Brand portfolio breadth and regionalization
in the Asian market
___________________________________________________________________________
MSc Business Administration – International Management
Amsterdam Business School
Master Thesis
Frank Veldkamp
10445072
STATEMENT OF ORIGINALITY
This document is written by student Frank Veldkamp who declares to take full responsibility for the contents of this document. I declare that the text and the work presented in this document is original and that no sources other than those mentioned in the text and its references have been used in creating it. The Faculty of Economics and Business is responsible solely for the supervision of completion of the work, not for the contents.
ABSTRACT
Finding the right regionalization strategy to be successful abroad has an important impact on firm performance and is often difficult to change or correct after strategic decisions are made. How a company decides to compose its brand portfolio, depends on a lot of factors and can provide both opportunities and challenges. These include cost benefits and meeting demands of heterogeneous countries. The purpose of this thesis is to investigate the relationship between brand portfolio breadth and the degree of regionalization and to determine the moderating role of host country contexts in this relationship. Economic openness, infrastructure quality and economic instability are used as host country context indicators. Secondary data of 45 Asian companies operating in the Asian beer, spirits and wine markets between 2012 and 2015 is collected to test the aforementioned relationship and obtain the results by using negative binomial regression models. The results show that there is a positive relationship between the domestic brand portfolio breadth and the degree of regionalization. The host country indicators economic openness and infrastructure quality have a negative moderating effect on this relationship, whereas economic instability had no significant effect. This thesis tries to extend the literature on brand portfolio strategy and to provide managers with empirical evidence to help them making future brand portfolio decisions.
Keywords: brand portfolio breadth/strategy; degree of regionalization; host country context; economic openness; infrastructure quality; economic instability
TABLE OF CONTENTS
1. INTRODUCTION ... 5
2. LITERATURE REVIEW ... 8
2.1 Brand portfolio strategy ... 9
2.2 Regionalization ... 12
2.3 Host country context ... 13
2.4 Research gap ... 16
3. HYPOTHESES DEVELOPMENT ... 17
4. METHODOLOGY ... 20
4.1 Data and sample ... 20
4.2 Dependent variable ... 22 4.3 Independent variable ... 22 4.4 Moderators ... 23 4.5 Control variables... 23 4.6 Empirical strategy ... 25 4.7 Model ... 26 5. RESULTS ... 26 5.1 Descriptive statistics ... 27 5.2 Correlations ... 28 5.3 Regression analysis ... 32 6. DISCUSSION ... 34
6.1 Key findings and contributions ... 34
6.2 Managerial recommendations ... 36
6.3 Limitations and future research ... 37
7. CONCLUSION ... 37
ACKNOWLEDGEMENTS ... 39
1. INTRODUCTION
Over the past decades, globalization started to have an increasing impact on international trade and became inseparable from the modern business environment. Decreased transportation costs and technological developments encouraged companies to expand abroad and serve markets outside their home country (Dunning, 1988). These companies that operate in multiple countries, multinational enterprises (MNEs), make product, market and development decisions that best suit their international expansion strategy. When they encounter different regulations, infrastructures and currencies, firms try to find a fit between these external contexts and the internal resources of the company (Chang & Wang, 2007; Dunning, 1988). Globalization forced MNEs to expand outside their domestic market and to figure out a successful way to operate in foreign markets. Finding the right internationalization strategy to be successful abroad, has a considerable impact on firm performance and is often difficult to change or correct after internationalization decisions are made (Brouthers & Hennart, 2007). Due to globalization, regionalism is considered increasingly important, as countries become more vulnerable and try to secure economic development and stability by establishing bonds within the region (Kim & Shin, 2002)
The brand portfolio strategy, what brands a company decides to offer in its current brand portfolio, is an important decision when expanding abroad (Geringer, Beamish & DaCosta, 1989; Douglas, Craig & Nijssen, 2001). It can provide both opportunities and challenges for MNEs, since all factors of operating in multiple countries, such as exploiting cost benefits and meeting demands of heterogeneous countries, have to be considered to be successful (Geringer, Beamish & DaCosta, 1989). Brands are a critical factor of a company’s international marketing strategy to establish a strong brand identity and customer base (Douglas, Craig & Nijssen, 2001). Due to globalization, markets became more integrated and MNEs have to put greater
effort in organizing their brands, coordinating their marketing efforts abroad and integrating activities worldwide (Douglas, Craig & Nijssen, 2001).
In previous research, brand diversification and internationalization are often considered as different strategies to pursue growth opportunities or increase a firm’s financial performance (Cantwell, Gambardella & Granstrand, 2004). Penrose (1959) described them as alternative routes to growth that are interrelated but seen as a strategic choice between two means. Many scholars focused on the relationship between brand diversification and firm performance, as well as the relationship between the degree of internationalization and firm performance or a combination (Chang & Wang, 2007; Hitt, Hoskisson & Ireland, 1994; Hitt, Hoskisson & Kim, 1997; Morgan & Rego, 2009; Tallman & Li, 1996; Zhao & Luo, 2002). This prior research on diversification strategies and performance outcomes did not lead to conclusive results and findings are often contradictory (Chang & Wang, 2007; Tallman & Li, 1996). Recently, a new perspective has emerged, that foresees a direct relationship between a firm’s diversification strategy and its degree of internationalization (Cantwell, Gambardella & Granstrand, 2004).
Little is known about the influence of host country contexts and whether they have an impact on strategic decisions (Chang & Wang, 2007). Host country characteristics include investment restrictions, political risk, economic development, access to capital, cultural traits and other factors that a company encounters when it expands to another country (Harzing & Pudelko, 2016). Meyer, Mudambi and Narula (2011) state that host country contexts, despite increasing globalization, remain important to consider for an MNE to be successful abroad. Global homogenization applies to a certain extent, but it remains important to adapt to regional, country, cultural and societal differences to be able to internalize market transactions, exploit resources and capabilities from the host country contexts and create a competitive advantage (Meyer, Mudambi & Narula, 2011). Until now, distance-based approaches such as cultural distance, dominated International Business research and are used to explain strategic decisions
made by MNEs. Furthermore, the exploratory power of home and host country contexts are underestimated. Studies in International Business should focus more on home and host country context, instead of using (cultural) distance as a catchall concept (Harzing & Pudelko, 2016). Contextual rich data is relevant to better understand countries, companies and the business system in general and it makes it possible to gain understanding of this deeply embedded contexts (Harzing & Pudelko, 2016; Shenkar, 2004)
Most studies focused mainly on single-nation studies, developed countries and U.S. firms (Erdal & Tatoglu, 2002; Geringer, Beamish & DaCosta, 1989; Morgan & Rego, 2009). The operations strategies of Asian MNEs differ from Western MNEs and theory development is limited (Sim & Pandian, 2003). The Asian market is very diverse in terms of host country context characteristics and stages of economic development, which makes it interesting to compare with American and European MNEs (Hoskisson et al., 2000; Sim & Pandian, 2003).
