Regional orientation of European commercial service banks and the importance of cities as a location choice

(1)

1

Regional orientation of European commercial service banks and

the importance of cities as a location choice

Master Thesis

MSc. Business Studies – International Management

Supervisor: Dr. J.P. (Johan) Lindeque

Second reader: Mr. D.J.H.M. (Daniel) van den Buuse

Student: Patrick van Eijck

Student ID: 10610146

(2)

2

Abstract

This research positions itself within current regionalization literature as a gap is found in Rugman & Verbeke (2004). Most Multinational Enterprises (MNE) are found not to be truly global but are more regionally oriented. In light of this literature, an extra dimension is given on the influence of the global city, on a sub national level (Sassen, 2001; Scott, 2001). This is important as the global city is an important factor in the current global economy (Sassen; Scott, 2001). The banking sector is embedded in financial hubs which seem to have a predetermined relation to the global city (Sassen, 1999). A qualitative multiple case study was done with a sample of six firms. So, each case focuses on a different internalization strategy: bi-regional, home region and host region oriented firms (Rugman & Verbeke, 2004). The findings of this research indicate the value of the global city for higher value creating activity and their possession of relevant market assets. The relevance of the home region in which banks operate is a critical factor in their decision making process and the large amount of acquisitions of midsize banks indicate the need for banks to gain access to institutional equity in lesser developed markets. Therefore, the characteristics of the lower level category of cities are of minor importance as the acquirer is more focused on country level and firm specific level characteristics.

(3)

3

Index of tables

Table 1 Categorization of cities ... 16 Table 2 Sample and regional orientation ... 23 Table 3 Overview of selected + different sources used

for the collection of data (2000-2014) ... 24 Table 4 List of cities entered between 2000 and 2013 ... 27 Table 5 Coding scheme ... 29 Table 6/7 Banco Santander ... 32/33 Table 8/9 Barclays ... 36/37 Table 10/11 Crédit Agricole ... 40/41 Table 12/13 Société Générale ... 44/45 Table 14/15 HSBC holdings ... 48/49 Table 16/17 Standard Chartered ... 52/53 Table 18 Cross-case comparison ... 57 Table 19 Working propositions results ... 63

(5)

5

1. Introduction

The European debate about the future of the Eurozone is a hot topic at the moment. The discussion is of whether the Eurozone should integrate further or whether member countries should retain sovereignty in their decision making. The outcome of this discussion has important implications for Multinational Enterprises (MNE) and their future strategy. This research does not focus on the political discussion of the regions nor their integration, but rather the relevance of the region and its influence on the decision making process of its location choice for an MNE.

This research positions itself within current regionalization literature, as due to the growing relevance of regionalization in current international business, more research is required. The critical position of the MNE in the global network is the result of rapid technological innovation (Brune & Garett, 2005). As a consequence of this rapidly changing world, the question remains whether MNEs are truly global in their activity. For instance, the existence of Regional Trade Agreements (RTA) suggests that the world is not globally connected (Banalieva et al., 2010). However, within this region (European Union) most goods, services and people move freely which provides the room for MNEs to engage in many countries. Hence, the world is not fully integrated nor fully dispersed (Ghemawat, 2003; 2005). In line with this thinking, current regionalization literature builds on the importance of the region in which a firm operates. Most importantly is the finding of Rugman & Verbeke (2004) which presents empirical evidence of the fact that a significant amount of Fortune’s top 500 MNEs have a minimum of 50% sales within their home region. These companies have a strategic orientation towards their home region, and only a few companies are found to be truly global.

Currently, regionalization literature is focused on the regional and country level effects on location choices by Multinational Enterprises (MNEs). Therefore, this research identifies a gap in the literature by evolving from national level of analysis to a sub national level of analysis. A report by McKinsey & Company (2013) points out that city leaders need to identify key factors of importance in developing their competitive clusters. Therefore, selecting the right sectors which are compatible with the competitive advantage of the city. Principally, the report indicates the importance of the development of the city region to establish sustained growth. Moreover, according to the growing urbanization rates (U.N.

(6)

6

Habitat, 2011); there is not only an opportunity but also a need to develop the literature on the influence of the city.

