Global Fast Food Brands: The Role of Consumer Ethnocentrism in Frontier Markets

Abstract

Purpose: Modern globalization and Western markets saturation has catalyzed the growth of culinary globalization into developing countries. The question was whether fast food consumers in frontier markets of Sub-Saharan Africa (Zimbabwe), either upholds national gastronomic tendencies in terms of consumer ethnocentrism and buy local or they adopt global fast food brands. Demographic consumer profiles were also analyzed as antecedents of consumer ethnocentrism. Research design, data and methodology: A sample size of 400 fast food-adult consumers was surveyed in the City of Harare. Data was captured on SPSS and Analysis of Moment Structure (AMOS). Hypothesis testing was done using sample t test (H₁), logistic regression (H₂) and multiple regression (H₃, ₄, ₅) analysis. Results: Consumer ethnocentrism in Zimbabwe was marginally above average and no statistically significant relationship between the levels of consumer ethnocentrism and adoption of foreign fast food brands was noted. Age had an inverse relationship; income had a positive association whilst gender had no statistical significance with consumer ethnocentrism. Conclusions: Despite the Zimbabwean consumers being marginally ethnocentric, international restaurateurs should invest in the Zimbabwean fast food market since their nature of being foreign has got an exotic appeal to the Zimbabwean consumers thereby enhancing their likelihood of success.

1. Introduction

The world is fast becoming a global village (Park, 2018; Park, 2020; Son, 2020, Vu & Huang, 2020) and the global fast food market has been growing steadily (Becker, Lee & Nobre, 2018; Cho, 2021) at 4.2% from 2017 and the net worth of the industry was expected to be well in excess of US$690 billion in 2022 (Zion Market Research, 2019). Western fast food consumption patterns have been shifting to organic and natural food to curb the growth rate of affluent diseases associated with junk food such as obesity and heart diseases which are prevalent among wealthy consumers. Such developments in Western markets have been coupled with the growth of fast food global brands in emerging and frontier markets. Musacchio and Werker (2016) advised multinational firms to invest in frontier markets so as to get first mover advantages. Global firms with fast food brands are supposed to be ‘direct and bootson-the-ground’ in terms of their presence in frontier markets (Musacchio & Werker, 2016). Hence, global fast food brands have had an impetus to spread their market penetration into Sub-Saharan Africa and Zimbabwe in particular. However, the entry of global fast food brands into developing countries need to be informed through studies on the peculiar consumer behaviour tendencies towards foreign brands as such studies can present huge theoretical departures from conventional wisdom from Western markets (Wright, Filatov, & Peng, 2009). There is dearth in literature on ethnocentric consumer behaviour in developing economies, especially in economically erratic frontier economies.

The fast food concept has been in existence for nearly a century, but was only prevalent in the western world. Fast food is defined as an inexpensive food prepared using assembly line technology and served over the counter with much emphasis on convenience (Fleischhacker, Evenson, Rodriguez, & Ammerman, 2011). In the United States of America (USA) fast food giants such as McDonalds started to operate in the 1940s, Kentucky Fried Chicken (KFC) in 1952, Wendys’ in 1969, Pizza hut in 1999, while Wimpy opened its doors to the public in 1999.

As globalization and civilization came to Africa the fast food concept conveniently found its way into the food markets in the emerging and frontier economies such as Zimbabwe. In Zimbabwe, the fast food restaurant industry is one of the most resilient economic sectors whose demand has been exponentially growing despite concerns from some quarters of the society that fast food is characterized by high levels of saturated fat, sugar, and salt, while deficient of important micro nutrients (Bhat, 2016; Fraser, Edwards, Cade, & Clarke, 2010). In fact fast food has serendipitously become staple food for most urbanites in Zimbabwe. In Zimbabwe, international brands such as KFC, Mag & Bean, Nandos, and Steers managed to penetrate into the fast food restaurant industry that is dominated by local restaurant chain brands such as Chicken Inn, and Chicken Slicer. Considering that the fast food restaurant concept has been adopted in Zimbabwe there are reasonable grounds to suggest that other global brands such as McDonalds which are already thriving in the neighboring South Africa may soon find the local market lucrative.

The adoption of a product is largely dependent on both product characteristics and consumer characteristics (Kotler, 2015; Rogers, 2003). Product characteristics that determine the rate of adoption of a new product in a market are outlined in literature as relative advantage, compatibility, complexity, divisibility, and communicability (Kotler & Keller, 2016; Rogers, 2003). Consumer characteristics that determine the adoption of new products include, but not limited to age, gender, levels of education, income levels and consumer ethnocentrism in the case of international brands (Schiffman & Kanuk, 2004). Central to consumer ethnocentrism is an individual’s decision to adopt or reject foreign made products (Shimp & Sharma, 1987).