This study tries to answer the question whether MNEs with a broad domestic brand portfolio have a higher degree of regionalization compared to MNEs with a narrow domestic brand portfolio. Furthermore, it investigates to what extent this relationship is moderated by host country context characteristics. To answer this question, a sample of 45 MNEs operating in the Asian beer, spirits and wine markets is examined. The relationship between brand portfolio breadth and the degree of regionalization is tested through multiple regression analyses and economic openness, infrastructure quality and economic instability are added as host country context characteristics to test whether they moderate this relationship.
This thesis contributes to the International Business literature in several ways. First, by focusing on the direct relationship between the breadth of the domestic brand portfolio and the degree of regionalization, as prior research mainly focused on the relationship between the separate variables and firm performance or as moderators. Second, the Asian alcoholic beverages market is examined, to extend prior research that primarily focused on western
developed countries like the U.S. and Japan. This could provide new insights for brand portfolio strategy in these countries. Third, host country context characteristics are included to see whether they moderate the before mentioned relationship, extending the literature and to possibly help managers in future brand portfolio decisions.
This thesis is structured as follows. In the next section, the relevant literature about brand portfolio strategy, regionalization and host country context is discussed and a theoretical framework is developed. Then, the data, sample and model are explained in the methodology section, whereafter the results are presented. Subsequently, the results, implications, contributions and limitations are discussed in the discussion section. Finally, the thesis ends with a conclusion.
2. LITERATURE REVIEW
Due to globalization and increasing competition, MNEs expanded the geographic scope of their operations and started to trade with countries outside their home country, often by opening subsidiaries or acquiring other companies abroad (Douglas, Craig & Nijssen, 2001; Dunning, 1988). At the same time, improvements in the quality of global infrastructure, communication technology and transport possibilities led to a more integrated global market. Therefore, the marketing strategy should be adjusted to achieve an integrated marketing strategy that fits multiple countries (Douglas, Craig & Nijssen, 2001).
A MNE has several options to coordinate and integrate their marketing efforts in the countries in which they are active. Focusing on the products, the company can choose to diversify or differentiate its product or brand portfolio. Product diversification can be defined as the way a company carries out its business activities and composes its product portfolio at foreign subsidiaries compared to the parent company (Zhao & Luo, 2002). Companies can choose to focus on one main line of business or to expand product portfolio’s and product
markets (Mudambi & Mudambi, 2002). A MNE can decide to diversify its product portfolio to lower and spread the risk of failure by offering more different products (Hitt, Hoskisson & Ireland, 1994). Managing MNEs with diversified product portfolio’s and operating in foreign markets has become increasingly complex and innovation is essential to maintain a sustainable competitive advantage and to be successful abroad (Porter, 1990). Besides the product portfolio, another important element in a company’s marketing strategy is the composition of the brand portfolio and the way a company structures this brand portfolio across borders (Douglas, Craig, Nijssen, 2001). A good branding policy consisting of strong brands can help a company to establish the company abroad and to develop a strong customer base. This becomes increasingly important, as international markets become more interlinked and companies have to consider the coherence of their overall brand portfolio strategy (Aaker, 1996; Douglas, Craig & Nijssen, 2001). This thesis focuses on the brand portfolio strategy of a company.
2.1 Brand portfolio strategy
Which brands and the number of brands a MNE decides to hold in its brand portfolio has important implications for the company, as they determine the visibility and position in the markets a company is active in (Douglas, Craig & Nijssen, 2001). Brand portfolio decisions depend on both firm and market-level factors, like the availability of internal and external resources, the stage of internationalization and the competitors (Townsend, Yeniyurt & Talay, 2009). A brand comprises all expectations and associations that are linked to experiencing company offerings (Petromilli, Morrison & Million, 2002). In marketing literature, brands are considered intangible resources that can significantly contribute to the performance of the company (Hunt & Morgan, 1995; Petromilli, Morrison & Million, 2002; Shah, 2015). The brand portfolio or brand structure refers to the current set of brands a company has in its portfolio across countries, businesses and product markets (Douglas, Craig & Nijssen, 2001). In contrast, several scholars use the concept brand architecture, which refers to the strategy
behind the brands, how they are rationalized and how they are applied if the company goes to market (Douglas, Craig & Nijssen, 2001; Petromilli, Morrison & Million, 2002). For the brand portfolio strategy this thesis focuses on the brand structures, explained as the number of brands a company has in its brand portfolio. How this brand portfolio is composed, depends on several factors. Companies make decisions concerning which brands they add to their brand portfolio or which brands should be retained, modified or discarded (Shah, 2015).
Companies can choose to follow either a global brand strategy or to adjust to local markets and compose a brand portfolio that fits local and regional needs (Schuiling & Kapferer, 2004). This is often a tradeoff between having several brands for different local markets or standardizing the brand portfolio, making few or no adjustments to local markets. Global brands are increasingly considered as a good way to achieve economies of scale and give MNEs the opportunity to reach an increasing group of consumers by creating a well-known identity and setting a certain quality standard that meets the expectations of existing customers (Hsieh, 2002). By having local brands in the brand portfolio in addition to global brands, MNEs can spread the risk and better serve consumers, since the brands are adjusted to the host country context (Schuiling & Kapferer, 2004).
Bordley (2003) states that having a highly diverse brand portfolio enables a firm to serve different consumer groups more precisely and it deters potential competitors to enter the market, since it becomes costlier to enter the specific market. A company should be sure that the broader brand portfolio serves the needs of the different customers sufficiently and that brands do not cannibalize each other (Petromilli, Morrison & Million, 2002). Aaker (1996) suggests that offering several brands across different markets and managing them independently might cause problems allocating the needed resources among the brands, which in the end affects the overall performance of the company. The ideal and effective number of brands in the portfolio is related to the overall profit of the company (Bordley, 2003). How a company integrates its brand
portfolio strategy across borders determines the overall success, this includes the choice between using the same brand worldwide or creating a common image for every country market a company is active in (Douglas, Craig & Nijssen, 2001). Often companies make these decisions on a country basis, which might cause incoherence in the overall brand strategy, but all brands serve different customer segments and target markets (Douglas, Craig & Nijssen, 2001). The general view of having a broad brand portfolio is that it is costlier, but the chance of succeeding abroad increases, as a company is better able to serve heterogeneous countries with different customers (Aaker, 1996; Bordley, 2003; Douglas, Craig & Nijssen, 2001).
A narrow brand portfolio often lowers overall costs of the brand portfolio, due to economies of scale, and it is often easier to manage (Bordley, 2003). Another reason to hold fewer brands is that if a company decides to focus on a limited number of brands, the quality of might increase (Bordley, 2003). Economies of scale can lead to cheaper production and increased profits, because companies can focus on a limited number of brands. Fewer brands lower the marketing costs, as well as it is easier to enter new foreign markets if a company decides to offer its complete brand portfolio (Schuiling & Kapferer, 2004). If some brands are not profitable or successful for the company, the management should consider discarding the brand and focus on the brands that are successful to save costs (Petromilli, Morrison & Million, 2002; Shah, 2015).