This research aims to seek signs of whether the city as a location choice is important in MNE decision making. This is done by relating foreign direct investment (FDI) motives, which are developed by Dunning (1998), with different categories of cities (Sassen, 1999). The commercial banking and savings sector is investigated as it is an interesting sector which is highly dependent on the regional institutions and financial centers. Since the financial crisis a new set of multinational banks have emerged as they are ‘disaggregating’ the value-creating activities of the financial services value proposition to allocate them within the network of globally dispersed operating units’ (Mudambi & Venzin 2010, p. 1521), this has led to an increase of their global presence. The banking sector is heavily technology dependent in their day-to-day business activity. Therefore, it is relevant to develop the regionalization literature of Rugman & Verbeke (2004) in relation to the influence of global financial centers (Sassen, 1999).

The sample draws commercial banks from the MNE Fortune 500 list. These MNEs are Banco Santander, Barclays, Société Générale, Crédit Agricole, HSBC Holdings and Standard Chartered. As banks are heavily dependent on their institutional environment, the environmental rules and cultural aspect of the services they deliver, this research adapts an institutional aspect on the internationalization of the banks in the commercial and savings sector. The institutional perspective that addresses the choices made by the banks must be related to the global city as a location choice. As indicated before by Dunning (1998), FDI motives are used to analyze the decision making process by the MNEs. Consequently, the following research question is developed:

‘To what degree are global cities important as a location for MNEs regionalization in the commercial banking sector?’

In terms of further aiming to answer this question, the research starts off with providing an overview of different literature on regionalization and location choice. As a consequence of this discussion, a conceptual framework is built on which the theoretical base derives working propositions. These provide the basis of the motives (Dunning, 1998) of the city as a location

(7)

7

choice that are specific to the commercial banking sector. Correspondingly, a categorization of the global city is made in which the different types of city have certain institutional aspects that relate to the different motivations of the city as a location choice. As a consequence of the theoretical framework, a methodology chapter follows on the philosophical base, method, sampling, data collection and analysis.

The following chapter continues with the analysis, as this research takes on a multiple case study approach. Therefore, three different cases with six samples are investigated from the same home region but with different regional strategies. This is done by analyzing the entry decisions within and across the different cases. First, the regional strategies of the different banks are set by using longitudinal data of their sales, employees and assets across multiple regions. Secondly, by using qualitative data the different choice of locations and reasoning behind these locations can be identified. The final part discusses the location choices by these MNEs and characterizes it by using profound variables to give a better perspective of their attractiveness as location choice.

2. Literature review

The first part of this literature review discusses the regionalization vs. globalization debate. The second part looks further into the current status of regionalization literature. Current literature falls short in describing the relevance of the city. And so, the third part goes further into the location choice in relation to the global city. Hence, the location choice literature starts off on national levels and country level, then goes to the host location level and to the sub-national level, focusing on the city and its region.

2.1 General overview of regionalization

Levitt’s (1983) work on the globalization of markets was an inspiration for further assessment of this subject. As he argued, the world is increasingly moving towards standardized products and a homogenization of demand due to the new facts of technology in communication, transport and travel. This leads to a critical stage in the rising importance of the multinational enterprise to coordinate worldwide activity in marketing, production, distribution and management by using economies of scale and scope. Furthermore, he argued that there is no exemption to the rule in terms of sectors, given the strength of globalization (Tedlow & Abdelal, 2003).

(8)

8

Garrett (2000, p. 941) defines globalization as “the international integration of markets in goods, services, and capital”, and while he recognizes that “other facets of the phenomenon (such as increased labor mobility and cultural homogenization) are surely essential, but [he leaves] their analysis to others.” However, as argued by Tedlow & Abdelal (2003), this work on globalization is excessive in its argumentation because globalization is not an unstoppable process. Undoubtedly, globalization and technological innovation has led to changes, but not as much as suggested above (Garrett, 2000).

Recently, the focus of literature has moved away from globalization and switched towards regionalization, where regionalization is a form of semi-globalization (Ghemawat, 2003, 2005), which implies that global markets are not fully integrated, but fragmented. As stated by Hymer (1972), the MNE is so influential in its worldwide coordination possibilities that it stresses social and political ties. The MNE cannot be stopped by the national borders and undergoes mainly a process of centralization, therefore pushing further inequality amongst the different countries. In this context, Garrett (2000) argues that one of the perspectives of globalization is technological determinism, which identifies itself within the banking sector. The perspective states that: “The shrinkage of time and space has been so persuasive that there is essentially nothing that can be done to stop it” (Garrett 2000, p. 942). In relation to the banking sector, FDI and international finance are influential, because both are applicable to the banking sector. Technological innovation has a contradictory influence on both. For FDI, only government policy can influence the MNEs decision making process. As for international finance, government policy is of minor relevance due to rapid technological innovation, rules can easily be bypassed.

