lies with a foreign company. According to signaling theory, the higher the investment of the foreign acquirer, the higher are its sunk costs. Moreover, a foreign majority enables the new foreign owner more effectively to initiate restructurings that may improve an IBPA product’s quality. Both points enhance credibility of the foreign signal for consumers (Connelly et al. 2011). Therefore, a small foreign stake in an emerging market company would not qualify the brands to belong to the IBPA type.

For the second premise, IBPA brands require a preserved local brand heritage. In general, it is more common for acquisitions to change the name and/or symbol of the acquired company (Ettenson and Knowles 2006) in order to be associated more readily with the (stronger) acquirer. Nevertheless, we require from an IBPA brand that the changes to name or symbol are only marginal, so that the local heritage of the acquired brand remains recognizable for the consumer. From a psychological reactance perspective, it is important to ensure that consumers are reminded of the local origin and continuous history of the brand, so that they may feel that their freedom to choose a traditional local brand without foreign influence is limited. Otherwise, consumers may see the acquired brand as a new one that merely builds on the old infrastructure. Moreover, termination of the old name or symbol may elicit different consumer reactions to the acquisition itself, which might dilute the results for IBPA brands.

Finally, for brands to qualify as IBPA, the foreign company should be active in the same market with purely foreign brands. This is important for the signal spillover and the psychological reactance effect. If a foreign company possesses expertise in a similar area, the transfer of quality from the foreign company to IBPA brands is more likely. Moreover, when foreign firms are active in the same market with purely foreign brands, the credibility of the signal rises, because it is easier for consumers to “punish” the foreign firm if IBPA brands do not fulfil the higher quality promise. From a signaling perspective, a foreign presence in the same market would also enhance the coherence, or fit, between the foreign and the IBPA brand. Coherence has been described as an essential parameter for successful signal spillover (Pina et al. 2013). Additionally, the presence of purely foreign brands could also increase consumer knowledge of the foreign company and in turn lead to a greater awareness of the foreign ownership of the IBPA brand. This again may raise consumer perceptions that their freedom to choose local brands is curtailed and thus lead to higher psychological reactance. We argue that brands which fulfill these premises create a new brand type that is distinct from the general effect of branded products in the market, as well as different from existing brand types, like foreign or local brands, in terms of consumer perceptions.

Because IBPA brands are a type that has not been described in the literature, we start with a very basic marketing model to analyze key differences between IBPA brands and the general effect of branded goods in the market (see Fig. 1). Attractive price and attractive brand quality are thus regarded as the key drivers of the marketing mix for consumer loyalty. It is well established that perceptions of quality and price as attractive are a precondition for loyalty intentions of consumers. Since they are often seen as the principal drivers, the literature confirming these links is abundant (see Parasuraman and Grewal 2000; Zhou and Wong 2008). Therefore, we do not hypothesize the effects of quality and price on loyalty. The relationship between price and quality is reciprocal (Erickson and Johansson 1985), so that we refrain from considering a direct path between both constructs and model a correlation instead.

Fig. 1

Conceptual model

Effects on brand loyalty

It has been reported that brand loyalty varies significantly according to the country of origin of the brand within a certain product category (Pappu et al. 2006). Cognitive, affective, and normative processes may be at play, but affective or normative cues are more directly related to loyalty intentions, compared to cognitive cues (Obermiller and Spangenberg 1989). The country-of-origin cue is especially influential for the FMCG product category, because this cue tends to be more important when consumers are less involved with a product (Josiassen et al. 2008).

For developed countries, research has described domestic products benefiting from home country bias (Verlegh and Steenkamp 1999). This bias is rooted in cognitive (economic protectionism) and affective/normative (in-group favoritism) processes (Verlegh 2007). There is no reason to believe that these motives underlying the home country bias are constrained to developed countries (Sharma 2011; Klein et al. 1998). On the contrary, the delayed development of emerging markets may lead to a latent minority complex that stimulates in-group favoritism. Also, the catching-up growth story of emerging markets is often associated with protectionism. Both effects would thus support a home country bias.

Brand-building capabilities of emerging market firms improve with the economic development of the country. Thus, it is reasonable to believe that the appeal of local brands also improves along with the development of an emerging economy (Chan et al. 2009). Such a rise in appeal may additionally increase the affective and normative arguments for consumers who speak in favor of local brands. This is supported by initial evidence of a preference for local products in the FMCG category (Alden et al. 1999; Bain & Company and Kantar Worldpanel 2014; BCG 2008; Özsomer 2012).

Additionally, consumers associate local brands as a “companion” who has been there all along (Schuiling and Kapferer 2004). This leads to a higher awareness level, which creates a loyalty advantage (BCG 2008). Most importantly, for FMCG products, emerging market consumers expect local brands to understand special consumer tastes or needs best, thus entailing the advantage of greater psychological proximity (BCG 2008).

Some studies have argued that there is preference for foreign brands in emerging markets. Such a preference by emerging market consumers is likely to be shaped by three factors: categories that are perceived as foreign (e.g., Western fast food), luxuries, and public consumption (Bain & Company and Kantar Worldpanel 2014; Okechuku and Onyemah 1999; Özsomer and Altaras 2008). For such product categories, foreign brands may be preferred by consumers for status reasons (Batra et al. 2000), or to demonstrate belonging to some kind of global community (Özsomer and Altaras 2008). However, most FMCG products are not subject to the factors benefitting foreignness (Bain & Company and Kantar Worldpanel 2014). Therefore, we expect an enhanced effect for local brands, but not for foreign brands and hypothesize:

• H1a: Local brands commonly possess a higher level of loyalty compared to the general effect of branded products in the market.

• H1b: Foreign brands commonly possess no statistical difference in the level of loyalty compared to the general effect of branded products in the market.

When it comes to IBPA brands, we expect local brands to forfeit their loyalty premium after being acquired by a foreign conglomerate. There may be a decline in loyalty levels, because (1) consumers may feel neglected, (2) local brands may simply be perceived as foreign, and (3) consumers might display reactance due to reduced freedom. The possible effects and their implications will be discussed subsequently.

A decline in consumer loyalty could be induced because consumers experience a feeling of neglect after an acquisition. Internal issues usually absorb considerable managerial energy immediately after the acquisition (Hitt et al. 1990). In this period, the firm may temporarily lose its customer focus. However, if the reduced loyalty is merely due to post-acquisition marketing and management problems, we would expect the decreased loyalty to prevail only for a limited time after the acquisition.