MNEs have a portfolio of brands where they make decisions about how to position the brands and how to deal with scope economies and competition. These decisions have an impact on both marketing and financial performance. Appropriate decisions depend largely on the performance goals that a MNE sets for itself (Morgan & Rego, 2009). This concerns both the marketing strategy in the home country, as well as the strategy it uses across borders. A company can choose to introduce its existing brands or to acquire brands from companies that are already active in the market abroad. Companies that expand through introducing brands that
are already successful in the domestic market often market their brands at the product level (Douglas, Craig & Nijssen, 2001).
2.2 Regionalization
Globalization of markets and a business environment characterized by global competition, fueled by decreased transportation costs and technological developments cause that companies expand their offerings across borders (Dunning, 1988). Expanding internationally provides an opportunity to capitalize on the success a company has in its domestic market (Hitt, Hoskisson & Ireland, 1994). Companies first introduce and exploit brands in the domestic market before it decides to expand abroad. This often starts with exporting the products and gradually moves towards foreign direct investments and subsidiaries. The MNE tries to leverage the expertise and experience it already possesses from the domestic activities (Hitt, Hoskisson & Ireland, 1994; Vernon, 1966).
Several scholars found a positive relationship between international diversification and firm performance (Rugman, 1979; Geringer, Beamish & DaCosta, 1989). In addition, it would stabilize returns, as companies operate in different geographic markets with other characteristics or stages in comparison to the domestic market (Hitt, Hoskisson & Ireland, 1994). MNEs are better able to exploit resources of the different countries they are active in and a big international network increases flexibility, but also improves the bargaining power of the MNE (Kogut, 1984). Johanson and Vahlne (1977) state that for the performance of marketing activities, both firm and market experience is required, which means that the to be successful abroad, employees should possess this knowledge and apply it to operations. They also agree that internationalization is a gradual process and companies gather specific knowledge in their domestic market to be able to apply this knowledge elsewhere (Johanson & Vahlne, 1977). The traditional route to expand abroad is to enter countries where the psychic distance between the home and the host country is smaller. Psychic distance includes different
languages, variations in educational systems, culture and level of economic development (Johanson & Vahlne, 1977). Brand managers are expected to lack the needed knowledge to expand abroad, as the knowledge it acquired in the domestic market often is less valuable abroad (Johanson & Vahlne, 1977). On the other hand, Cuervo-Cazurra (2010), claims that companies develop specific knowledge in the domestic market that can be used to overcome the difficulties of foreign expansion if it has experience with managing multiple operations. He claims that foreign expansion is not only a gradual learning process and that a firm can prepare starting in their domestic market. This includes managing complex operations and transferring local resources to a new country market (Cuervo-Cazurra, 2010).
Between 1959 and 1996 the world became increasingly globalized (Kim & Shin, 2002). This globalization process caused markets became more integrated and it increased the vulnerability of the companies that are active in these markets. This is why companies try to cooperate and form bonds with companies within the region to secure a certain level of economic development and stability (Hurrell, 1995; Kim & Shin, 2002). Regionalism might be interesting for companies to increase the negotiation power with countries outside the region, as you become less vulnerable (Kim & Shin, 2002). Dicken (1992) on the other hand claims that regionalization might cause protectionism and hinders free trade between countries, as companies are more connected to the region.
2.3 Host country context
Interesting might be to see whether the number of brands within the domestic brand portfolio and the degree of regionalization are influenced by the host country context. Host country context and differences between countries and cultures are very important in International Business research. The host country context can be defined as the specific aspects of a country that distinguishes it from other countries and that explain the difference between international and domestic business (Meyer, Mudambi & Narula, 2011). Characteristics that define the host
country context include investment restrictions, political risk, economic development, access to capital, cultural traits and other factors that a company encounters when it expands to another country (Harzing & Pudelko, 2016). Although some scholars expect that globalization will cause a shift to centralized global strategies, Meyer Mudambi and Narula (2011) expect that local contexts will most likely become more important and companies that best succeed to respond to local values and needs will perform best. Wan and Hoskisson (2003) found that home country context influences the decisions concerning diversification strategies. An unstable political environment or cultural differences for example could lead to different entry and investment decisions.
Harzing and Pudelko (2016) state that prior studies have focused mainly on the impact of distance dimensions on empirical relationships and phenomena and stress the importance of home and host country contexts within International Business research. Contextually rich and thick data is needed to gain contextual information and better understand how companies and countries interact within a domestic market (Shenkar, 2004; Tsui, 2007). Instead of focusing on academic concepts such as distance, it would be interesting to consider variables that can be of value for managers operating in foreign markets. These variables might have more explanatory power and could guide managers in future decisions regarding their regionalization and brand portfolio strategies (Harzing & Pudelko, 2016).
MNEs operating in different contexts that have to cope with large cultural, economic and psychic distance may benefit in particular from understanding host country contexts (Meyer, Mudambi & Narula, 2011). A MNE must adapt its organization to the host country context to be successful in the country they are operating in. The attractiveness of a host country context largely depends on the costs and benefits that occur during the adaptation process (Meyer, Mudambi & Narula, 2011). Both formal and informal institutions set rules and determine how MNEs interact with each other. Additionally, they have considerable influence
on the costs and benefits MNEs encounter by operating in that specific context (Rodrik, Subramanian & Trebbi, 2004). These local institutions and their attitude towards MNEs partly affect the location choice and thus leads to competition between institutional systems. Host country contexts that are more open and transparent are more attractive for MNEs and are better able to achieve growth (Rodrik, Subramanian & Trebbi, 2004). Another dimension of host country context is the availability and accessibility of resources. An important part of entering a new host country context is aligning firm-specific and location-specific resources (Dunning, 1988).
MNE headquarters and subsidiaries interact with both home and host country contexts and that leads to complex relationships between these MNEs and the local contexts and institutions (Forsgren, Holm & Johanson, 2007). Meyer, Mudambi and Narula (2011) provide three perspectives on how MNEs interact with their local context. First, MNEs are shaped by their home country context, which determines how it operates in a foreign market. Second, MNE subsidiaries are embedded in both the host country context and the MNE network. Therefore, it is exposed to both institutional pressures and the parent. MNEs have to adapt their strategy to this pressures within the host country context. Third, it is important to see how MNEs cope with the different contexts of their subsidiaries (Forsgren, Holm & Johanson, 2007; Meyer, Mudambi & Narula, 2011).
Erdal and Tatoglu (2002) have researched the relationship between several location-related determinants or host country context variables to test how they influence the level of foreign direct investments (FDI) in a country. If a company decides to expand abroad it encounters several factors of the host country. These factors for example include the natural resources of the country and the availability of labor, but also the business environment, which concerns the political, economic and infrastructural factors (Erdal & Tatoglu, 2002). Erdal and Tatoglu (2002) expected that economic openness, infrastructure and economic instability all
have an impact on FDI and found positive outcomes for economic openness and infrastructure. The economic openness or trading regime of a host country and the willingness to provide a decent infrastructure are key determinants to be successful in attracting investments from abroad (Erdal & Tatoglu, 2002; Vernon, 1966). In an open economy import and export are an important part of a country’s GDP and it becomes easier to export the products of the company to the specific country. In addition, better infrastructure facilitates the distribution throughout the country and makes it possible to reach more potential customers (Erdal & Tatoglu, 2002). Nordås and Piermartini (2004) found that the quality of infrastructure had a significant and a relatively large impact on the bilateral trade flows between countries. They also found that increased infrastructure not always make up for distance between countries and there is a positive link between infrastructure quality and transportation costs (Nordås & Piermartini, 2004). If less developed countries decide not to improve their infrastructure this will most likely lead to a further decline of their share in regional trade (Nordås & Piermartini, 2004). Economic instability increases both the risk and the cost of capital for the company, as it become more expensive to borrow money (Erdal & Tatoglu, 2002).