The literature on semi-globalization provides the basis for the first empirical evidence of the existence of triad regions (Rugman and Verbeke, 2004). Furthermore, the distinction is made between home region oriented, host region oriented, bi-regional and truly global firms. The different types of regional strategies lay the foundation for analyzing the location choice by MNEs. The significance of the home region is indicated by the low intra-regional sales. As a result, three triad regions can be distinguished. These triad regions have distinctive characteristics that can influence the location choice decision of firms (Rugman & Verbeke, 2004).

(9)

9

One crucial finding is that MNEs have certain firm specific advantages (FSA) that are region bound because they are not easy exploitable across the borders of the home region. Therefore, it presumes that MNEs have location bound FSA. Consequently, most firms internationalize within their home triad region and only a few are found to be truly global. This results in four distinct types of firms, first, home oriented firms with a minimum of fifty percent of sales in their home region. Secondly, the bi-regional firms, with a minimum of twenty percent of sales in two regions but no more than 50% in these regions. Thirdly, the host region oriented firms that have fifty percent of sales in a region that is not their home region. Finally, the global firms with twenty percent sales in all three regions but no more than fifty percent in one region (Rugman & Verbeke, 2004).

Criticism on Rugman and Verbeke (2004) comes from the lack of use of macro-data in their investigation (Dunning et al., 2006). Examples are the FDI, GDP and trade between countries as some indicated measures of globalization indicated by Garrett (2000). Dunning et al. (2006) precedes using macro-data to extend and confirm the model by Rugman and Verbeke (2004). Yet, the expansion shows that outward FDI is more widely spread in geographical reach. A possible explanation is that the data excludes the home countries of the selected MNEs. In addition, exports from the home region are included as well as the sales to foreign affiliates which, of course, increase the intra-regional share of total sales.

In response, Rugman and Verbeke (2007) show a solution to the country and firm level conceptual problems by using macro-economic data in line with Dunning et al. (2006). The large preference for home region orientation is explained by Rugman and Verbeke’s (2007) work drawing on Zaheer’s (1995) liability of foreignness (LoF) in their research. The liability of foreignness is the cost of doing business outside the home country that is in addition to those liabilities experienced by host country firms; usually these liabilities are associated with unfamiliarity and discriminatory costs. Correspondingly, Rugman and Verbeke (2007) discovered that “the liability of intra-regional expansion appears to be much lower than the liability of inter-regional expansion” (Rugman & Verbeke 2007, p. 201). It implies that the LoF is lower within the home region (intra-regional) than outside and between regions (inter-region). This view is supported by Kolk et al. (2013) study with the focus on the electrical utility sector. Correspondingly, the influence of the home country region is derived from a lower LoF. As a result, the MNE is restrained to stay within their home region. One possible

(10)

10

reason is that government policy has an impact on the internationalization process of electric utility companies. Furthermore, Rugman and Oh (2012), argue the occurrence of a regional path dependency in the life cycle of a firm. This path dependency is the outcome of human resources, infrastructure and most importantly institutions. Therefore, the next part explores the institutional perspective in relation to regionalization literature.

2.2 Drivers of regionalization

Both the resource based view (Peng, 2001) and the institution based view (North, 1991; Scott, 2001) have been used to explain the degree of intra- and inter-regional LoF. However, for this research the institutional based view is adopted. The importance of the institutional environment is supported by research from Dikova et al. (2009); it indicates that formal and informal institutional differences are of substantial influence on purchasing or abandoning, at the completion stage, an acquisition. The likelihood of the successful completion of an acquisition depends on the past experience in acquisitions of an MNE within this institutional environment. Furthermore, past experience cuts back the duration of acquisitions in a more distant institutional environment.

The difference between formal and informal institutions is developed by North (1991) and is viewed from an economic perspective. This study does not adopt the sociological perspective of institutions as developed by Scott (2001), rather adopting North’s (1991, p.97) definition of institutions as “humanly devised constraints that structure political, economic and social interaction.” The informal institutions are the sanctions, taboos, customs, traditions and codes of conduct within an institutional environment. Contrarily, the formal institutions consist of constitutions, laws and property rights. The institutional environment is formed by an incremental process to reduce uncertainty within a historical context. MNEs adapt their behavior towards the institutional environment in which they operate, as it determines costs and (long term) profit. From an economic perspective this can be viewed from the game theory where each individual, in this case MNEs, maximizes its profit.