Another argument could explain a reduction in consumer loyalty, namely that IBPA brands are simply seen as foreign. As laid out in the development of H1a and H1b, local brands are expected to enjoy higher levels of loyalty compared to the general effect of branded products in the market, while foreign brands are not expected to have such a premium. This is a valid argument, since a brand is foreign to a developing country consumer, if it is not purely local in terms of conception, labor, ownership of patents, trademarks, and production facilities (Tian and Dong 2011). If this were the sole reason for a decline in loyalty levels, the common level of loyalty for IBPA brands should be situated somewhere between foreign and local brands, but it should not drop below the level of the general effect of branded products in the market. This would correspond with the argument of Funk et al. (2010), who demonstrate that purchase intentions decrease for products which are manufactured partly in countries the consumers hold animosity against. Funk et al. (2010) argue that this effect would develop similarly to the effect witnessed if these products wholly originate from these countries.

Psychological reactance may be a third reason for a decline in consumer loyalty (Algesheimer et al. 2005). Consumers have a preference for maintaining an activity to which they have grown accustomed (Samuelson and Zeckhauser 1988) and dislike a forced alteration of the status quo (Thørbjornsen and Dahlén 2011). The theory of psychological reactance (Brehm 1966) explains this behavior and assumes that consumers are inclined to preserve and restore their personal freedom. Fong et al. (2013), for example, find low repurchase intentions toward a post-acquisition target if consumers hold animosities against the acquirer’s country of origin. Similarly, Papavasileiou et al. (2008) report less identification of consumers with an acquired company if they perceive the acquisition of a socially responsible company as a sellout to a dominant acquirer that seeks to polish its image. For acquirer-dominant M&As, Thørbjornsen and Dahlén (2011) show that consumers develop negative attitudes toward the acquiring brand and intentions to switch from the target brands. As outlined in the above theory section, reactance may be evoked, even if the choice object which is threatened/eliminated is relatively low in preference, as long as the threatened freedom is important to the individual (Clee and Wicklund 1980; Brehm 1966). Applied to IBPA brands, we may thus expect reactance, even if the IBPA brand was not the top choice of the consumer before the acquisition, as long as it is important for consumers that local FMCG brands stay local. We have already indicated a rising nationalism in emerging markets that could trigger such a preference of emerging market consumers.

The effects are well illustrated by the attempt of Coca-Cola to acquire Chinese juice producer Huiyuan in 2009, which was eventually blocked by the Chinese government. Coca-Cola refused to agree to relinquish the local brand after the acquisition and, in turn, Beijing was worried “about public opposition to a foreign company taking over a leading brand” (Tucker et al. 2009). An online poll by a leading Chinese web site reveals the dimensions of public discontent with the attempt of Coca-Cola to add Huiyuan as an IBPA brand to its portfolio: Over 120,000 respondents participated, and more than 80% strongly agreed with the rejection by the Chinese government. In addition, over two-thirds were of the opinion that foreign investment in Chinese firms damages domestic brands (Tucker and Anderlini 2009).

Summing up, psychological reactance is the most compelling reason for a decline in consumer loyalty. Compared to consumer neglect after an acquisition, psychological reactance could be similar to animosity and linger for decades (Klein et al. 1998), since the local brand heritage associated with the brand name would remind consumers of the foreign acquisition. Compared to the foreign perception argument, psychological reactance can also explain why consumer loyalty would sink even below the general level of branded products in the market. Therefore, we hypothesize:

• H1c: IBPA brands commonly possess a lower level of loyalty compared to the general effect of branded products in the market.

Effects on the quality–loyalty path

According to the COO paradigm, normative and affective processes may bypass the cognitive process and have an immediate effect on consumer attitudes and behavior (Obermiller and Spangenberg 1989; Verlegh and Steenkamp 1999). For developing country consumers, this is best demonstrated in the animosity literature, which argues that consumers may have unbiased judgments of a product’s quality (cognitive) but display animosity toward the country of origin (normative) and thus refuse purchase (Klein et al. 1998). We similarly argue that consumers may well see a potential cognitive benefit from a foreign acquisition, e.g., quality enhancement, but still display low levels of general loyalty, because their personal decision freedom is curtailed. Therefore, the mechanisms determining the strength of perceived quality and a perceived attractive price as drivers of loyalty (cognitive) are independent of the reactance that determines the general level of loyalty (affective/normative).

The country of origin of a brand has been shown to act as a signal of unobservable quality (Han 1989). Foreign brands have been identified as more credible signals than local brands in emerging markets (Zhou et al. 2010). A reason could be that it is more costly to build a foreign or global brand, because it has to be present in multiple countries. These investments thus raise the signaling costs, which are directly related to signal credibility (Connelly et al. 2011). Therefore, the consumers’ quality expectations of foreign brands are higher than for local brands in emerging markets (Batra et al. 2000; Kinra 2006). Additionally, foreign brands that have expanded to an emerging market are likely to possess higher brand equity than local brands. Foreign brands usually have a longer history and are necessarily present in countries beyond the emerging market, while local brands have had less time to develop their brand and their international activity. Because brand equity signals product quality (Erdem and Swait 1998; Kirmani and Rao 2000), the quality claim may be more credible for foreign brands. For both these reasons, a higher level of quality is expected from a foreign brand. This makes it much harder to cross a satisfaction threshold which can induce loyalty (Oliver 1999). Therefore, a certain increase in quality for foreign brands is likely to induce a proportionately lower increase in loyalty, compared to the general effect of brands in the market. In other words, the responsiveness of loyalty to an increase in quality perception is relatively lower for foreign brands. On the other hand, for local brands, quality expectations are not that high, so that the same increase in quality would result in a higher increase in loyalty, compared to the general effect of brands. Therefore, we propose:

• H2a: The loyalty-enhancing effect of quality is stronger for local brands compared to the general effect of branded products in the market.

• H2b: The loyalty-enhancing effect of quality is weaker for foreign brands compared to the general effect of branded products in the market.

Clear and credible brand signals may increase quality perceptions and decrease perceived risk, both of which increase consumer utility (Erdem and Swait 1998). Signaling theory has already been extended to brand allies, where brands in an alliance influence consumer attitudes toward the alliance (Simonin and Ruth 1998; Rao et al. 1999). Also, due to the presence of foreign brands in the same market, to emerging market consumers, there is high perceived coherence between IBPA and foreign brands (Pina et al. 2013), which facilitates signal spillover. Thus, an application of signaling theory to IBPA brands is straightforward; the foreign company’s brand name acts as a bond (Wernerfelt 1988). In other words, if the foreign brand company sends a false quality signal (claiming good quality, while in fact it is disappointing), it would put itself in a worse position than if it had not sent a signal at all. IBPA brands thus profit from the foreign corporate brand name acting as collateral, which leads to higher quality expectations from IBPA brands, compared to local brands. According to the above logic, the expected quality rises and thus the strength of quality as a driver of loyalty declines for IBPA brands.