2.4 Research gap
In prior research, scholars have mainly focused on the relationship between product, brand or geographic diversification and firm performance, but findings are inconclusive, as both positive, negative and neutral outcomes were found (Chang & Wang, 2007; Hitt, Hoskisson & Ireland, 1994; Hitt, Hoskisson & Ireland, 1994; Hitt, Hoskisson & Kim, 1997; Morgan & Rego, 2009; Tallman & Li, 1996; Zhao & Luo, 2002). Others focused on cultural distance (Harzing & Pudelko, 2016). Cantwell, Gambardella and Granstrand (2004) suggested that these independent variables could be interrelated and foresee a direct relationship, apart from firm performance. This thesis tries to contribute to International Business literature by focusing on the relationship between a firm’s brand portfolio breadth and the degree of regionalization of a
MNE. In addition, I will examine the influence of a country’s host country context on the relationship between a company’s brand portfolio breadth and the degree of regionalization in the Asian alcoholic beverages market. This because as mentioned before, the emphasis used to be on American and EU companies, but developing markets in Asia are considered interesting for theory developing (Sim & Pandian, 2003). This leads to the following research question:
Do MNEs with a broad domestic brand portfolio have a higher degree of regionalization compared to MNEs with a narrow domestic brand portfolio and to what extent is this relationship moderated by host country context characteristics?
3. HYPOTHESES DEVELOPMENT
In this chapter the hypotheses are defined and explained. The following part will provide a theoretical justification for every hypothesis. In total four hypotheses are developed and tested in this thesis. The dependent variable is the degree of regionalization and the independent variable is the domestic brand portfolio breadth.
Companies make brand portfolio strategy decisions that best suit their projected target groups and can choose between having few global brands or adjusting to the local markets they operate in (Schuiling & Kapferer, 2004). Before a company decides to expand abroad, it should be successful first in the domestic market to be able to capitalize on the success the company already established in the home country (Hitt, Hoskisson & Ireland, 1994). This often starts with exporting the products in the domestic brand portfolio (Vernon, 1966). When companies first go abroad, it is assumed that they often lack the needed knowledge to do so and it is seen as a gradual learning process (Johanson & Vahlne, 1977). Firms can prepare for regional expansion and might benefit from having experience managing a complex domestic market with multiple brands (Cuervo-Cazzura, 2010). In addition, maintaining a broad brand portfolio provides the company with more options to satisfy heterogeneous groups of consumers, as it
will be able to adjust the brand to meet divergent expectations (Bordley, 2003). If a company has a broad domestic brand portfolio it might also discourage competitors to pursue an expansion strategy (Bordley, 2003). Most scholars suggest that having a broad brand portfolio is costlier, but it increases the chance of being successful in different markets, as potential customers can choose what they like best (Aaker, 1996; Douglas, Craig & Nijssen, 2001).
Hence, I expect that if a company has a broad domestic brand portfolio, the likelihood of being successful abroad increases, as the company can meet the demands of different customers. This would mean that a company is active in more countries and thus increases the degree of regionalization. This leads to the first hypothesis:
H1: There is a positive relationship between the breadth of a firm’s domestic brand portfolio and a firm’s degree of regionalization.
The other hypotheses test the influence of the host country context on the main relationship between the breadth of the domestic brand portfolio and the degree of regionalization. By testing for a possible moderating effect of host country context determinants, I would like to test which exact determinants influence this relationship and in what way. In this thesis I test the moderating effect of economic openness, infrastructure quality and economic instability, as these determinants are considered important (Erdal & Tatoglu, 2002; Nordås & Piermartini, 2004; Vernon, 1966). These determinants explain the differences between doing business in the domestic market in comparison to the market that a company would like to enter (Meyer, Mudambi & Narula, 2011). Understanding these concepts are especially interesting for companies operating in different countries in Asia, where cultural, economic and psychic distance is often considerable (Meyer, Mudambi & Narula, 2011).
The economic openness says something about the import and export indicators in a country and provides information about the convenience of doing business with a country. A
high economic openness means that a country is not only focused on the domestic market, but laws and other circumstances create a favorable business climate, for example caused by favorable trading tariffs (Erdal & Tatoglu, 2002; Vernon, 1966). In this way it becomes easier for companies to export products if countries have a higher economic openness. Infrastructure quality has a similar contribution to creating a favorable business environment, as companies are better able to distribute their products across the country and it increases the reach of possible customers within a country (Erdal & Tatoglu, 2002; Nordås & Piermartini, 2004). As Johanson and Vahlne (1977) describes regionalization as a gradual learning process for companies, I expect that MNEs first expand to countries that are easy to access and have a favorable business environment. In this way the overall economic openness and infrastructure quality of the combined host countries will decrease as the degree of regionalization increases. In this way the relationship between the domestic brand portfolio breadth and the degree of regionalization will be weaker if economic openness and infrastructure quality of the combined host countries increases and thus turn out to be negative moderators. This leads to the following hypotheses:
H2: Higher economic openness of host countries negatively moderates the relationship between brand portfolio breadth and the degree of regionalization.
H3: Higher infrastructure quality of host countries negatively moderates the relationship between brand portfolio breadth and the degree of regionalization.
Economic instability is expected to have the opposite effect, as a higher value leads to a higher risk and an increased cost of capital for the company (Erdal & Tatoglu, 2002). I expect that companies expand to companies with lower economic instability first, as the business environment is more favorable and only later decide to expand to countries with higher instability (Johanson & Vahlne, 1977). Hence, I expect higher economic instability for the
combined host countries of the companies with a higher degree of regionalization. Therefore, I expect a positive moderating effect, which this leads to the fourth hypothesis:
H4: Higher economic instability of host countries negatively moderates the relationship between brand portfolio breadth and the degree of regionalization.
The following figure shows the conceptual model with domestic brand portfolio breadth as main independent variable, the degree of regionalization as the dependent variable and the host country context as proposed moderators. These include economic openness, infrastructure quality and economic instability.
Figure 1. Conceptual model
4. METHODOLOGY
The research approach adopted for this thesis is quantitative. This chapter presents information about the data and the sample used in this thesis. The variables, the empirical strategy and the model to test the hypotheses are explained next.