Consequently, in relation to the institutional perspective, this part elaborates on the main factors of influence on internationalization by MNEs within their home region. As argued by Driffield et al. (2013), the regional dimension has some valuable contextual factors that can be of influence on the decision making process by MNEs. First, the technological development within the region, as argued, is of great importance because most of the core

(11)

11

competences of firms stay within their home region. Therefore, firms gain knowledge within their home region through reverse technological spillovers. Secondly is the development of the institutional environment. The third factor of influence is the development of a system that preserves MNEs patents. The fourth factor of substance is the quality of the financial system, as it corrects the firms embedded in the market. If this is of high quality, firms are more efficient or if not the financial system corrects it by lowering investments.

The fifth and last factor is the strength of the regional trading block. In this regard, Enright et al. 2005 argue that MNEs internationalization within the home region is mainly the outcome of regional trading blocs and innovations in information systems. As a consequence, the RTA within the highly integrated European Union is crucial (Banalieva et al., 2010). Therefore, MNEs that are home oriented are perceived as being more efficient. Hence, MNEs relocate value chain activity to more efficient locations because the RTA results in lower trade barriers between countries within the region. This in turn provides a basis for the importance of the institutional environment of the European Union, as it stimulates MNEs to expand their business within this region. As supported by Driffield et al. (2013), the aforementioned factors are well developed within the European Union. Therefore, the factors of influence intensify the effects on internationalization within the home region.

2.3 Motives of regionalization

As argued by Khan (2010), the motives of an MNE for pursuing a global strategy can be either market or production oriented. An MNE needs to be responsive towards their customers and also integrate its value chain to be more efficient towards the competition. This means that if you look from a market perspective, regionalization of the strategy means being responsive towards customers in those regions. The underlying phenomenon is that homogenization of preferences is restrained by local taste which is embedded into culture. Thus, contributing towards the informal institutions in which the MNE operates. Moreover, next to culture, the public trust of a bank within a region is also crucial for a commercial bank (Deephouse, 1996). On the other hand, the production perspective takes us back to the regional block argument as mentioned before. This limits the integration possibility for an MNE which makes it more attractive to continue producing within a region because the firm can better organize the environmental uncertainty within its home region. However, also cost and possible agglomeration advantages are of significance (Arregle et al., 2009). In particular, for commercial banks the production perspective takes us back to the institutional

(12)

12

environment in which it operates. Hence, the regulatory rules it has to follow and the periodic checks by capital market regulators (Deephouse, 1996).

2.4 Location choice

The next part discusses, in short, the country level influence on an MNE, then continues with an elaboration of host country literature from an institutional perspective and resource endowments of host locations. The second part moves towards a lower level of analysis in terms of the region and in particular the city in which an MNE can locate.

The differences between countries demonstrate the variation in performance of foreign subsidiaries of MNEs. Porter (1990) argues the importance of a nation’s competitive advantage due to localized processes. Nations have influence which makes them more successful in particular industries due to unique aspects of the national values, culture, economic structures, institutions, and history. However, this model fails to incorporate the international aspect of a nation’s competitive advantage. Therefore, Moon et al. (1998) show the relevance of multinational activity (FDI) for a nation’s competitiveness. Factors of importance here on the competitiveness of a country are factor conditions and demand conditions, are equally important, firm strategy, structure and rivalry. Lastly, the presence of related and supporting industries is essential.

As a result of these differences among countries, and due to semi-globalization as discussed before, there is a need for the MNE to engage in multiple locations immediately. As first stated by Barlett & Ghoshal (1988), a tradeoff exists between local responsiveness and global integration of an MNE. Rugman et al. (2011) refined this framework by introducing the complex situations which MNEs find themselves and the different local contexts in which they operate. As a consequence, Meyer et al. (2011) argues that there is a need for a tradeoff between external and internal embeddedness. An MNE needs to capture local opportunities and at the same time adequately embrace the internal network of the MNE. As a result, the external benefits can be moved internally to other parts of the MNE. The next part moves towards the relevant aspects of host locations that attract MNEs. This is done in respect of Meyer et al. (2011), which describes the local context by the differences between institutional frameworks and available resources. The local context as described below pays attention to the CAGE distance as developed by Ghemawat (2001), where firms have lower geographic distance between countries within the home region of Europe. However, some local

(13)

13

differences still and are of relevance between certain countries within particular regions (Meyer et al., 2011).