However, we argue that IBPA brands do not reach the level of quality expectations of foreign brands, but lie somewhere in between local and foreign brands. The reason is that there are two signals at work in IBPA brands. First, the former local brand signal that is not strongly associated with high quality and second, the foreign brand signal that is strongly associated with high quality. Thus, a brand acquisition creates a mix of the local and the foreign brand signal, and therefore, would be associated with only moderate quality. Accordingly, foreign companies acquire local brands, in order to position them differently, compared to their foreign brands (Douglas et al. 2001). For emerging markets, this would imply a lower quality positioning, compared to the foreign brand. Such a spillover effect of brands after an acquisition has already been indicated in the literature (Lee et al. 2011; Wiles et al. 2012). Therefore, we hypothesize:

• H2c: The loyalty-enhancing effect of quality does not differ significantly between IBPA brands and the general effect of branded products in the market.

Effects on the price–loyalty path

According to signaling theory, the sunk costs of building a credible brand signal can be converted into a price premium (Shapiro 1983). This theoretical notion is confirmed empirically by a positive relationship between the quality of a product and its price (Tellis and Wernerfelt 1987). Consumers are well aware of this tradeoff between what they “get” (quality) and what they “give” (price) (Zeithaml 1988). Such a weighing is a rather rational sequence and thus belongs to the cognitive process in the COO paradigm (Dodds et al. 1991).

Foreign brands, due to their relatively higher investments, signal high product quality. As foreign brands are associated with higher quality, higher brand equity, and prestige, consumers expect to pay for these extra benefits (Zhou and Hui 2003; Zhou et al. 2010). Accordingly, a higher quality signal from foreign brands is associated with higher price expectations by consumers. This can turn a lower price into a pleasant surprise, so that a lower than expected price for the same quality and prestige may increase consumer loyalty. Therefore, we expect a more attractive price to increase loyalty more for foreign brands than the average effects of brands in the market.

Local firms in emerging markets are associated with lower quality signals, due to their lower sunk costs for maintaining quality and building a good reputation. However, the price premium they can charge for the brand is also smaller. Hence, they aim to deliver acceptable quality at lower costs in order to remain competitive (Ger 1999). Accordingly, the lower quality signal of local brands is related to lower price expectations from consumers. Evidence from the literature supports this argument; for example, Sharma (2011) finds that value-conscious consumers have more positive associations with products imported from emerging countries than for those imported from developed markets. Therefore, contrary to the quality expectations, consumers in emerging markets have relatively high value-for-money expectations of local brands (Dmitrovic et al. 2009). Also, Batra et al. (2000) and Kinra (2006) argue that a lower price is anticipated from local brands. As a result, the loyalty elasticity for attractive prices of local brands is expected to be lower than the one of the general effect of branded products. In other words, the price–loyalty path is expected to show the exact opposite effects for local and foreign brands than the quality–loyalty path. We thus propose the following hypotheses:

• H3a: The loyalty-enhancing effect of an attractive price is significantly weaker for local brands compared to the general effect of branded products in the market.

• H3b: The loyalty-enhancing effect of an attractive price is significantly stronger for foreign brands compared to the general effect of branded products in the market.

As laid out above, consumers have higher quality expectations of IBPA brands after their acquisition. To fulfill these expectations, investments in actual quality increases must follow. From a revenue perspective, these investments then may be recouped by the company, if consumers purchase an IBPA brand more frequently, or pay more for the brand. Due to psychological reactance, however, a quality improvement dividend in the form of increased consumer loyalty is very questionable. Thus, a price increase for IBPA brands after an acquisition is likely.

This is also in line with the positive quality–price relationship inherent in signaling theory (Shapiro 1983; Tellis and Wernerfelt 1987); the high quality signal of foreign ownership is attenuated by the low quality signal of the local past, resulting in only moderate quality expectations. These expectations are then matched by expectations of a moderate price increase for IPBA brands after the acquisition. This seems reasonable, since it has been shown that consumer willingness to pay is higher for higher quality brands. For example, Blattberg and Wisniewski (1989) show that price promotions provide greater returns for high-quality brands (compared to low quality brands), and Sivakumar and Raj (1997) demonstrate that high-quality brands suffer less demand decreases from price increases. Following the above logic, we argue that IBPA brands will lose the low-price expectations associated with local brands. Thus, an attractive price becomes a stronger driver for loyalty intentions for IBPA brands after an acquisition. However, we do not expect an attractive price of IBPA brands to be as conducive as an attractive price of foreign brands because the price point of IBPA brands is likely to be lower than that of foreign brands (Douglas et al. 2001). Therefore, we hypothesize:

• H3c: The loyalty-enhancing effect of an attractive price does not differ significantly between IBPA brands and the general effect of branded products in the market.

Empirical study

China was selected as an appropriate location for our study. It is the largest emerging market, both in terms of population and growth momentum, and has attracted the attention of practitioners and scholars alike (Kumar and Steenkamp 2013). More importantly, Chinese consumers are known for their fascination with foreign products, as well as for their nationalism and cultural pride (Ewing et al. 2002; Tian and Dong 2011). They are thus open to both foreign and local brands, which is a decisive trait for emerging market consumers (Özsomer 2012). Moreover, Chinese consumers’ attitudes toward local and global brands are relatively representative, compared with other countries (Steenkamp and de Jong 2010). All these issues make China ideal for our IBPA brand investigation.

Multilevel models require a certain number of level-two units to ensure the accuracy of parameter estimates. Maas and Hox (2005) and Ozkaya et al. (2013) recommend no less than 30. In order to identify a sufficient number of relevant brands, hypotheses were tested using FMCG brands in four different categories: shampoo, facial cream, toothpaste, and mineral water (9 brands per category). FMCG products have been widely used for consumer loyalty studies (Rundle-Thile and Bennet 2001). The 36 brands also represent local, foreign, and IBPA brands. In emerging markets, local and foreign brands determine the main brand categories in the market (Batra et al. 2000; Lannes and Booker 2013). Other established categories from Western countries, such as private label brands, are negligible (Euromonitor International 2014). According to the respective average market share of local and foreign brands of the selected four brand categories in China (Bain & Company and Kantar Worldpanel 2014), we included a similar proportion of local and foreign brands. Together with IBPA brands, they are used to calculate the general effect of branded products in the market. The brands were selected with an emphasis on wide availability and consumer familiarity, using desk research and focus group interviews held in five Chinese cities. We finally narrowed the sample to three cities, where all 36 brands are well established. A second pre-test in all three cities (n = 30) validated consumer familiarity with the pre-selected 36 brands.