4.1 Data and sample
To test the relationship between brand portfolio breadth and the degree of regionalization, this thesis focuses on the Asian alcoholic beverages market. Traditionally, scholars have focused on Western and Japanese MNEs when developing theory and explaining internationalization strategy (Geringer, Beamish & DaCosta, 1989). Emerging Asian MNEs became increasingly
interesting for theory development, as their operations strategies differ from Western MNEs and little research is done on the matter (Sim & Pandian, 2003). Furthermore, the Asian market is heterogeneous in terms of for example culture and economic development and includes all developed, developing and transition economies (Hoskisson et al., 2000). Examining the impact of host country context is therefore more interesting. Further research examining Asian MNEs is necessary to explain differences with American and European MNEs (Sim & Pandian, 2003). Asian MNEs are MNEs that have their domestic market in Asia and are operating in multiple countries inside Asia. This setting is used to test which brand portfolio strategy might be more successful to operate within the Asian market. This could also provide managers with new insights of doing business in Asia and making decisions concerning brand portfolio and internationalization strategies.
The research is conducted using secondary quantitative data collected from the databases Passport, Bureau Van Dijk’s Orbis, the World Trade Organization, the World Bank and the CIA World Factbook.. The sample consists of Asian MNEs operating in the alcoholic beverages market, specialized in beer, spirits and wine. The main dataset is retrieved from the Euromonitor Passport Global Market Information Database and covers the period between 2006 and 2015. The database includes information about brand names, home country, host country, market share and sales. Passport is a market research database, providing detailed information on several countries, economies and markets to analyze trends and market contexts (Euromonitor International, 2017). The database has been used in prior market research in an international context (Spicka, 2016). This thesis uses Passport for data of the Asian beer, spirits and wine market, which makes it possible to compare different countries in Asia. The countries that are included in this research are China, Hong Kong, India, Indonesia, Japan, Kazakhstan, Korea, Malaysia, Philippines, Singapore, Taiwan, Thailand and Vietnam. These countries are linked to the companies in the existing database, so I did not select the countries myself. Data
is available for 105 companies active in the Asian market. After excluding incomplete data and data of companies that have their domestic market outside of Asia, the final sample for this thesis consists of 45 companies. MNEs that have their headquarters outside of Asia or companies that only operate in their domestic market are excluded from the sample. A lot of the companies in the dataset were European or American and were thus excluded. Data on host country context is obtained from the World Trade Organization, the World Bank and the CIA World Factbook. This data includes economic openness, infrastructure quality and economic instability, as used in prior research (Erdal & Tatoglu, 2002; Nordås & Piermartini, 2004; Kumar, 2006).
4.2 Dependent variable
The dependent variable is the degree of regionalization, which is measured by the number of countries a company is active in within the Asian market (Calof, 1993). This sample only consists of companies that have their domestic market in Asia and operate within the Asian market. Home and host country information is retrieved from the Euromonitor International Passport database and includes both the company’s home country and all the countries in Asia the company is active in. The number of countries provides a measure of the degree of regionalization.
4.3 Independent variable
The independent variable is the breadth of the domestic brand portfolio, which is the number of brands in the domestic brand portfolio. The number of brands is frequently used as a measure of portfolio breadth (Dacin & Smith, 1994; Morgan & Rego, 2009). I decided to only consider the brands in the domestic brand portfolio. Normally these are the brands they use to introduce to other markets and to avoid including brands multiple times, as companies often use the same brand abroad. The data is retrieved from the Euromonitor International Passport database, by counting the number of brands the companies have in the brand portfolio in their home country.
4.4 Moderators
This To test whether host country context has an impact on the relationship between brand portfolio breadth and the number of markets served, this thesis includes economic openness, infrastructure quality and economic instability (Erdal & Tatoglu, 2002; Nordås & Piermartini, 2004; Kumar, 2006). The Economic openness is measured by dividing the total merchandise trade, import plus export, by the country’s GDP. Import and export data is found on the World Trade Organization database. The quality of the infrastructure is measured by constructing a measure, consisting of several infrastructure indicators (Nordås & Piermartini, 2004). These include fixed telephone subscriptions per 100 people, kilometers of paved roads per square kilometer, kilometers of railways per square kilometer and airports per 1000 square kilometer. The data on infrastructure is retrieved from the World Bank and CIA World Factbook databases. All indicators are normalized by the mean, added up and the average is taken to create a measure for overall infrastructure quality (Nordås & Piermartini, 2004). Economic instability is measured by the real interest rate (Erdal & Tatoglu, 2002).
4.5 Control variables
The GDP per capita of the home country (GDP Home), the market size of the home country (MS Home), the market size of the host country (MS Host) and both the size (Firm Size) and the age (Firm Age) of the firm are used as control variables to account for differences among the observed companies and the markets in which they are active that might cause differences in the brand portfolio breadth or the degree of regionalization.
GDP per capita data for all the countries are retrieved from the World Bank database. The natural log is used in the analyses. This measure is used to control for differences in purchasing power between the countries, which might influence the number of brands a company has in its domestic brand portfolio (Hsieh, 2002). It is expected that a higher GDP per
capita indicator has a positive outcome on the number of brands a company has in its domestic brand portfolio.
Market size is measured in millions of US dollars at current prices and retrieved from the Euromonitor International Passport database (Erdal & Tatoglu, 2002). For the control variables, the natural logs are used in the analyses. The size of the market provides information about the possibilities to sell your products and the demand of the products you have in your brand portfolio (Talay, Townsend & Yeniyurt, 2015). The size of the host market is also perceived as a pull factor to invest in a country. A bigger market size often leads to an increase in the demand of a company’s products and services and could ultimately help to achieve economies of scale (Erdal & Tatoglu, 2002). I therefore expect that both MS Home and MS Host are positively correlated with the independent variable Brands and the dependent variable Countries.
The data on firm size and firm age are retrieved from Bureau Van Dijk’s Orbis database. Firm size is measured by the firm’s total assets and the natural log is computed to control for differences in firm size and the access to capital and scale economies of the company that are not measured by market size (Morgan & Rego, 2009). Firm age is measured by the number of years since the company’s incorporation. Both control variables are expected to be positively correlated to both the number of brands in the domestic brand portfolio and the degree of regionalization.
For this thesis the companies are observed from 2012 to 2015, therefore I created a dummy variable for each year to control for differences between years. The dummy has a value of 1 if it concerns the specific year and a value of 0 for any other year. I control for yearly differences to make sure no errors are made throughout the years (Chang & Wang, 2007). All the variables are listed and explained in Table 1.