First of all, Meyer et al. (2011), argues that MNEs need to adapt to the institutions embedded in a location because they are unlikely to change anything institutionally. More relevant is the stability of the host country because it determines the effects on performance due to the cost of monitoring and enforcement of contracts. On the other hand, the formal institutional environment can provide incentives to attract MNE by reducing taxes or providing investment opportunities (Chan et al., 2008). This is supported by Laamanen et al. (2011) who addresses the importance of the tax level on HQ-relocations, especially within the European Union where there is not an equal taxation level throughout the countries of the European Union. Therefore relocation of HQ within these borders is quite common. In addition, a host country can provide a strong juridical system that stimulates a clean business environment. On the other hand, the host country can favor home country companies and therefore can have stricter legislation for the foreign affiliates of MNE (Chan et al. 2008). In addition, Claessens & van Horen (2008) argue the possibility for banks to create an institutional competitive advantage. In this way, banks can locate more easily in host countries that have the same level of institutional development. As a result, more institutionally experienced banks have an advantage over their competitors in well-developed or underdeveloped institutional environments. Looking at the informal institution, this can be attributed towards the attractiveness of the host locations with regard to its social norms and behavior. Therefore it is linked to differences in national cultures as argued by Hofstede (1994). As a consequence, MNEs are more likely to choose host locations which have more overlapping layers with their culture as it makes it easier for MNEs to adapt within this local context. This results in less complexity within the organization as it adapts better to the external and internal embeddedness of the MNE, as discussed before (Meyer et al. 2011).

More critical in choosing a host location is the availability of resources. Consequently, with the use of Dunning’s theory (1998) more insight is given into the decision making process of a host location by MNEs. There are four motives that explain location choice by MNEs. First, resource seeking FDI, are the traditional motives as are natural resources, infrastructure, certain government restrictions on FDI/capital and dividend remissions and tax incentives. Furthermore, provision of local opportunities that are available for the quality and processing

(14)

14

of resources and the presence of local partners. The second motive is the market seeking FDI, the demand of the large markets or regions such as the European Union. Also important are, the presence of skilled labor, competitiveness, quality of infrastructure, less market failure, agglomeration advantages, local service support facilities, nearby users in knowledge intensive sector, regional and local development agencies. As important, is the efficiency seeking motive, mainly cost efficiency ideas, as well as freedom in trading in order to internationalize efficiently. Other possibilities are the opportunities of agglomeration advantages and investment incentives. More relevant is the role of government in reducing market failure and the opportunities given by specialized clusters. Fourthly, strategic asset seeking FDI is a strategic decision to invest in certain types of assets. Examples are technical knowledge, learning experience, management expertise and organizational competence. These are typically found in developed countries, which therefore provide MNEs with the best economic and institutional attributes. In addition, there are also agglomeration advantages, because of cluster possibilities and the presence of specialized services (Dunning, 1998).

The innovation possibilities result in easier coordination possibilities across borders. Consequently, the increase in preference of MNE to co-locate their activity with other firms to form alliances has to lead to the concentration of activity in some locations (Dunning, 1998). As a result, not only location choice can be viewed from a regional level in terms of a country versus international but also on a sub national level. Hence, literature is falling short in analyzing further regionalization and location choice on a sub national level.

The concentration of activity implies the growing value of the city as a level of analysis. As supported by Sassen (2001), there must be an increasing importance of global cities because of the growth in speed of information and capital flows. Global cities and their regions, as argued by Scott (2001), are of great importance in this (semi) globalized world. However, first the global city must be defined, before progressing to the different institutional aspects of the city. The global city is defined as an important part of the strategic function in cross border networks of MNE as it is the outcome of localized processes which have a political, economic and cultural base (Sassen, 2004). Consequently, categorizing global cities is necessary to establish different types of global city as there is a broad range of differences among cities (Goerzen et al., 2013). Hence, bona fide global cities are divided into three groups based on ‘global capacity’ developed by Sassen (2001). This global capacity refers to the capacity of

(15)

15

global cities to exert global control (Robinson, 2009). The first category refers to the cities which possess the most global control; therefore they are more likely to retain strategic assets. On the other hand first and second tier global cities tend to retain the largest markets; therefore being very attractive for MNEs.