The cities include two first tier cities and one second tier city from the northern, central, and southern parts of China, to make our findings representative across regional differences. Beijing and Shanghai belong to the tier one cities of China, which classify the most developed cities, whereas Chengdu was selected to represent the central region and the tier two cities, which are regional economic powerhouses (Woetzel 2004; Virasami 2013). In each city, households were randomly selected from the inhabitant list, which was provided by the Chinese registration office in the three cities. We restricted the list to the urban population aged between 18 and 45. The reason is that we wanted to avoid distortion effects due to income differences, since China has a relatively young retirement age and older consumers tend to have benefited less and suffered more from economic reforms. Additionally, some older Chinese may display differences in shopping behavior and, for example, still shop at wet markets instead of modern shopping formats (Gamble 2011). We later tested whether the respondents are homogeneous in their behavior in terms of age, and found that age has virtually no effect on the results. Through a specific counting procedure, households were selected first by district, then block and finally streets in each city. Trained interviewers from a local research agency were provided with this list of households to conduct face-to-face interviews at respondents’ homes, based on a standardized questionnaire. The interviewers were trained by one of the co-authors over a one-day period in each city. Only respondents who were at least rather familiar, familiar, or very familiar with the respective brand were included in our sample. Each respondent was only questioned about one brand, in order to limit fatigue (Zhou et al. 2010). Altogether, the sample comprised 1188 valid questionnaires (generally 11 per city, for each of the 36 FMCG product brands) (Table 1).

Table 1 Demographic profile of study respondents

Measurement

The three main constructs of our model are as follows: loyalty measures loyalty intentions (conative loyalty) (Oliver 1999), whereas quality and price measure how the consumer evaluates the quality and the price, respectively. The measurement scales were adapted from established research (Table 2), and we used a seven-point Likert scale where 7 equals “strongly agree.” We tested our scales rigorously for validity, reliability, and possible bias.

Table 2 Measurement

Using our pre-tests, we assessed face validity. Our standardized factor loadings for the three constructs are, without exception, above 0.7 (Hair et al. 2010). Cronbach’s alpha coefficients range between 0.824 and 0.889, which is well above the recommended threshold of 0.6 (Bagozzi and Yi 1988). As all CFA factor loadings were above 0.7, and the average variance extracted (AVE) easily exceeded the threshold of 0.5, so convergent validity was supported (Bagozzi and Yi 1988). The correlation matrix and AVE in Table 3 indicate discriminant validity and nomological validity of the measures, according to the Fornell Larcker criterion (Fornell and Larcker 1981).

Table 3 Correlation matrix

In order to ensure idiomatic equivalence, we used the translation-back-translation approach (Hult et al. 2008). First, the original scales were translated into Chinese by a bilingual market researcher. The scales were then translated back into English by a bilingual graduate student. We then compared the original English version and the back-translated version, and corrected the Chinese version. The process was repeated until the two English versions corresponded with one another. Additionally, we tested the scales for comprehension in the pretest. We used an appropriate questionnaire design to minimize common method effects from the beginning (MacKenzie and Podsakoff 2012).

After the data collection, we conducted Harman’s single factor test by running a confirmatory factor analysis, including all factors, and restricting the factor loadings to equality, while setting the variance of the common factor to one. We reached a common method variance of 0.217. This value does not explain the majority of covariance among the measures, so that we do not see common method testing as biasing our results (Podsakoff et al. 2003). To reduce the model’s complexity, we modeled each construct using a weighted single indicator, after taking the satisfactory measurement model into account (Bandalos 2002) (CFI = 0.969; TLI = 0.960; χ ² (81) = 335.347).

To provide a more rigorous test for our model, covariates were taken into account to control for exogenous variables. At the individual level, we identified five. Since people with a local identity may be more loyal to or keener to learn about a brand with local origin, we included local identity. This variable controls deviations induced by people who strongly identify with their “own local community” (Zhang and Khare 2009, p. 524) and might thus display stronger latent nationalism. Brand familiarity could lead to differences in information processing and brand evaluation (Simonin and Ruth 1998) and could therefore influence loyalty intentions. The items used to measure local identity and brand familiarity are also listed in Table 2. The measurement for local identity yields two weak factor loadings and the construct thus only reaches an average variance extracted of 0.395 (Table 3). However, we retained the items, because of high construct reliability (Cronbach’s alpha > 0.70) and semantic proximity between the construct and the items.

For differences between the three brand types in terms of a higher rate for correct country of origin identification (Samiee et al. 2005), Correct identification was included. Coded as 1 for correctly identified and 0 for falsely identified origins of the brand’s holding company, correct identification controls for the effects on loyalty, if a brand type’s country of origin should be easier to identify than the others. Gender (coded as 1 for male, 0 for female) was included, since female consumers tend to rate foreign products more favorably (Wall and Heslop 1986). However, it has also been argued that there is no reason to believe that reactance effects differ by gender (Brehm and Brehm 1981). To account for possible differences in macroeconomic development levels, we controlled for first and second tier cities in China (tier one cities coded as 0, tier two cities as 1). On the brand level, we included possible differences between the product categories of personal care products (toothpaste, facial cream and shampoo) (coded as 0) and bottled water (coded as 1).

Before developing the model, we analyzed our data to search for possible bias. One bias we would like to address is population heterogeneity (Lubke and Muthén 2005), which could lead to systematic underestimation of path and regression coefficients (Shugan 2006). To consider population heterogeneity, we use a method originally established by Steenkamp and Baumgartner (1998) to test for measurement invariance. Measurement invariance exists for different levels and according to Steenkamp and Baumgartner (1998), partial factor covariance invariance implies that factor correlations across subpopulations are invariant. In other words, neither measurement nor factor correlation differences between possible subgroups should then distort the model results. More precisely, we wanted to ensure measurement invariance between the three brand types, local brands, IBPA brands and foreign brands. We followed the stepwise procedure of Steenkamp and Baumgartner (1998) and compared whether the fit of an unconstrained measurement model is significantly better than increasingly constrained multi-group models. We found configural invariance, full metric invariance (χ2 difference of 0.262), full scalar invariance (χ2 difference of 0.263) and partial factor covariance invariance (χ2 difference of 0.255). In accordance with Steenkamp and Baumgartner (1998), we therefore conclude that differences between the brand group’s measurement do not bias our results.