Table 1. Variables
Code Definition Source
Brands Number of brands in the
domestic brand portfolio (breadth)
Euromonitor International Passport
Countries Number of countries served
outside the home country (degree of regionalization)
Euromonitor International Passport
GDP Home Gross domestic product per
capita of home country
World Bank
MS Home Market size in US$ of home
country
Euromonitor International Passport
MS Host Market size in US$ of host
country
Euromonitor International Passport
Firm Size Size of the firm:
total assets
Van Dijk’s Orbis
Firm Age Age of the firm
in years
Van Dijk’s Orbis
Openness Economic openness: sum of
import and export divided by GDP
World Trade Organization, World Bank
Infrastructure Overall infrastructure quality: construct of fixed telephone subscriptions, paved roads and airport density
CIA World Factbook, World Bank
Instability Economic instability: real interest rate
World Bank
4.6 Empirical strategy
To examine the relationship between brand portfolio breadth and the degree of regionalization, the descriptive statistics, a correlation matrix and multiple negative binomial regression models are computed. This will provide information about the dispersion of the variables and helps to detect highly correlated variables and possible errors in the data. As the dependent variable is a count variable, a negative binomial regression model is used, instead of a normal Poisson model, because the mean does not equal the variance. To test whether host country context
moderates the before mentioned relationship, the interaction terms for Openness, Infrastructure and Instability are added to the model. Multiple years (2012-2015) are studied to smoothen annual errors in the data (Chang & Wang, 2007). IBM SPSS Statistics is used for the analyses and to test the hypotheses.
4.7 Model
Degree of regionalization = b0 + b1 * Brands + b2 * GDP Home + b3 * MS Home + b4 * MS
Host + b5 * Firm Size + b6 * Firm Age + b7 * Year + b8 * Openness + b9 * Infrastructure +
b10 * Instability + e
5. RESULTS
After performing all the statistical tests in SPSS, the results are presented in this chapter. The descriptive statistics of all the variables, except the Year dummy variables are included in Table 1. The correlation matrix is presented in Table 2 and shows the bivariate correlations between the dependent variable and all the independent variables, including control variables. The regression analysis is presented in Table 3.
I ran a frequency test to see if there were any errors in the data and no errors were found. To check if there were any multicollinearity issues in the dataset, the variance inflation factor (VIF) was used. The stability of the parameter estimates is not significantly influenced if the VIF is lower than 10 (Dielman, 1991). This was the case for all the independent variables, as the VIF ranged between 1,153 and 2,358.
Table 2. Descriptive statistics
N Minimum Maximum Mean Std. Deviation
Brands 180 1,00 20,00 5,1333 5,28347 Countries 180 1,00 11,00 2,4667 2,36926 GDP Home 180 7,28 10,94 9,9912 ,93999 MS Home 180 6,03 11,31 8,9399 1,52898 MS Host 180 5,55 11,92 8,4536 1,74670 Firm Size 180 4,33 17,46 13,6958 2,89137 Firm Age 180 1,00 119,00 45,3889 35,11217 Openness 180 ,27 4,20 1,4213 1,03708 Infrastructure 180 ,28 4,60 1,6889 1,15470 Instability 180 -,88 8,08 2,9872 1,70129 Valid N (listwise) 180 5.1 Descriptive statistics
The descriptive statistics show that there are no missing values for any of the variables used in the analysis. Every variable has a valid N of 180. The breadth of the domestic brand portfolio of the companies in the dataset varies between 1 and 20 brands, with on average 5,1333 brands in the portfolio. The companies are active in at least 1 country other than their home country and a maximum of 11 other countries. The mean of 2,4667 and the frequency test show that there is a limited number of companies that serve more different countries. GDP Home (GDP per capita) and MS Home and MS Host are measured in US dollar. The descriptive statistics show a high variability in Firm Size, which proves the importance to control for these differences. The values of GDP Home, MS Home, MS Host and Firm Size, listed in the table, are natural logs. The year dummies are not listed in Table 2, but the results show that no mistakes are made in constructing the year dummy variables 2012, 2013, 2014 and 2015, as they have a minimum of 0 and a maximum of 1 and the frequency test did not show any other values.
5.2 Correlations
The correlation matrix in Table 3 shows the Pearson Correlations and the two-tailed significance of all the variables used in this thesis. This table was used to check if there were any variables that were too highly correlated (Pearson Correlation > ,700). This and the VIF test show that there are no correlations that are too high and show that there are no multicollinearity problems.
All dependent, independent and control variables are listed in Table 3. Several significant correlations are found. The independent variable Brands is positively and significantly correlated with the dependent variable Countries (,495**). This indicates that the number of brands in the domestic brand portfolio moves in the same direction as the degree of regionalization. This could mean that companies with a broader domestic brand portfolio are active in more countries or that companies that are active in more countries have a broader domestic brand portfolio. GDP Home is positively and significantly correlated with Brands (,177*). A possible interpretation could be that if the GDP per capita in a country is higher, the more brands a company has in its product portfolio, as people have more money to spend. MS Home is positively and significantly correlated with Brands (,356**). This shows that market size of the home country relates to the breadth of the domestic brand portfolio. MS Host is both positively and significantly correlated at the 0,05 level with Brands (,173*) and Countries (,149*), which suggests that a bigger market size of the host country relates to more brands in the domestic brand portfolio and a higher degree of regionalization and the other way around. Firm Size is positively and significantly correlated with both Brands (496**) and Countries (,207**), which indicates that bigger firms have more brands in their brand portfolio and operate in more countries and that the opposite goes for smaller firms. Surprisingly, Firm Age is not correlated with Brands, Countries or Firm Size. The moderator variables Openness,
Infrastructure and Instability move in the same direction, but the correlations are below 0,7 and should not cause any problems concerning the reliability of the model.
Table 3. Correlation matrix
1. 2. 3. 4. 5. 6. 7. 8. 9. 10.
1. Brands Pearson Correlation 1
Sig. (2-tailed)