The concept of consumer ethnocentrism has been widely studied in extant international business literature (e.g. Kvasina & Crnjacckac, 2018; Makanyeza & Du Toit, 2017; Shimp & Sharma, 1987). Most of the focus of those studies was on the antecedents of consumer ethnocentrism. For instance, both socio-psychological antecedents such as patriotism, collectivism, nationalism, conservatism, animosity, and materialism, and demographic antecedents such as age, gender, education, and income had been the focus of most of the previous studies (e.g. Makanyeza, 2017). However, equally important are the outcomes of consumer ethnocentrism such as product or service adoption which were largely neglected in previous research. Moreover, previous research on consumer ethnocentrism has an unbalanced focus on product categories such as clothing, automobiles, electronic and electric gadgets leaving out other important categories such as fast food category (e.g. Kamwendo, Corbishley, & Mason, 2014). Moreso, research on restaurants have been largely local (e.g. Kim, Choi, & Shin, 2019) and also local brands (Kim, Kim, Kim, Mehyaoui & Youn, 2014). Therefore, this study seeks to fill in that important research gap using data collected from the fast food restaurant industry in Zimbabwe. The study was focused on the following questions:

What are the levels of consumer ethnocentrism in frontier markets of Sub-Saharan Africa?

What is the influence of consumer ethnocentrism on adoption of global fast foods brands?

Are demographics (age, gender and income) antecedents of consumer ethnocentrism?

The next sections are arranged as follows: review of relevant literature leading to specification of hypotheses, followed by an outline of the methodology adopted. After the methodology lay out, data analysis and then discussion of the results attained shall conclude the study.

2. Literature Review

2.1. Consumer Ethnocentrism

Consumer ethnocentrism is a psychological construct of economic nature that was conceptualized by Shimp and Sharma (1987) with regards to the appropriateness and morality of patronizing foreign made products and services. Consumer ethnocentrism is a derivative of Sumner (1906)’s ethnocentrism concept. Ethnocentrism is a sociological concept that is firmly rooted in the social identity theory. Consumer ethnocentrism is a general disposition that is different from an attitude, making ethnocentric tendencies more stable and durable (Schifman & Kanuk, 2004). Central to the concept of consumer ethnocentrism is the belief that one’s group is a point of reference for evaluation of offers from other cultural groups (Shimp & Sharma, 1987). The concept of consumer ethnocentrism appeals to the morality of citizens when it comes to decisions such as consuming or patronizing foreign brands. The basis for this appeal is that foreign products tend to retard the growth of the domestic economy leading to increases in unemployment rates and related social ills (Shimp & Sharma, 1987).

Consumer ethnocentrism does not necessarily imply a negative attitude towards foreign products, but rather it is simply a bias in favor of domestically produced products (Balabanis & Diamantopoulos, 2004). Moreover, the levels of consumer ethnocentrism are not static, but vary depending on the product category under consideration (Kim & Pysarchic, 2000; Herche, 1992; Sharma et al., 1995). For instance, a consumer may exhibit high tendencies of ethnocentrism towards foreign food items, but may have different tendencies towards electronic gadgets. Vida and Readon (2008) identified three components of consumer ethnocentrism which are namely cognitive, affective, and normative. The cognitive component entails one’s perceptions that domestically produced products are superior to imported products. Affective component relates to one’s affection for locally produced goods. The normative component refers to the moral obligation that one feels with regards to purchasing locally produced goods as opposed to imported goods.

2.1.1. Levels of Consumer Ethnocentrism Based on Nature of Markets

The levels of consumer ethnocentrism tend to vary depending on the nature of the marketunder consideration. Scholars have categorized markets in the binary form as developed markets and emerging markets (Burgess & Steenkamp, 2006; Sheth 2001). However, emerging markets are too broad and amorphous such that this binary approach should be refined by dividing emerging markets into prime emerging markets, frontier markets and least developing economies. Retaining the binary categorization would lead to combining BRICS which are fast growing with frontier economies characterized by lack of stability with regards to their macroeconomic environments. Unlike emerging markets in Eastern Europe and Asia, African economies are characterized by stagnant growth leading to consumers inevitably preferring foreign made products (Kibret, 2016).

Extant consumer behavior literature has emphatically revealed that consumer ethnocentric attitudes and tendencies are highly prevalent among consumers in developed markets than those in emerging markets. Consumers in emerging and frontier markets generally prefer brands of foreign origin (Hamin & Elliot, 2006). The general preference for foreign products has found empirical support from studies on emerging markets such as Crotia (Kvasina & Crnjacckac, 2018), China and Russia (Klein, Ettenson, & Krishnan, 2006), and Hungary (Papadopoulos, Heslop, & Beracs, 1990). Empirical evidence has also revealed that in emerging markets foreign brands are assessed to be superior to domestic brands (e.g. Good & Huddleston, 1995; Suphellen & Rittenburg, 2001). More so, research findings also emanated from several studies (e.g. Javalgi, Khare, Gross, & Scherer, 2005; Sharma et al., 2005) where it was concluded that consumers from developed markets are more ethnocentric than consumers from emerging markets. The lack of research on consumer ethnocentrism in frontier markets in Sub-Saharan Africa becomes apparent from the foregoing empirical studies discussion.