In addition to our level-two moderators, different product categories may be associated with specific effects. Consumers might, for example, perceive our four product groups as having different levels of risk. In particular, bottled mineral water, as a food product, might be perceived as more risky than non-food products such as shampoo, facial cream, or toothpaste. Therefore, we also undertook a similar procedure to that used to assess the brand groups, in order to find out whether respondents followed this line of reasoning about the risk associated with different product groups. Configural invariance, full metric invariance (χ2 difference of 0.069), full scalar invariance (χ2 difference of 0.130) and full factor covariance invariance (χ2 difference of 0.060) could all be established. Thus, we conclude that differences between the riskiness of our various product categories do not bias our results.

Furthermore, it is possible for familiarity with the brand to create a response bias. Participants who have identified the region of origin of the brand correctly might thus differ in their responses from those who attributed the brands to an incorrect region, for example, thinking that local brands are foreign. However, our results establish configural invariance, partial metric invariance (χ2 difference of 0.126), partial scalar invariance (χ2 difference of 0.767) and partial factor covariance invariance (χ2 difference of 0.671) between those groups (Table 4).

Table 4 Measurement invariance

Before introducing the model, we also wanted to determine whether consumers are actually able to recognize the IBPA brands as such. This is an important factor for the hypothesized reactance, as well as for the signal spillover. After responding to the questionnaire items, we asked respondents to identify the region of origin of the company owning the brands as “foreign,” “domestic,” or “don’t know.” From the 370 respondents in our survey who answered a questionnaire belonging to an IBPA brand, 67.6% thought the brand was foreign owned (28.6% opted for the domestic ownership, 3.8% stated “don’t know”). This is significantly higher than the local brand (7.4% of respondents chose foreign company ownership, 91.3% opted for domestic ownership). For the respondents answering with respect to the foreign brand, 78.6% identified the company behind the brand as foreign and 16.0% as domestic. The recognition rate of the origin of the owner might be perceived as unusually high. This may be explained by the familiarity of consumers with the brand in question and the high sensitivity of foreign brand ownership in emerging countries, as laid out in the hypothesis development section. The high recognition rate of foreign ownership of IBPA brands thus supports our reasoning. To eliminate further bias due to the recognition of IBPA brands, as indicated above, we checked for measurement invariance between those respondents identifying the origin of the brand-owner correctly, and also included a control variable labeled correct identification. Table 5 shows that the high recognition rate is evenly distributed among brands. One IBPA brand, Zhonghua, diverges from this pattern with only a third of respondents identifying this brand as having a foreign owner. The reason might be that this brand is merely licensed and was not actually acquired. We leave this brand in the IBPA section, because our reasoning, both for psychological reactance and a quality upgrade, might also apply to licensing. We also checked the model, both including and excluding the brand, and found no difference in results.

Table 5 IBPA brands validity assessment

In the hypothesis development section, we argued that a reduced effect of loyalty for IBPA brands even below that of foreign brands might be due to two reasons: psychological reactance and post-acquisition managerial problems. We based H1c on reactance and thus needed to control for the effect of managerial problems. Post-acquisition integration problems are generally supposed to be solvable over time, while we expect the effect of psychological reactance to linger on for years. To focus on effects due to psychological reactance, we selected IBPA brands for which acquisitions dated back no less than 3 years at the time of study. Moreover, we ran an SEM model of the individual level data, only including the IBPA brands. The additional control variable time of acquisition did not have a significant impact on the results.

Multilevel model specification

We tested the hypotheses with the software MPlus (Muthén and Muthén 2010), which can analyze hierarchical linear models. We chose the method in accordance with our sampling procedure with non-independent observations at the individual level. We thereby avoid an underestimation of the standard errors of individual parameters and decrease the probability of Type 1 errors (i.e., a false rejection of the null hypothesis) (Heck and Thomas 2009). The proposed model has two levels; the individual level distinguishes between individual customers (n = 1188), and the brand level differentiates one brand from another (n = 36). By capturing random slopes (coefficients) and random intercepts (means) for the variable brand type, we simultaneously control for two sources of variance. The level-one equation for brand loyalty intention is:

$$ {\mathrm{L}}_{\mathrm{ij}} = {\hat{\mathrm{a}}}_{0\mathrm{j}} + {\hat{\mathrm{a}}}_{1\mathrm{j}}\left({\mathrm{Q}}_{\mathrm{ij}}\right) + {\hat{\mathrm{a}}}_{2\mathrm{j}}\left({\mathrm{P}}_{\mathrm{ij}}\right)+{\hat{\mathrm{a}}}_{\mathrm{controls}} \ast {\mathrm{FControls}}_{\mathrm{ij}}+{\mathrm{r}}_{\mathrm{ij}} $$

(1)

Where i represents individuals and j brands, L_ij denotes individual i’s loyalty intention toward brand j. Q_ij reflects individual i’s perceived quality of brand j. P_ij represents individual i’s price perceptions of brand j. FControls include individual-level control variables, such as correct brand identification, familiarity, local identity, gender and city. â_0j stands for the intercept and â_1j and â_2j for the regression slopes that are allowed to vary across brands. r_ij denotes the error term at the individual level. Following the suggestion of Raudenbush and Bryk (2002), the independent variables were group–mean centered for a better numerical stability and to avoid model misspecification. Additionally, we checked the variance inflation factor (VIF = 2.75), which was below the cutoff point of 10 suggested by Hair et al. (2010). Therefore multicollinearity should not be an issue for the model.