2. Countries Pearson Correlation ,495** 1 Sig. (2-tailed) ,000
3. GDP Home Pearson Correlation ,177* -,020 1
Sig. (2-tailed) ,017 ,789
4. MS Home Pearson Correlation ,356** -,118 ,073 1
Sig. (2-tailed) ,000 ,115 ,330
5. MS Host Pearson Correlation ,173* ,149* ,038 -,348** 1
Sig. (2-tailed) ,020 ,046 ,613 ,000
6. Firm Size Pearson Correlation ,496** ,207** ,108 ,307** ,347** 1
Sig. (2-tailed) ,000 ,005 ,147 ,000 ,000
7. Firm Age Pearson Correlation ,031 -,141 ,048 ,315** -,199** ,017 1
Sig. (2-tailed) ,677 ,060 ,520 ,000 ,007 ,821
8. Openness Pearson Correlation -,067 -,032 -,194** ,376** -,634** -,145 ,140 1
Sig. (2-tailed) ,375 ,670 ,009 ,000 ,000 0,053 ,060
9. Infrastructure Pearson Correlation -,114 -,189* -,227** ,160* -,495** -,106 ,289** ,588** 1
Sig. (2-tailed) ,127 ,011 ,002 ,032 ,000 ,157 ,000 ,000
10. Instability Pearson Correlation ,139 ,277** -,267** -,318** ,041 -,151* -,072 ,153* ,156* 1
Sig. (2-tailed) ,062 ,000 ,000 ,000 ,582 ,043 ,340 ,040 ,036
**. Correlation is significant at the 0.01 level (2-tailed). *. Correlation is significant at the 0.05 level (2-tailed).
Table 4. Negative binomial regression
Model 1 Model 2 Model 3 Model 4 Model 5
Variables Beta Wald Sig. Beta Wald Sig. Beta Wald Sig. Beta Wald Sig. Beta Wald Sig. (Intercept) ,041 ,003 ,960 3,514 16,776 ,000 4,093 21,434 ,000 4,095 21,778 ,000 3,467 16,167 ,000 GDP Home ,012 ,052 ,819 -,146 6,944 ,008 -,179 9,818 ,002 -,206 12,565 ,000 -,144 6,710 ,010 MS Home -,107 5,923 ,015 -,269 33,285 ,000 -,285 39,365 ,000 -,282 39,258 ,000 -,268 33,066 ,000 MS Host -,064 2,234 ,135 -,048 1,321 ,250 -,066 2,379 ,123 -,019 ,185 ,667 -,048 1,305 ,253 Firm Size ,134 31,990 ,000 ,066 7,142 ,008 ,082 10,479 ,001 ,062 6,319 ,012 ,068 7,430 ,006 Firm Age -,001 ,369 ,544 ,001 ,152 ,697 ,001 ,212 ,645 ,002 2,010 ,156 ,001 ,189 ,664 Openness -,189 6,452 ,011 -,300 14,838 ,000 -,186 4,489 ,034 -,325 17,161 ,000 -,299 14,901 ,000 Infrastructure -,297 22,534 ,000 -,302 21,402 ,000 -,279 18,670 ,000 -,360 29,089 ,000 -,306 21,676 ,000 Instability ,207 31,182 ,000 ,062 2,222 ,136 ,059 2,062 ,151 ,045 1,202 ,273 ,070 2,752 ,097 2013 -,121 ,792 ,374 -,052 ,146 ,702 -,060 ,197 ,657 -,049 ,129 ,719 -,063 ,211 ,646 2014 -,079 ,341 ,559 -,029 ,046 ,830 -,047 ,120 ,729 -,032 ,054 ,816 -,037 ,075 ,784 2015 -,034 ,063 ,803 -,003 ,001 ,981 -,058 ,175 ,676 -,011 ,007 ,934 -,028 ,040 ,842 Brands ,089 72,324 ,000 ,081 56,018 ,000 ,079 50,594 ,000 ,083 47,761 ,000 Int_1 -,283 7,198 ,007 Int_2 -,306 13,672 ,000 Int_3 ,071 ,874 ,350 Model AIC / BIC 692,739 / 731,054 626,521 / 668,030 621,349 / 666,050 614,197 / 658,898 627,657 / 672,358 Likelihood (Sig.) 96,552 (,000) 164,769 (,000) 171,942 (,000) 179,094 (,000) 165,634 (,000) N 180 180 180 180 180
Dependent variable: Countries. Wald = Wald Chi-Square.
5.3 Regression analysis
Negative binomial regression is used to test the hypotheses, as the dependent variable Countries is a count variable. A count variable as dependent variable usually requires a Poisson regression model, but the assumption that the mean equals the variance is not met (µ = 2,4667 /σ2 = 5,613).
This explains the choice for a negative binomial regression model. Five models are created to test the hypotheses. Model 1 only includes the control variables GDP per capita of the home country (GDP Home), market size of the home and host countries (MS Home, MS Host), firm size (Firm Size), firm age (Firm Age), economic openness (Openness), infrastructure quality (Infrastructure) and economic instability (Instability). For model 2, the independent variable, domestic brand portfolio breadth (Brands), is added to test hypothesis 1. Models 3, 4 and 5 subsequently include the interaction terms Int_1 (Brands x Openness), Int_2 (Brands x Infrastructure) and Int_3 (Brands x Instability) to test hypotheses 2, 3 and 4. The results are presented in Table 4.
To compare the different models and to test whether the used models are a good fit to the data, I used Akaike’s information criterion (AIC) and the Bayesian information criterion (BIC). Lower values of AIC and BIC indicate an improvement and a better fit with the data (Kuha, 2004). In addition, I used the Omnibus test to check the Likelihood ratio and the significance of the fitted models to be able to predict meaningful results. The AIC and BIC scores are respectively 692,739 and 731,054. The Omnibus test for Model 1 shows that the model is significant and has a likelihood ratio of 96,552 (,000), so it can be used for further analysis. GDP Home has a positive but insignificant Beta in Model 1. Both MS Home and MS Host have a negative Beta, but only MS Home is significant (B = -,107; Sig. = ,015), which means that if the market size of the home country increases, the company is active in less countries outside its home market. Firm Size has a positive Beta and is statistically significant (B = ,134; Sig. = ,000). This indicates that bigger companies operate in more countries outside
their home market. The variables Openness (B = ,189; Sig. = ,011) and Infrastructure (B = -,297; Sig. = ,000) have a negative and statistically significant effect on the dependent variable Countries. This indicates that if the overall economic openness and the overall infrastructure quality in the host countries increases, the company is active in less countries. The opposite goes for Instability (B = ,207; Sig. = ,000). The year dummies have no significant Beta, which indicates that the influence of yearly differences is negligible.
To test Hypothesis 1, Brands is added to the model in Model 2. The AIC of 626,521 and the BIC of 668,030 show an improvement of the model in comparison to Model 1. The Omnibus test gives an indication of the significance of the model. The Likelihood ratio increases to 164,769 and is still significant (,000). This indicates that the model better explains the variance in the dependent variable Countries. Interesting to see is that the GDP Home now is significant (B = -,146; Sig. = ,008), which indicates that a higher GDP per capita in the home country, the number of countries served by a country decreases. To test whether the number of brands in the domestic brand portfolio (Brands) has a positive effect on the degree of regionalization (Countries), the variable Brands is added to the model. Model 2 shows a positive and statistically significant relationship between the independent variable Brands and the dependent variable Countries (B = ,089; Sig. = ,000). This indicates that if the number of brands in the domestic brand portfolio increases, a company serves more countries outside its home country. This provides support for Hypothesis 1.
In Model 3 the interaction term of Openness x Brands (Int_1) is added to the model to test Hypothesis 2. Both AIC (621,349) and BIC (666,050) improve in comparison to the previous model, which indicated a better fit. The Likelihood ratio of the Omnibus test is 171,942 (,000) , higher than the previous model and still significant at the 0,01 level. The model is used to test whether Openness negatively moderates the relationship between Brands and Countries.
Model 3 shows that Openness is indeed negatively moderating the relationship between Brands and Countries (B = -,283; Sig. = ,007). This means that Hypothesis 2 is supported.
In Model 4 the interaction term of Infrastructure x Brands (Int_2) is added to the model to test Hypothesis 3. AIC and BIC indicate that the model is an improvement in comparison to previous models (614,197 / 658,898). The model is still statistically significant, as the Omnibus test shows a Likelihood ratio of 179,094 that is significant at the 0,01 level (,000). Model 4 is used to test Hypothesis 3, whether Infrastructure negatively moderates the relationship between Brands and Countries. The results show that the Beta is negative and statistically significant at the 0,01 level (B = -,306; Sig. = ,000) and thus support Hypothesis 3.