The growth of these city regions in both developed and developing countries are the outcome of growing urbanization rates and global economic integration (Scott, 2001). These regions are dealing with a growing number of problems and have to respond to this with economic and political measures. However, city regions provide certain features that make them unique because they embed certain local firms. This creates an agglomeration of creativity, learning and innovation which is supported by the cities’ for example distinct, formal and informal institutional features (O’Neill, 2003). Therefore, as discussed before, the agglomeration advantages offered by the (city) region are meaningful in the decision making process of MNEs (Arregle et al. 2009). Furthermore, “As national states erode, so they say, global city regions forge cross-national networks feeding from the opportunities created by new multinational geopolitical spaces” (O’Neill 2003, p. 326).

Furthermore as argued by Hymer (1972), business is the heart of the city in relation to its geographical specialization. Sassen (1999) further explains this in terms of presence of resources and advanced technology. Highly talented accountants, lawyers and data analysts are needed for the complexity of the data that banks have to process. However, nowadays more than half of the business is done by electronic information systems. This can be attributed to the formal institutional facilities that are present within the city. For example, the presence of prestigious universities where businesses can attract talented employees. Moreover, the occurrence of supporting mechanisms of local government to attract certain types of businesses by adapting legalisation or providing subsidies is extremely important.

Alongside to the formal institutional aspect within the city region, the informal aspects are important, For instance, the substantial value of a city’s (cultural) history must not be neglected as this lays the foundation for cultural connection among countries and cities. Good examples of this are the ties linked to the colonial backgrounds between countries. The city’s history and colonial ties are arguably the most profound reason of the substantial urbanization growth rates in different cities (McCann & Acs, 2011).

(16)

16

2.5 Conceptual Framework

Consequently, the conceptual framework to develop the working propositions is developed. First, the location choice and entry mode decision to analyze the internationalization process of an MNE within the commercial banking industry is discussed. The second part develops the working proposition on which the analysis is based.

This research is based on the fact that there is no clear reasoning on how MNEs choose a location and the influence of the city. By using Dunning’s location motives, the decision making process by MNEs is analyzed further. From an institutional perspective a clear overview is provided of the location choice by MNEs on a sub national level. In relation to the sub national level of analysis, global cities seem to be a critical aspect of the commercial banking sector. In relation to the previously named definition of the global city, Sassen (1999) argues the existence of global financial centers. However, companies cluster into these global financial city centers, but seem to be not truly global in their focus (Grosse, 2005).

Global Capacity National consolidation Major institutional equity holding Liberal market

Category 1 Very High x x x

Category 2 High - - -

Category 3 Normal

Table 1 Categorization of cities

As indicated previously, there are different categories of global capacity to define a global city, to refine this for the commercial bank sector; a distinction is made between three categories: leading global financial centers, secondary smaller economic capitals and cities that have no direct financial-economic influence and global capacity (table 1). First, national consolidation is of importance, for example the growing concentration of investment banks in New York, which diminished the presence of investment banks in other cities within the U.S (Sassen, 1999). Secondly, the presence of major institutional equity holding as for example the stock exchange market is relevant (Sassen, 1999). Also the presence of prestigious universities for a wide range of talented potential employees has to be taken into account. The third relevant aspect is more related to emerging markets as the deregulation supports growth of particular cities as they function as a bridge between the liberated home market and international business activity. What can be noted here is that not all three aspects mentioned before have to be met for a city to be defined as a global financial center (Sassen, 1999).

(17)

17

After deciding the location in which the multinational enterprises wishes to enter, the mode of entry must be decided by the MNE. In this regard, Pan & Tse (2000), argue the difference between equity and a non-equity mode of entry. An equity mode includes wholly owned subsidiaries (WOS) and equity joint ventures (EJV); these types exert the most control and ownership. As argued by Lehner (2009), both these entry modes are common within the banking sector. Furthermore, banks can decide to pursue cross border lending opportunities, however this seems less decisive in this research. The banking sector seems mostly dependent on two factors, the host country development and its market size. Regarding the size of the market, acquisition is often chosen in markets that are smaller and less developed. Consequently, both WOS and EJV are used to analyze the location choice decision.