At the brand level, effect coding was applied to differentiate the effects between the three brand types (Cohen et al. 2003). The effects of the individual brand types were measured against the overall mean, which is intended to resemble the general effect of branded products in the market. Dummy variable D¹ was assigned with the values of local brand = 1, IBPA brand = −1, foreign brand = 0, and D² was constructed as local brand = 0, IBPA brand = −1, foreign brand = 1. While the Mplus software automatically comes up with estimates and p-values for differences of local and foreign brands against a global average, we used the model constraint function to find the corresponding figures for IBPA brands. The level-two models were specified as follows:

$$ {\hat{\mathrm{a}}}_{0\mathrm{j}} = {\tilde{\mathrm{a}}}_{00} + {\tilde{\mathrm{a}}}_{01}\left({\mathrm{D}}_{\mathrm{j}}^1\right) + {\tilde{\mathrm{a}}}_{02}\left({\mathrm{D}}_{\mathrm{j}}^2\right) + {\mathrm{u}}_{0\mathrm{j}} $$

(2)

$$ {\hat{\mathrm{a}}}_{1\mathrm{j}} = {\tilde{\mathrm{a}}}_{10} + {\tilde{\mathrm{a}}}_{11}\left({\mathrm{D}}_{\mathrm{j}}^1\right) + {\tilde{\mathrm{a}}}_{12}\left({\mathrm{D}}_{\mathrm{j}}^2\right) + {\mathrm{u}}_{1\mathrm{j}} $$

(3)

$$ {\hat{\mathrm{a}}}_{2\mathrm{j}} = {\tilde{\mathrm{a}}}_{20} + {\tilde{\mathrm{a}}}_{21}\left({\mathrm{D}}_{\mathrm{j}}^1\right) + {\tilde{\mathrm{a}}}_{22}\left({\mathrm{D}}_{\mathrm{j}}^2\right) + {\mathrm{u}}_{2\mathrm{j}} $$

(4)

The error terms u are normally distributed. Substituting the Eq. 2–4 into Eq. 1 yields the following model, which we used to test the above hypotheses.

$$ {\mathrm{L}}_{\mathrm{ij}}={\tilde{\mathrm{a}}}_{00} + {\tilde{\mathrm{a}}}_{01}\left({\mathrm{D}}_{\mathrm{j}}^1\right) + {\tilde{\mathrm{a}}}_{02}\left({\mathrm{D}}_{\mathrm{j}}^2\right) + \left[{\tilde{\mathrm{a}}}_{10} + {\tilde{\mathrm{a}}}_{11}\left({\mathrm{D}}_{\mathrm{j}}^1\right) + {\tilde{\mathrm{a}}}_{12}\left({\mathrm{D}}_{\mathrm{j}}^2\right)\right]\left({\mathrm{Q}}_{\mathrm{ij}}\right) + \left[{\tilde{\mathrm{a}}}_{20} + {\tilde{\mathrm{a}}}_{21}\left({\mathrm{D}}_{\mathrm{j}}^1\right) + {\tilde{\mathrm{a}}}_{22}\left({\mathrm{D}}_{\mathrm{j}}^2\right)\right]\left({\mathrm{P}}_{\mathrm{ij}}\right)+{\mathrm{Y}}_{\mathrm{controls}}\ast {\mathrm{FControls}}_{\mathrm{ij}} + \mathrm{error} $$

(5)

Results

We followed Zhou et al. (2010) and applied a stepwise procedure. The null-model (model without predictors) divides the variance of the dependent variable into individual and brand levels. Although the brand-level variance (0.085) is only a fraction of the individual-level variance (0.718), it has been shown that even much smaller level-two variances are sufficient to distort results (Cohen et al. 2003). The One Way Random Effect Model first adds controls (baseline) and then the predictors from the individual level (full individual). The Intercepts and Slopes as Outcome Model first adds the brand level control to the individual model (baseline) and then includes the random intercept and slopes, as specified in Eq. 3. The different models reveal a steady decline of residual variance over the individual steps, which indicates good support for the proposed model (Zhou et al. 2010).

At the individual level, only two control variables were found to be significant (correct identification: b = −0.140; p < .01; familiarity: b = 0.292; p < .001). The other three controls of local ID (0.000), gender (−0.023), and city (0.036) were found to be insignificant. Because the foreign and IBPA brands in our study have different countries of origin, it could be argued that certain outliers could distort our results. For example, outliers such as Japanese brands could change the results, because of animosity toward Japanese products (Klein et al. 1998). We reran the model excluding the one Japanese brand in the study and found only minor changes in the results. The other foreign brands in the study are from Western countries and South Korea. Chinese consumers are not known to display any animosity toward these countries; the effects are thus not distorted by country-of-origin animosity with respect to the foreign brand or the acquirer’s brand, as suggested by Fong et al. (2013). Moreover, it may be that consumer sentiment is more positive toward some of these foreign countries than others, which may impact on the results. To rule out this possibility, we ran several SEM models based on the individual-level model and included only foreign and IBPA brands. In these models, we also control for an aggregate of cultural distance, calculated with Hofstede et al.’s (2010) cultural dimensions and for the macro and micro country image (Pappu et al. 2007). None of these constructs had a significant influence on consumer loyalty, so that possible consumer preferences for certain developed countries therefore also do not distort our results.

As expected, the covariance between quality and price is relatively high (b = 0.405; p < .001). When assessing both loyalty predictors, we find support that perceptions of both good quality (b = 0.501; p < .001) and a good price (b = 0.116; p < .001) drive loyalty intentions. These results indicate that the general effect of branded products in the market is positive for both drivers of loyalty. Thus, the level-one model receives good support (Table 6).

Table 6 Results of hierarchical linear modeling

At the brand level, while controlling for the product category (b = −0.248; p < .01), we tested for differences of local, foreign and IBPA brands from a grand mean. Attitudes have been proven to be relatively stable psychological constructs (Fishbein and Ajzen 1975), so these differences, especially concerning IBPA brands, are likely to be only marginally detectable. Hypothesis 1 predicted a higher level of loyalty intentions for local brands (a), an insignificant deviation from the overall brand effect for foreign brands (b), and a lower level of loyalty for IBPA brands (c). According to our results, the loyalty intercept of local brands is well above the grand mean (b = 0.118; p < .01), while the intercept for IBPA brands deviates negatively from the grand mean (b = −0.072; p < .01). For foreign brands, no significant deviation was detected. Therefore, H1 is supported. The findings can best be summarized with the equation: IBPA brands < general level of loyalty toward brands in the market = foreign brands < local brands.

According to H2, good quality is a stronger driver for loyalty intentions for local brands, compared to the general effect of branded goods in the market (a), a weaker driver for foreign brands (b), and an average driver for IBPA brands (c). Our results show that for local brands, the path of quality on loyalty has a significantly higher slope (b = 0.092; p < .05), foreign brands display a significantly lower slope (b = −0.142; p < .001), and IBPA brands do not differ significantly from the average brand effect (b = 0.050; p = .182). All findings are consistent with H2. The findings can be summarized with the equation: foreign brands < general brand effect of quality on loyalty = IBPA brands < local brands.