In Model 5 the interaction term of Instability x Brands (Int_2) is added to the model to test Hypothesis 4. In comparison to the previous models with the other interaction terms, this model has the highest AIC (627,657) and BIC (672,358) values, which indicates that the fit with the data is not as good. The AIC and BIC values are higher than Model 1, but lower than Model 2 with only the independent variable Brands. The Omnibus test shows the model has a Likelihood ratio of 165,634; but is still significant (,000). Model 5 is used to test whether Instability positively moderates the relationship between Brands and Countries. The Beta is positive but not statistically significant (B = ,071; Sig. = ,350), so Hypothesis 4 is not supported.
6. DISCUSSION
6.1 Key findings and contributions
This thesis examined the effect of the domestic brand portfolio breadth and the degree of regionalization. Cantwell, Gambardella and Granstrand (2004) suggested there could be a direct relationship between the two variables. Prior research mainly focused on the relationship between several diversification strategies and firm performance (Chang & Wang, 2007; Hitt, Hoskisson & Ireland, 1994; Hitt, Hoskisson & Ireland, 1994; Hitt, Hoskisson & Kim, 1997;
Morgan & Rego, 2009; Tallman & Li, 1996; Zhao & Luo, 2002), but were unable to find conclusive results. Brand portfolio breadth is measured by the number of brands in the domestic brand portfolio and the degree of regionalization is measured by the number of host countries a company is active in. Besides the direct relationship, I examined the moderating effect of the host country determinants economic openness, infrastructure quality and economic instability to get a better understanding of the underlying effects, as these determinants are considered most important if it concerns country contexts (Erdal & Tatoglu, 2002).
I included several control variables to check their effect on the dependent variable. The market size of the home country and the size of the firm both had a significant effect on the dependent variable degree of regionalization. If the market size of the home country of a company was bigger, this lead to a smaller degree of regionalization. This could mean that the need to expand abroad is smaller, as companies can make enough profit in their own country. A positive effect of firm size on the degree of regionalization was expected, as companies have more resources to expand abroad. Surprisingly, firm age was not significant in any of the models, which indicates that the age of a company does not affect the degree of regionalization of that company.
The results show that the relationship between domestic brand portfolio breadth and the degree of regionalization is positive and significant and thus Hypothesis 1 is supported. If a company holds a higher number of brands in its domestic brand portfolio, it also has a higher degree of regionalization. This supports the suspicion of Cantwell, Gambardella and Granstrand (2004) that there is also a direct link between these constructs. This was also expected, because firms with a broader brand portfolio are expected to be better able to meet the demands of heterogeneous groups of customers (Bordley, 2003).
The moderator variables economic openness, infrastructure quality and economic instability are tested in models 3, 4 and 5. Hypotheses 2 and 3 were supported, as they showed
a negative and statistically significant Beta. This means that economic openness and infrastructure quality indeed negatively moderate the relationship between domestic brand portfolio breadth and the degree of regionalization. If the economic openness of the combined host countries is higher, the relationship between becomes weaker. A possible explanation could be that companies first expand to companies that are relatively easy accessible and thus the openness decreases if a company is active in more countries (Johanson & Vahlne, 1977; Vernon, 1966). The same effect was found for infrastructure quality, which indicates that companies first expand to countries with a higher infrastructure quality, learn from this and then expand to countries with a lower infrastructure quality. This also indicated that infrastructure quality is an important factor of trade success (Nordås & Piermartini, 2004). Hypothesis 4 was not supported, as no significant effect of economic instability was found.
This thesis contributes to International Business literature to extend the literature on brand portfolio strategy within the Asian market (Sim & Pandian, 2003). It tries to explain the relationship between brand portfolio breadth and the degree of regionalization and to see which strategic choices positively or negatively influence a company’s regionalization process. There still is little research on the direct relationship between brand portfolio breadth and the degree of regionalization.
6.2 Managerial recommendations
This thesis might be useful for managers that work for a company specialized in consumer goods and provide them with insights of internationalization strategy and decisions concerning the breadth of a company’s brand portfolio. The results of this thesis indicate that managers should consider having a broad domestic brand portfolio before expanding abroad. In this way they are better able to serve different customers in other countries.
If managers experience having a broad domestic brand portfolio and thus a more complex home market, they become better prepared to expand abroad and capitalize on the success they already have in their domestic market.
6.3 Limitations and future research
A limitation of this thesis is the sole focus on the Asian beer, spirits and wine market. Different outcomes may hold in other countries and other product markets. Because data was not available of the domestic brand portfolios of the countries that were in the dataset, I had to exclude these countries. It might be interesting to see if the findings are similar in Europe and regions around the world.
Another limitation of this thesis is that the sample only consists of 45 companies and is measured over the years 2012 to 2015. To be able to generalize the results it is recommended to increase the sample which might lead to results that are more precise to interpret and provide information about what is a desirable domestic brand portfolio breadth. Other determinants of both home and host country context might be interesting to include in future research.
7. CONCLUSION
This thesis aims to provide a better understanding of brand portfolio strategy and how it affects the regionalization strategy. Scholars appeared unable to reach consensus on the relationship between several diversification and brand portfolio strategies and firm performance. The results show that there is a positive relationship between domestic brand portfolio breadth and the degree of regionalization. Furthermore, it supports the hypotheses that the host country context determinants economic openness and infrastructure quality negatively moderate the before mentioned relationship. Economic instability appears to have no significant effect. Brand portfolio strategy and regionalization remain an important area for future research, as brand
become increasingly important to distinguish the company in a integrated global market with tough competition.
ACKNOWLEDGEMENTS
I would like to thank my supervisor, Dr. Vittoria Scalera, for her contributions to this thesis. Her valuable comments and suggestions during the meetings were greatly appreciated and enabled me to complete this thesis. In addition, I would like to thank my family and friends for their support and encouragement throughout my studies.
REFERENCES
Aaker, D. A. (1996). Measuring brand equity across products and markets. California
management review, 38(3), 102.
Bordley, R. (2003). Determining the appropriate depth and breadth of a firm’s product portfolio. Journal of Marketing Research, 40(1), 39-53.
Brouthers, K. D., & Hennart, J. (2007). Boundaries of the firm: Insights from international entry mode research. Journal of Management, 33(3), 395-425.
Calof, J. L. (1993). The impact of size on internationalization. Journal of small business
management, 31(4), 60.
Cantwell, J., Gambardella, A., & Granstrand, O. (Eds.). (2004). The economics and
management of technological diversification (Vol. 34). Routledge.
Cuervo-Cazurra, A. (2011). Selecting the country in which to start internationalization: The non-sequential internationalization model. Journal of World Business, 46(4), 426-437. Chang, S. C., & Wang, C. F. (2007). The effect of product diversification strategies on the
relationship between international diversification and firm performance. Journal of
world business, 42(1), 61-79.
Dacin, P. A., & Smith, D. C. (1994). The effect of brand portfolio characteristics on consumer evaluations of brand extensions. Journal of Marketing research, 229-242.
Dicken, P. (1992). Global shift: The internationalization of economic activity. SAGE Publications Inc.
Dielman, T.E. (1991). Applied regression analysis for business and economics. Boston: PWS-Kent Publishing Company.