Foreign brands’ superiority tends to be embedded in aesthetic store image, fancy brand names, and innovative product attributes (Bruning, 1997; Sharma, 2011). This suggests that foreign brands are regarded as more valuable if they are from developed markets (Maina, Kibera, & Munyoki, 2015). Consumers in emerging markets are likely to have less favorable attitudes towards local brands due to their inherent disposition towards status oriented consumption behavior (Batra, Ramaswamy, Alden, Steenkamp, & Ramachander, 2000; Chen & Hsin-Hui, 2010; Nguyen & Smith, 2012; Wong & Chen, 2004). However, most of the previous research was disproportionately tilted towards emerging markets in Eastern Europe and Asia. It is therefore imperative to test the validity of the previous research findings on samples drawn from frontier markets in Africa. Therefore based on previous research findings, it is prudent to suggest that;

H1 : The levels of consumer ethnocentrism are low in frontier markets

2.2. Product Ethnocentrism Adoption and Consumer

Adoption is defined as an individuals’ conscious decision to regularly patronize a product or service (Rogers, 2003). In accepting or rejecting a new product, a potential consumer moves through the stages of an adoption process that starts with awareness, interest, evaluation, trial and eventually adoption or rejection depending on product characteristics (Rogers, 2003).

Product characteristics that influence adoption are relative advantage, complexity, trialability, compatibility, and Observability (Rogers, 2003). Relative advantage refers to the degree to which innovation is regarded as being better than the prevailing competing products. Complexity refers ease of use of a new brand. Trialability is a degree to which potential consumers may use a product on experimental basis without incurring costs. Compatibility is regarded as being consistent with the potential customers’ needs, values, and previous experiences. Observability is conceptualized as a multidimensional construct made up of demonstrability and visibility. Demonstrability refers to tangible benefit arising out of using a brand, while visibility relates to brand’s relative exposure (Wei, 2001).

The adoption process is also influenced by consumer characteristics such as the socio-economic factors, demographic factors, personality traits, and consumer ethnocentrism in the case of global brands. Consumer ethnocentrism has been modeled as an antecedent of a number of psychological constructs such as purchase intention, consumer choices, brand preferences, and more importantly consumer adoption (Sharma et al., 1995). Previous research has shown that consumers exhibit ethnocentric tendencies towards foreign products (e.g. Makanyeza, 2017), and acknowledged how this affects the levels of foreign brands adoption (e.g. Hsu & Nien, 2008; He & Wang, 2015; Qing, Lobo & Li, 2012; Wei, Wright, Wong, & Yu, 2009).

The relationship between consumer ethnocentrism and product adoption is best explained by the application of the extended Theory of Planned behaviour (TPB). Maksan, Kovacic, and Cerjak (2019) referred to the fusion of consumer ethnocentrism on the TPB as; consumer ethnocentrism extended model of the TPB. The TPB (Ajzen, 1991) is an extension of the theory of Reasoned Action (TRA) (Ajzen & Fishbien, 1980). The TPB contends that behaviour is a function of intention (Choi & Lee, 2012; Guo, Kim, & Kim, 2020; Ngan & Khoi, 2020), which in turn is a function of the three weighted constructs: attitude, norms, and control (Ajzen, 1991). Since consumer ethnocentrism is related to attitudes (Maksan, Kovacic, & Cerjak, 2019) and adoption is a behavioural activity, a clear link is established between these two constructs using the TPB. Thus, consumers develop ethnocentric tendencies through either commercial or social means which will in turn influence their behaviour in the form of new brands adoption. This view received empirical support from several studies that found the prevalence of an inverse relationship between the levels of consumer ethnocentrism and adoption of foreign brands (.e.g. Netemeyer et al., 1991; Shimp & Sharma, 1987). Recently, Maksan, Kovacic, and Cerjak (2019) noted a positive relationship between consumer ethnocentrism and consumer purchase of foreign wine.

It is therefore hypothesized that;

H2 : Consumer ethnocentrism has an influence on foreign product adoption.