However, because of the focus on a specific sector: commercial banking, sectorial asset specificity needs to be taken into account. Therefore, the commercial banks are conceptualized as intangible FSA driven financial MNEs. Banks within this commercial service sector are knowledge driven MNEs which possess certain type of intangible assets which they can use as leverage across or within borders.

To further understand the location choice by MNEs, and in particular by banks, this part develops three propositions on which the next part of this research can build. As argued by Focarelli & Pozzolo (2005), banks generally open branches and not subsidiaries in countries with financial capitals. This can mostly be attributed to the history, organizational culture and core business. Moreover, the past experience must not be neglected as this can cuts costs (Ball & Tschoegl, 1982). By further elaborating on the third aspect of a global financial centre a research which looked at German firms found a positive effect of European Union on the financial centres and its appeal to foreign banking (Buch, 2000). One factor of influence is the liberal regulatory regime, but also the low information asymmetries between countries which are relevant and must not to be neglected. These low information asymmetries are the result of similarities between informal institutions of the different countries.

Substantially, in the case of the E.U., formal institutions similarities are critical, as is the influence of the euro which terminated exchange volatility and promoted FDI in the non-banking sector (Buch, 2000).

(18)

18

More importantly, the commercial banking sector is mainly focused on the consumers/clients and being close to them. Because of their focus on traditional lending and savings activities and the deliverance of more non-traditional products like mutual funds sales, data processing services and mortgage servicing, there is a need for commercial banks to be near to the consumers (DeYoung & Roland, 2001). Moreover, there is a need for centralization of the decision making process due to the complexity of financial services. Goerzen et al. (2013) argue that the presence of firms in global cities is mostly driven by market seeking behavior by MNEs. The connectivity of the global city to the global network reduces the LoF for entering the host country by an MNE. Furthermore, asset and efficiency seeking behavior seems to focus on the metropolitan areas outside the global city. However, as indicated before, a difference is made between branches and subsidiaries of commercial banks. Opening a branch to deliver financial services in financial capital to the customer in that market, would indicate market seeking behavior by the MNE. More critical for a bank is the location convenience for its customers. However due to technological innovation, the possibilities of internet banking, telecommunication and ATM’s minimizes this preference (Zineldin, 1996). Hence, proposition one is developed below.

WP1: Intangible FSA driven financial MNEs open branches in the first two categories of cities because of market seeking motives.

The next proposition focuses on the strategic asset seeking motives of the banks. As stated before, opening a subsidiary or pursuing an acquisition by the bank will need to gain more than a market seeking motive. As Buch (2000) argues in his research about the German banking sector, German MNEs may acquire local banks because of the size of the market and to gain local market knowledge, thus, a strategic asset seeking motive. However, this is more on a country level and not on a sub national level of analysis. Hence, it is argued by Sassen (1999), that global cities retain the presence of resources and advanced technology. Thus, high value creating activities are performed in these global financial centers or otherwise called financial service hubs (Mudambi & Venzin, 2010). Furthermore, higher level institutional decision making is done in these global cities as they possess major institutional equity. For the banking sector this means the presence of a capital market within the city, as for example the securitization of loans can be performed here (Mudambi & Venzin, 2010). Therefore, proposition 2 is explained below.

(19)

19

WP2: Intangible FSA driven financial MNEs open subsidiaries in first category global cities because of strategic asset seeking motives.

On the other hand, the lower value-added and standardized processes of the bank are more likely to be moved to emerging countries as they possess lower costs in relation to skilled human resources or process capabilities (Mudambi & Venzin, 2010). Certain activities are off shored because they are the ‘backbone’; examples are call centers, administrative work or software development (Mudambi & Venzin, 2010). These activities can be related to an efficiency seeking motive as they are cost related and administratively efficient. Off shoring relates to avoiding extra layers within the organization and consequently bettering the service towards the customer (Mudambi & Venzin, 2010). In this case, the cities do not need to possess specific attributes of the global city and do not need an exceptional global capital. Therefore, category three cities are more related to an efficiency seeking motive of MNEs as is shown in working proposition three.

WP3: Intangible FSA driven financial MNEs open subsidiaries in the third category cities because of efficiency seeking motives.

Consequently, the next part will continue to develop a research design that is able to look for indications of these proportions in the form of a multiple case study design.