Hypothesis 3 predicts that the effect of an attractive price on loyalty is significantly weaker for local brands, compared to the average effect of branded goods in the market, and higher for foreign brands, while IBPA brands lie between foreign and local brands and thus do not differ from the average effect. Hypothesis 3b is supported, since the deviation of foreign brands is significantly above the grand mean (b = 0.133; p < .001). However, local brands do not fall significantly below the grand mean (b = −0.082; p = .106). IBPA brands behave as predicted and also do not differ from the average brand effect (b = −0.051; p = .178). Therefore, only H3b and H3c are supported, while H3a has to be rejected. To further investigate this hypothesis, we ran a Wald test of parameter constraint to determine whether the brand types differ significantly from each other. The results reveal that local brands differ significantly from foreign brands (Wald χ2 (1) = 6.702, p = .010) and that IBPA brands differ significantly from foreign brands (Wald χ2 (1) = 9.559, p = .002), but local brands do not differ significantly from IBPA brands (Wald χ2 (1) = 0.143, p = .71). Therefore, the different brand types align as follows in terms of a higher coefficient for the path from price to loyalty: local brands = IBPA brands = general brand effect of price on loyalty < foreign brands.

Discussion

The study indicates that acquiring a local FMCG brand in an emerging market may decrease brand loyalty and may lead consumers to expect higher quality from the acquired brand, while persistently demanding a low price. Thus, it is not an advisable strategy for foreign brand conglomerates. These findings are remarkable, since previous research has suggested positive abnormal returns for cross-border acquisitions (Morck and Yeung 1992) and brand acquisitions (Wiles et al. 2012).

More specifically, the findings indicate that consumers tend to display lower loyalty intentions toward IBPA brands, compared to the general level of branded goods in the market. In fact, the loyalty intentions are lower than those for both local and foreign brands. This indicates decreasing loyalty when local brands become IBPA brands after an acquisition. Deductive reasoning suggests that one main reason for this effect is that local consumers regard their freedom as being reduced by the foreign acquisition, and thus react by devaluing the imposed alternative. These results extend the observations of Thørbjornsen and Dahlén (2011) to cross-border acquisitions. More importantly, this study provides external validity to the applicability of psychological reactance in an M&A situation, because we refrain from potentially priming participants with reactance.

The present study develops the notion that the observed reactance of consumers is likely to be triggered by growing nationalism in emerging markets, a phenomenon also noted in a number of studies (e.g., Sheth 2011; Cavusgil et al. 2012; Becker 2013). According to the theory of psychological reactance, the magnitude of reactance depends on the importance of freedom to the individual (Brehm 1966). If nationalism continues to rise in emerging markets, it may induce local consumers to perceive the foreign takeover of a local brand as an even greater threat to their freedom. It is, therefore, plausible that the reactance following foreign acquisitions will increase further in the future.

Even though previous studies have indicated positive effects of both foreign and local brands, no study has explicitly tested which one has a stronger effect on loyalty. Our results suggest that in an emerging market such as China, local FMCG brands commonly enjoy a loyalty premium, compared to the general effect of brands in the market. However, contrary to previous findings (e.g., Batra et al. 2000; Strizhakova et al. 2008), foreign brands do not appear to enjoy such a premium. The reason could be that local brands benefit from their proximity to the market (BCG 2008), home country bias (Verlegh and Steenkamp 1999), and nationalistic tendencies in emerging markets (Sharma 2011). Additionally, other effects that may favor foreign brands (categories perceived as foreign, luxuries, and public consumption) are less dominant for FMCGs than in other product categories (Bain & Company and Kantar Worldpanel 2014).

The results from FMCG brands in China support the above reasoning. However, both influences, home country bias and psychological reactance, are also likely to apply in emerging markets beyond China. Catching up growth, developing brand-building capabilities, and rising nationalism are phenomena witnessed in all emerging markets and they tend to increase home country bias and psychological reactance. Moreover, we expect the influence of home country bias and of psychological reactance to be present in other product categories. However, these effects may then be covered, because the effects favoring foreign brands might dominate more in other product categories than FMCGs (Verlegh 2007).

The second finding indicates that for IBPA brands, the strength of quality as a driver of loyalty tends to be weaker than for local brands. This means that consumers’ quality expectations rise when local brands become IBPA brands after an acquisition, which is in accordance with the literature (Batra et al. 2000; Steenkamp et al. 2003). Özsomer and Altaras (2008, p. 10), for example, have argued that global brands “have greater credibility because of greater brand investments.”

The third finding indicates that the strength of an attractive price as a loyalty driver appears to be comparable between IBPA and local brands. This means that consumers continue to expect low prices from IBPA brands after the acquisition. This result is in line with Clemente and Greenspan (1997), who report that customers are very sensitive to price changes after a merger or acquisition. Combined with increased quality expectations, this finding is particularly important, because the low price expectations toward IBPA brands suggest a stagnant willingness to pay for the increased quality. This would drastically limit the potential to profit from the acquisition.

Theoretical implications

Introducing a new brand type – IBPA brands – and the theoretical and practical assessment of consumer reactions to this brand type are the main contributions of the present study. IBPA brands are created as a result of cross-country acquisitions in an emerging market, by a developed-country corporation. These brands are indeed recognized by consumers and treated distinctively from local and foreign brands, both in terms of attitudes (e.g., quality elasticity) and intentions (e.g., loyalty), as the present study demonstrates.

The introduction of IBPA brands to the literature is a major contribution, because the dominant theoretical notion in the consumer focused cross-border acquisition literature, spillover effects (e.g., Lee et al. 2011; Fong et al. 2013), is able to explain only a fraction of our findings. This part is the observation that higher quality expectations from consumers toward foreign brands seem to spill over to the newly acquired IBPA brands. We develop this idea theoretically by means of signaling theory, because the overall investment of foreign firms signals high quality. The investment then turns into a guarantee of the quality promise of IBPA brands and thus the signal literally spills over to them.

However, the spillover idea is incapable of explaining the two other important findings of this study. First, the spillover effect does not explain why price expectations of IBPA brands stay similarly low to those of local brands, whereas quality expectations rise. Apparently, consumers tend not to translate the rising costs of an increased quality level into a higher price expectation. In other words, the feedback loop from the receiver back to the sender of the signal is constrained, because the quality expectation–willingness-to-pay correlation (Tellis and Wernerfelt 1987) is suspended. Therefore, in order to further assess IBPA brands in the future, the theoretical focus must move beyond the spillover effect and pay special attention to price elasticity. Second, according to the spillover literature, there should not be a lower level of loyalty intention for IBPA brands beyond that for foreign brands. This effect, however, may be explained by the theory of psychological reactance, according to which consumers object to the selling out of their local brand icons to international conglomerates.