2.3. Age and Consumer Ethnocentrism

There are reasonable grounds to believe that age has an effect in consumer ethnocentrism. More specifically, the assumption is that old aged adults are more ethnocentric than young adults. This is based on the understanding that old aged adults are usually more conservative and less exposed than young adults. Empirical studies have provided mixed findings on the influence of age to the levels of consumer ethnocentrism. In a study by Jain and Jain (2013) in India, Hamelin et al. (2011) in Morocco, Jalvalgi et al. (2005) in France, Balabanis et al. (2002) in Turkey, Supphellen and Rittenburg (2001) in Poland, age was found to be positively related to consumer ethnocentrism. However, an insignificant relationship was also noted on studies by Pannanen et al. (2017) in Russia, Mangnale, Potluri, and Degufu (2011) in Ethopia, and Balabanis et al. (2002) in Czech Republic. Based on a balance of probability, it can therefore be hypothesised that;

H3 : There is positive and significant relationship between age and consumer ethnocentrism in the fast food restaurant industry.

2.4. Gender and Consumer Ethnocentrism

Gender differences have been studied infood selection (Yim, 2020). It is generally assumed that women are more consumer ethnocentric than men. This assumption emanate from the fact that women, particularly those from emerging markets are less exposed to global marketing information that is relevant for understanding the products in the market. Consequently, women tend to trust locally manufactured products whose production standards are more visible to them. Moreover, women are more conservative, comfortable with established practices (Shankarmahesh, 2006), and are inclined towards preserving social harmony than men (Sharma, Shimp & Shin, 1995). Studies by Javalgi et al. (2005), and Balabanis et al. (2002) had shown that consumer ethnocentrism is influenced by gender. However, other studies indicated that a non-significant relationship exist between gender and consumer ethnocentrism (e.g. Vida & Damjan, 2001). Based on the reviewed literature it can be seen that most of the previous studies’ findings are tilted towards confirming the existence of a significant relationship between gender and consumer ethnocentrism. It can therefore be hypothesised that;

H4 : There is a relationship between gender and consumer ethnocentrism in the fast food restaurant industry.

2.5. Income and Consumer Ethnocentrism

There is a proposition in the extant marketing literature that income is significantly related to consumer ethnocentrism. That relationship indicates that low income consumers are high on ethnocentrism, while high income consumers are low on ethnocentrism. Shimp and Sharma (1987) indicated that low income earners feel threatened by foreign products which in most cases are unaffordable to them. The unaffordability of foreign products may lead to high levels of consumer ethnocentrism. In contrast high income earners are less ethnocentric probably due to the fact that they have opportunities to travel abroad and purchase foreign manufactured products. It can therefore be hypothesised that;

H5 : There is a negative relationship between the levels of income and the levels of consumer ethnocentrism.

The next section presents information on the methodology followed in testing the hypotheses developed in the previous section. It includes sampling and data collection procedures, measures used, and the data analysis procedures adopted.

3. Research Methods and Materials

3.1. Sample and Data Collection Procedures

The respondents surveyed were contacted in the City of Harare based on the fact that almost all the international fast food restaurant brands operating in Zimbabwe are situated in Harare. The sample size was 400 which was determined using the requirements of exploratory factor analysis (EFA) (Hair, Black, Babin, & Anderson, 2014). Convenience sampling was used to contact respondents. Convenience sampling is common in surveys done in general (Pradana & Wisnu, 2021) as well as in restaurant and hospitality industry (Barger & Grandey, 2006) due to the fact that most of the target respondents are not willing to participate in studies which potentially delay their excursion or disturb their social outings.

3.2. Measures

The levels of consumer ethnocentrism were measured using the 10 item shortened version of the Consumer Ethnocentric Scale (CETSCALE) (Bearden & Netemeyer, 1999; Sharma et al., 1995). The CETSCALE has been developed as an etic measure with a universal applicability that allows for comparisons of findings across varying cultures. The strength of this scale is that it measures the consumers’ disposition to act rather than attitudes (Sharma et al., 1995). Tendencies are dispositions to behave in a specified consistent fashion (Teo, Mohamad, & Ramayah, 2011). Furthermore, the use of the CETSCALE in the extant literature has been without major psychometric concerns. The scale has been found to exhibit validity, reliability, and uni-dimensionality in several studies across the globe (e.g. Sepehr & Kaffashpoor, 2012).

The respondents were requested to indicate their levels of agreement with several statements describing consumer ethnocentric on a 7 point Likert type scale ranging from (1) strongly disagree to (7) strongly agree. In line with tradition widely followed in literature, product category was not included in the statements of the original scale items (Jimenez-Guerrero et al., 2014). The CETSCALE items were coded for subsequent analysis as follows;

Table 1: The CETSCALE Items and Coding

Adoption was operationalized as a dichotomous variable represented by the response categories indicating an acceptance or rejection of a foreign brand of fast food in the market, and coded with 1 representing acceptance and 0 indicating rejection.

3.3. Data Analysis Procedures

The data analysis process was conducted in two phases using Statistical Package for Social Scientists (SPSS) version 23, and Analysis of Moment Structure (AMOS) version 21 data analysis statistical packages. The first phase involved validation of the latent variable (consumer ethnocentrism) using both exploratory factor analysis (EFA) and confirmatory factor analysis (CFA) for dimensionality and convergent validity tests respectively. The second phase comprised of hypotheses testing using one-sample t test for the first hypothesis and binary logistic regression analysis (BLRA) for the second hypothesis. The choice of BLRA is based on the fact that BLRA is suitable for testing hypotheses where outcome variables are dichotomous (Leech, Barrett, & Morgan, 2005; Peng, Lee, & Ingersoll, 2002). Most consumer decisions are categorical in nature such as to buy or not to buy, to use or not to use, and to adopt or not to adopt, making them amenable to the use of BLRA. Demographics, that is, age, gender and income were analyzed as antecedents of consumer ethnocentrism using multiple regression analysis in SPSS.

4. Results

4.1. Demographic Profile of the Respondents

The demographic profile of the respondents who participated in this study is shown in Table 3.

Table 3: KMO and Bartlett's Test

The sample of the respondents in this study had the composition of age groups shown in table 2 represented as follows: 18-29 (32.5%), 30-39 (30%). 40-49 (25%), 50-59 (10%), and 60+ (2.5%). Females (57.5%) outnumber males at 42.5%. The level of education for most of the respondents was secondary education (55%), followed by tertiary education (25%), and primary education (20%).

Table 2: Demographic Profile of Respondents

4.2. Measurement Scale Validation

Measurement scale validation was conducted using EFA to test unidimensionality and CFA focusing on convergent validity and internal consistency reliability.

4.2.1. Factor Analysis Statistical Evaluation

The appropriateness of factor analysis was evaluated using the Bartlett’s test of sphericity (Bartlett, 1954), the KMO measure of sampling adequacy test (Kaizer, 1974) (see table 3), and multivariate normality tests. A Bartlett’s test that is significant and a higher KMO index imply that there are some correlations among items; a condition that signifies the presence of underlying factor(s) (Burns & Burns, 2008). The Bartlett’s test indicates whether a correlation matrix of the items in the data collection instrument is significantly different from an identity matrix (Hair et al., 2014). An identity matrix is characterized by items badly correlating with each other (Field, 2013). A Bartlett’s test has a null hypothesis that the original correlation matrix is an identity matrix (Bartlett, 1954). A KMO value closer to 1 indicates a higher degree the compactness of the patterns of correlations suggesting that distinct and reliable factor(s) can be yielded from EFA (Field, 2013).

The Bartlett’s test of sphericity and the KMO measure of sampling adequacy yielded the indices of 1163.879 with 10 degrees of freedom at a significant value of 0.000, and 0.866 respectively which all indicated that the data was appropriate for EFA.The KMO statistic of 0.866 attained in this study falls in the Kaiser (1974)’s meritorious region and this suggests the adequacy of the sample size necessary for factor analysis (Field, 2013).

The appropriateness of the sample data for use in factor analysis was further tested using multivariate normality tests since the estimation method was the maximum likelihood (ML). CFA in AMOS as a covariance structure analysis (Arbuckle & Wothke, 1999) with a ML estimation method has the requirements of multivariate normal probability distribution of data (Izquierdo, Olea, & Abad, 2014; Oke, 2012). ML is a default parameter estimation method in most data analysis software’s (Schermeller- Engel et al., 2003). ML works well with large sample sizes of around 400 (Boomsma & Hoogland, 2001), but its strength is that it is scale free and scale invariant which implies that it is not affected by the nature of data being original or transformed (Bollen, 1989; Schermeller-Engel et al., 2003). Multivariate normality tests results are shown in Table 4.

Table 4: Normality Tests

The normality tests in the form of skewness and kurtosis indicated that the data was normally distributed as all the indices for all the items had their absolute values falling within the standard range of 2 for both skewness and kurtosis (Field, 2013). Furthermore, the presence of normality in data distribution was evidenced by the scores attained after dividing the values of skewness and kurtosis by their respective standard errors being within the +/_1.96 range for all the items.

Having ascertained the appropriateness of factor analysis, EFA was conducted (see table 5) to test the CETSCALE dimensionality (Fabrigar, Wegener, MacCallum, & Strahan, 1999). ML was used for items extraction. Besides the benefits of ML discussed earlier on, ML as an extraction method is appropriate for EFA tests whose results shall be further subjected to CFA (Hair et al., 2014). Items with the factor loadings of less than 0.5 were suppressed and excluded from further analysis (Hair et al., 2014).