The study also makes two contributions to the broader theory of psychological reactance. First, the study relates to the question of subconscious psychological reactance. The literature has only briefly discussed this phenomenon, with a focus on reactance toward another individual in a relationship (Chartrand et al. 2007). While we find that a large share of consumers in our study identified IBPA brands correctly and indicated lower levels of loyalty, those who did not consciously recognize IBPA brands also displayed an effect. Hence, our results support the notion of subconscious psychological reactance and also suggest that this form of reactance is not limited to an interpersonal relationship. Second, we show that psychological reactance may extend signaling theory. A number of phenomena in the marketing literature (such as product scarcity, brand alliances, and acquisitions) have been analyzed with both psychological reactance and signaling theory, but the theories have been applied separately (e.g., Rao et al. 1999; Thørbjornsen and Dahlén 2011). Our results suggest that the theories are actually related to each other. The relationship can best be understood by considering psychological reactance as a form consumer feedback to the signaling action. A signal loop is only complete if the receiver provides feedback to the sender. This feedback has so far been understood as a way to improve the efficiency and reliability of signaling (Connelly et al. 2011). We suggest that psychological reactance may also constitute feedback to a signal in the form of disliking the sender’s signaling action, for affective/normative reasons. Considering psychological reactance as feedback to a signaling action turns signals into a double-edged sword. As we have shown for IBPA brands, despite the uncertainty decreasing function of the foreign ownership signal, an acquisition may not be advisable. Therefore, to provide a more holistic picture of the effects of a signal, we suggest that psychological reactance needs to be added to signaling theory as a possible feedback process.

In connecting both theory streams back to the COO paradigm, we support its core idea that the cognitive process (signaling) may be hindered by affective/normative processes (psychological reactance) (Obermiller and Spangenberg 1989; Verlegh and Steenkamp 1999). This possible detachment of the individual COO processes has been addressed by a number of studies (e.g., Klein et al. 1998; Verlegh and Steenkamp 1999). However, to the authors’ best knowledge, no study has extended the argument to the theoretical extreme that the effects point into opposite directions. Using IBPA brands as an example, we demonstrate that consumers can simultaneously exhibit a positive cognitive and a negative affective/normative response to a foreign takeover of a local brand.

Managerial implications

Many M&A studies stress the consolidation of resources as the key reason for the success of an acquisition (Capron et al. 1998; Capron and Hulland 1999; Swaminathan et al. 2008; Shimizu et al. 2004). However, the present study indicates that customer-related factors are also highly important. For example, for emerging market consumers who oppose the sales of local brands to foreign companies, even subtle motives like a perceived loss of decision-making freedom may lead to reactance and decreased consumer loyalty. A possibility to decrease psychological reactance by consumers would be to involve them in the acquisition process as much as possible. Hence, customers should be treated as active partners whose engagement is valuable and who have a say in the decision on whether or not to acquire an IBPA brand. This implies following the discussion in social media, but it could also involve an opportunity for customers to voice their opinions on whether to keep the old local brand name or create a new one. Such options may reduce psychological reactance, because consumers regain some lost freedom. If customers articulate that their local brand should remain untouched, the foreign company could try to grant relatively independent decision processes to the local entity and assure consumers that the essence of their local brand will persist.

However, the present study also demonstrates that customers notice not only the negative but also the positive aspects of the acquirer; this opens opportunities for the foreign firm. For example, consumers may see the foreign acquirer not only as a hostile threat to the heritage of the local brand but also as a facilitator of enhanced quality. This may pave the way for the foreign company to brand the M&A as an opportunity for the further development of the IBPA brand. For example, communicating that the IBPA brand will be introduced to other countries might create the image of a facilitator and additionally satisfy the national pride of customers.

When Beiersdorf acquired the C-bons group, it tried to increase prices and upgrade the new brands by advertising with high profile actresses and TV-show sponsoring (Madden 2009). Despite their efforts, the development of their newly acquired brands was still unsatisfactory (Heidenreich 2015). Our findings indicate that Beiersdorf may have misjudged the challenges of the decreasing loyalty and continuing low-price expectations. A better strategy might be to focus IBPA brand investments on quality upgrades rather than on endorsements and sponsorships. This way, the firm can focus on customer expectations and cut unnecessary costs to keep the price low. If the foreign company seeks to reposition the brand to target more affluent customer segments after the acquisition, it should consider introducing a new local brand instead. In this manner it could utilize the acquired new local competencies, create synergies between its current brands and new local brands, and prevent decreased loyalty due to psychological reactance from the beginning.

Limitations and future research

Several limitations should be noted that point to opportunities for future research. First, the study has a relatively narrow focus on China and FMCG brands. Second, the study controls only for time; it does not provide an assessment of the development of the effects over time. These effects, however, may be essential for a company’s evaluation of an acquisition decision and may also pave the way for more precise remedy measures by foreign companies. Finally, despite many acquisitions of local brands, the study does not control for, or uncover possible motives of foreign companies to enrich their portfolio with an IBPA brand.

This study is the first to examine IBPA brands, so we constrained the analysis to a very narrow field to isolate the factors examined. Future research should extend the focus beyond FMCG brands in China. Different product categories, as well as services, may reveal other effects that compensate for local bias or psychological reactance. Moreover, to put IBPA brands to the test, we recommend a replication in an emerging market that differs substantially from China, as well as in developed countries.

Different effects have been suggested as following psychological reactance over time, e.g., overreaction, helplessness, and alienation (Brehm and Brehm 1981). The present study controlled and tested for time effects but did not investigate them in detail. It remains to be tested how long customer perceptions of an IBPA brand linger on and whether, at some point, they become more closely associated with foreign brands. Additionally, consumers might anticipate a certain time frame for quality improvements to commence. Hence, their demands for higher quality from IBPA brands may also be characterized by a time delay. A longitudinal study could thus produce promising results. After achieving a greater understanding of the development over time, remedy measures like promotional activities before and after the acquisition would provide another promising future research area.

There are other reasons for acquiring an IBPA brand than those tested here. The literature discussing these reasons is rich (e.g., Harzing 2002; Lee and Lieberman 2010; Damoiseau et al. 2011). Notable ones include synergies for existing foreign brands of the acquirer (Capron and Hulland 1999) and access to segments that might be difficult to target with a foreign brand image (Rao et al. 2004). The present study demonstrates that, from a consumer perspective, for the IBPA brand alone, the payoff of an acquisition is highly questionable. Therefore, it should be valuable to investigate the possible positive effects of IBPA brands, such as synergies or the redeployment of marketing capabilities from the IBPA brand to other brands of an acquirer.