Table 5: Factor Analysis

The original CETSCALE’s construct domain was hypothesized as uni-dimensional (Shimp & Sharma, 1987). Unidimensionality refers to a set of items representing one underlying construct (Hair et al., 2014), and is a prerequisite to the assessment of validity and reliability (Dietvorst, Willem, Verbeke, Bagozzi, Yoon, Smits, & van der Lugt, 2009). Unidimensionality is best assessed using EFA (Germain, Droge, & Daugherty, 1994). The items representing the CETSCALE yielded a single factor made up of 5 items out of the original 10 items showing further evidence of the scale’s uni-dimensionality. The items CE1, CE2, CE3, CE4, and CE5 poorly loaded on the scale and were therefore excluded from further analysis. The loading of the 5 items on the CETSCALE and the exclusion of the other 5 items that poorly loaded left the scale with the evidence uni-dimensionality (Bearden et al., 2011). Having less than 10 items in measuring the levels of consumer ethnocentrism is not without precedence e.g. Klen, Ettenson, & Krishnan, 2006; Kwamwendo et al., 2014). Furthermore, the single factor extracted from EFA explained 66.42 % of the total variance, and had an eigenvalue of 3, 321. An eigenvalue is a conceptual representation of variance accounted for by a factor (Burns & Burns, 2008).

After conducting EFA the items were further subjected to CFA in order to assess the hypothesized factor structure in the form of convergent validity. Results shown in Table 5 indicate that the CETSCALE bears some convergent validity as evidenced by all the items loading significantly (Anderson & Gerbing, 1988). The validity of the one factor model for CETSCALE was further evaluated using fit indices. Model fit relates to the extent to which implied covariance matrix closely resembles the empirical covariance matrix (Hair et al., 2014). The most commonly used fit indices in assessing measurement models are the chi-square (X2), Root Mean Square Error of Approximation (RMSEA), Standardized Root Mean Square Residual (RMR), Goodness-of-fit Index (GFI) (Jereskog & Sorbom, 1989), Comparative Fit Index (CFI) (Bentler, 1990), and the Adjusted Goodness Fit Index (AGFI) (Jereskog & Sorbom, 1989), Tucker-Lewis Index (TLI) (Tucker & Lewis, 1973), and Normed Fit index (NFI) (Bentler & Bonnet, 1980). An insignificant chi-square signifies model fit while, the values for GFI, and NFI, CFI, NNFI, and PNFI range from 0 to 1 with values closer to 1 indicating a good fit (Byrne, 2010).

All the fit indices shown in Table 6 indicate the fitness of the one factor measurement model of consumer ethnocentrism except for the X2 which was highly significant. A significant X2 suggests that there are considerable differences between the predicted model and the actual model. However, X2 is sensitive large sample sizes (Hair et al., 2014), therefore, an alternative suggested by Jereskog and Sorbom (1993) which involves examining the ratio of X2 to the degree of freedom was assessed. The X2/df attained was 4 which were still above the rule of thumb suggested by Kline (1998) as 3. As a result the researchers had to rely on the other indices for assessing model fit.

Table 6: Fit Indices

The reliability of the CETSCALE in the form of internal consistency obtained in this study was determined through the widely used Cronbach’s alpha coefficient (Bearden, Netemeyer & Haws, 2011). The extracted Cronbach’s alpha coefficient was 0.95 was within the range of 0.88 and 0.96 which was attained in most of the previous studies (e.g. Kamwendo et al., 2014; Shimp & Sharma, 1987).

4.3. Hypotheses Testing

The first hypothesis (H1) stated that the levels of consumer ethnocentrism are low in frontier markets. The results from hypotheses testing are shown in Table 7. This hypothesis was tested using one-sample t test.

Table 7: One-Sample T Test

The results in Table 7 indicate that H1 was not supported (t (399) =0.977; p=0.329). This means that the mean value of 3.5395 attained on a 7 point Likert type scale from this study is not statistically different from a test value of 3.5.

The second hypothesis (H4) sought to determine the relationship between the consumers’ ethnocentric tendencies and the adoption of fast food foreign brands. A BLRA was used to test this hypothesis. However, before hypothesis testing, statistical assumptions related to BLRA were assessed. BLRA has got only one assumption that data follows a binomial distribution (Field, 2013). The binomial distribution for BLRA is shown in Table 8.

Table 8: Binomial Test

a. Alternative hypothesis states that the proportion of cases in the first group < .470000

The results shown in Table 8 indicate that data for the dependent variable (Adoption) in this study follows a binomial distribution as evidenced by a probability value of 52%. Having satisfied this statistical assumption, the researchers proceeded to test H2.

5. Results and Discussion

The results shown in Table 9 reveal that there is no statistically significant relationship between the consumers’ levels of consumer ethnocentric tendencies and adoption of fast food foreign brands (B-0.059, p=739). The Exp (B) of 0.943 suggests that the odds of consumers rejecting are the same as the odds of accepting foreign fast food brands. This suggests that probability of adopting or rejecting foreign fast foods in the Zimbabwean fast food market in particular, and in frontier markets in general is not a function of consumers’ levels of ethnocentrism.

Table 9: Variables in the Equation

a. Variable(s) entered on step 1: Consumer ethnocentrism.

The results shown in Table 9 reveal that there is no statistically significant relationship between the consumers’ levels of consumer ethnocentric tendencies and adoption of fast food foreign brands (B-0.059, p=739). The Exp (B) of 0.943 suggests that the odds of consumers rejecting are the same as the odds of accepting foreign fast food brands. This suggests that probability of adopting or rejecting foreign fast foods in the Zimbabwean fast food market in particular, and in frontier markets in general is not a function of consumers’ levels of ethnocentrism.

The model for consumer ethnocentrism and adoption shown in Table 10 has got a low explanatory power as evidenced by the Nagelkerke R square of 0.001, which implies that less than 1 % of the variation in foreign fast food brands is explained by consumer ethnocentrism.

Table 10: Model Summary

a. Estimation terminated at iteration number 2 because parameter estimates changed by less than .001

Using multiple regression analysis in SPSS, it was noted that as one’s age increases, consumer ethnocentrism tendencies decrease. The third hypothesis (H3) was therefore not supported. This was a departure from majority of findings in extant literature. Gender, for the fourth hypothesis (H4), had no statistically significant effect on consumer ethnocentrism tendencies. Therefore, hypothesis 4 was also not supported. As one’s income increases the levels of consumer ethnocentrism tendencies also increase (see Table 11). These findings were contrary to the fifth hypothesis (H5) where a negative relationship had been proposed.

Table 11: Demographic Antecedents of Consumer Ethnocentrism

a. Dependent Variable: Consumer Ethnocentrism

6. Discussion

Fast foods have been adopted in Zimbabwe despite the overwhelming evidence that links fast food to non-communicable diseases such as diabetes, cancer, and high blood pressure (Maholtra, Noakes, & Phinney, 2015). In fact, fast food has become a staple diet for most working class people who are time constrained (Oluwafemi & Dastane, 2016). The ongoing global economic integration is opening up markets previously restricted by various forms of protectionism. Consumer ethnocentrism is one such form of self-imposed protectionism that has a weakened form in frontier markets such  as Zimbabwe. Therefore consumer ethnocentrism is one of the variables that need to be assessed when evaluating the lucrativeness of foreign markets (Jimenez-Guerrero, Gozquez-Abad, & Linares- Aguero, 2014).

The success of foreign brands in international markets relies upon the levels of consumer ethnocentrism (Savitha & Dhivya, 2017). Consumer ethnocentrism is generally viewed as having a negative influence on purchase behaviour towards foreign brands (Balabanis & Siamagka, 2017). The lower the levels of consumer ethnocentrism, the higher the chances of success for foreign brands, and vice versa since consumers high on ethnocentrism believe that products produced domestically are superior to those produced abroad. In fact, more often than not, country of origin serves as a cue for indicating the quality of a product or a service. However, the results in this study revealed that consumers’ levels of ethnocentrism did affect the adoption of foreign fast food brands. Thus, the results in this study revealed a stark contrast between consumers in frontier markets of developing countries such as Zimbabwe and developed countries such as the USA in terms of how they evaluate foreign products (Batra, Ramaswamy, Alden, Steenkamp, & Ramachander, 2000; Chen & Hsin-Hui, 2010; Nguyen & Smith, 2012; Wong & Chen, 2004). In frontier markets of developing countries, foreign brands which in most cases originate from developed countries are viewed as superior in quality and their consumption entails a symbol of high social status. This is contrary to how foreign brands are evaluated in developed countries where purchase and consumption of foreign products is regarded as unpatriotic and retrogressive (Shimp & Sharma, 1987).

In this study it was revealed that consumers in Zimbabwe generally have marginally above average levels of consumer ethnocentrism. The success of foreign fast food brands in Zimbabwe is a testimony of the wide adoption of foreign brands as claimed by most of the respondents. This aligns well with the inherent fact that most of the frontier markets with transitional economies have a sense of patriotism and nationalism as a result of being weaned from colonialism. Ethnocentrism has been modelled as function of patriotism and nationalism (Balabanis, Diamantopoulous, Mueller, & Melewar, 2001; Tsai, Lee, & Song, 2013). Therefore ethnocentric tendencies in the general populace are a result of incessant political mantras awash in the mainstream media peddled by political leadership trying to solidify their protectionist rhetoric. Patriotism tends to side-step rational decision making to the extent of invariably adopting domestic products that may be visibly inferior (Sharma, 2011). The marginally higher levels of consumer ethnocentrism found to be prevalent among urbanites in Zimbabwe smacks of regrettable and retrogressive connotations of salient and implicit isolationism, nationalism, and protectionism policies, which any progressive oriented nation should repeal.

Paradoxically, the marginally higher levels of consumer ethnocentrism recorded from the respondents did not translate into general preference for local brands. Rather most of the respondents indicated that they have adopted foreign fast food brands than local brands, despite being high on consumer ethnocentrism. This may suggest the shortcomings of the CETSCALE in terms of nomological validity. The findings in this study that consumer ethnocentrism is not related to the adoption of foreign brands is an additional support to the accumulated empirical evidence from some of the previous research (e.g. Bi, Gunessee, Hoffman, Hui, Loner, Ma, & Thompson, 2012; Tsai, Lee, & Song, 2013; Wong, Yim, Polonsky, Garma, 2008). However, since another stream of research has also indicated an emphatic relationship between consumer ethnocentrism and adoption of foreign brands, there is a probability that this relationship may be moderated by other variables such as psychographic and psychological factors.

Moreover, the differences in results obtained in this study with those found in most of the previous studies can be traced back to the data collection instrument. Although the CETSCALE has been designed as etic measure, it is possible that in some developing economies it may not be fully applicable. The use of a questionnaire in data collection may be unnatural leading to the introduction of limited ecological validity (Bryma & Bell, 2015). The fact that the most of the respondents indicated that they are consumer ethnocentric, despite adopting foreign brands may point to the challenges associated with consumer behaviour studies in frontier markets. Issues like ethnocentrism and patriotism are very sensitive and deeply entrenched in some emerging and frontier economies to the extent that some respondents tend to give socially desirable and politically correct responses. However, such responses are often not indicative of their behaviour.

The weak link between the levels of consumer ethnocentrism and adoption of foreign brands may also be accounted by the fact that in most cases in emerging and frontier markets there are no levels of equivalency of foreign brands. In such cases consumers have no option but to simply adopt foreign brands since there are no available domestic alternatives. Even where alternatives are available they are often vastly inferior to their imported substitutes. This is sustained by the declining means of production, and chronic inability to be innovative in some of the frontier markets particularly those in Africa.

7. Managerial Implications

Global fast foods brand managers targeting to establish their brands in Zimbabwe may need to take note of the unique characteristics of the market. High income earners are less interested in global brands but are reverting back to organic foods which are home-grown. Most affluent diseases are associated with junk food, so these high income earners tend to reduce consumption of such foods. Moreso, older consumers are an easy target for global fast foods in Zimbabwe frontier market. Older consumers may be nostalgic with regards to working with former colonisers unlike younger ages that might have been swayed by nationalistic and patriotic marketing communications sponsored by the Government of Zimbabwe. Older consumers can be employed as a beach-head to launch a global brand into the market. Given that Zimbabwean consumers are marginally ethnocentric, international restaurateurs should invest in the Zimbabwean fast food market since their nature of being foreign has got an exotic appeal to the Zimbabwean consumers thereby enhancing their likelihood of success.

8. Limitations, Conclusions and Future Research

The sample for this study was drawn from Harare, the capital city of Zimbabwe where most fast food international brands are hosted. However, Harare is a cosmopolitan city which inherently is a brew of different cultures and attitudes. Cosmopolitan refers to the openness towards cultural diversity (Roudometof, 2005). Cosmopolitanism tend to have a buffering effect on the intensity of consumer ethnocentrism (Nijssen & Douglas, 2008), since cosmopolitan consumers hold universal aspirations that are not limited by blind allegiance towards their nation’s products (Riefler, Diamantopoulos, & Siguaw, 2012). This cultural diversity in Harare makes the level of consumer ethnocentrism to be less intense than they would have been had the sample been drawn from less cosmopolitan regions of Zimbabwe such as Masvingo. Moreover, the acceptance or rejection of foreign brands cannot be solely linked to consumer ethnocentrism alone. There are several factors that may also affect the adoption of foreign brands such as price, quality, taste etc.

This study has provided some theoretical and practical relevance in the accumulated body of knowledge with regards to consumer behaviour in the realms of international business. Firstly, this study established or re-emphasized some of the boundaries of the consumer ethnocentrism theory. It demonstrated that while in general consumers have ethnocentric tendencies towards foreign products, there are some boundaries to this generalizability. In this study it was revealed that consumer ethnocentric tendencies and attitudes are less apparent in consumers from frontier markets. Moreover, consumer ethnocentric tendencies are also weak with regards to foreign products in frontier economies that are plagued with shortages, and quality issues.

Practically, this study revealed that consumers bear marginal levels of ethnocentric tendencies which however, do not affect their adoption of foreign products. This implies that international businesses especially from developed markets are more likely to succeed in emerging and frontier markets. Thus, consumers in emerging and frontier markets are favorably disposed towards foreign brands which they regard as innovative and superior. This gives a green light for investors from established markets to spread their tentacles to frontier markets such as Zimbabwe.

Future research may focus on unraveling antecedents of consumer ethnocentrism in the context of frontier markets especially psychographic variables such as beliefs, Ubuntu and culture. Cross-cultural studies on antecedents of consumer ethnocentrism in Southern African countries would be invaluable so as to provide a clear consumer ethnocentric tendencies map to global brand managers encompassing several industries.