3. Methodology

This chapter elaborates on the methodology though the research question is answered. The first part describes the ontological and epistemological foundation on which this research design is based on. The second part focuses on the multiple case study design in light of the Yin (1981, 2009) and Eisenhardt (1989b) work that adopt an approach that Shepherd and Sutcliffe (2011, p.362) refer to as “combining deductive approaches with bottom-up theorizing”. Moreover, the quality criteria, case selection and sample criteria that are associated with the multiple case study design are discussed. After this the proceedings include the data collection methods which will address the quality criteria issues associated with multiple case studies. Thus, the data analytical approach and sources of the collected data are shown.

(20)

20

3.1 Ontological and Epistemological Foundations of Study

The ontological foundation for this research is objective. Ontology is the manner in which the world can be addressed. Objectivity attains to a theory-neutral language in which the social and natural science exist independently from human cognition (Brannick & Coghlan, 2007). In other words, this objective foundation views the world objectively without the interaction effect of humans taking an objective perspective in qualitative research is the philosophical assumption of Yin (2009).

The epistemology is concerned with the nature of knowledge, what constitutes valid knowledge, what can be known and who can know (Ryan 2006, p.15). The objectivist view on epistemology is that the world can be accessed externally (Brannick & Coghlan, 2007). Thus, this comes forwards in the epistemological approach of positivism. The post-positivist perspective (Ryan 2006, p. 18) takes on a learning role instead of testing. This type of research does not seek to achieve an overall truth but acknowledges the complexity of a phenomenon. This perspective is consistent with formal deductive procedures which provides in depth understanding for a theoretical explanation of a general phenomenon (Hyde 2000, p. 84).

3.2 Qualitative Multiple-Case Study Design

This research draws on the approach of Yin (2009) to develop a multiple case study which is explanatory in nature. Yin (1981, 2009) is at the core of this research design as it seeks to answer a ‘How’ / ‘Why’ research question (Yin, 2009). . The case study is different from other methods and can be defined as suitable when studying “(a) contemporary phenomenon in its real-life context, especially when (b) the boundaries between phenomenon and context are not clearly evident” (Ying 1981, p. 58). Moreover, a strong theoretical foundation is followed up with developing propositions important to allow generalization to theory (Yin, 2009). This research is theoretically driven while still being open to new insights that might allow theory extension. For the ‘deductive bottom-up theorizing’ approach (Shepherd & Sutcliffe, 2011) this is one of its main strengths, as it takes on a position of being open to new insights while having a predetermined set of constructs, as in the research question in combination with the working propositions, to guide through a large amount of data.

(21)

21

Furthermore, this research draws on techniques used in the inductive approach of Eisenhardt (1989b) and Eisenhardt & Graebner (2007), as they support the multiple case design and are equally applicable to the deductive approach (Yin, 2009; Hyde, 2000). However, this research does not seek to build theory, but is tightly scoped within its context, as theory driven research that extends existing theory (Eisenhardt & Graebner, 2007). Thus, this paper draws on both research design features that are applicable to both inductive and deductive case studies.

3.2.1 Quality criteria for the Multiple Case Study

The strength of a multiple case study approach is that certain decision making processes or events can be described and explained in detail. Furthermore, the use of multiple methods makes the arguments and findings more compelling (Yin, 2009). On the other hand, often the reproduction of particular evidence fails and has been criticized for lacking generalizability (Yin, 2009). This can lead to questioning of conclusions and rigor of the study (Yin, 2009). Furthermore, extensive use of empirical evidence in multiple case studies can lead to too much complexity, too narrow and idiosyncratic theory as they are unable to identify the most important relationship between observations (Eisenhardt, 1989b). Hence, the multiple case study approach needs to follow a certain set of criteria to enhance the credibility of the research. To increase the rigor of a multiple case study, four quality criteria must be considered; construct validity, internal validity, external validity, and reliability (Yin, 2009; Gibbert et al. 2008).

First of all, construct validity, is the way that the variables that are made operational match the concept of the research objective. As a result the concept that is studied precisely has conclusions that increase in credibility and generalizability (Yin, 2009). In this case, the IBV is use as the overreaching theoretical foundation between the regionalization and location literature. For this reason, the case study builds in a sufficient base for further analyses. In this multiple case study, two tactics to increase the construct validity are adapted to this research (Gibbert et al. 2008). First, the same chain of evidence needs to be repeated at the collection of data within and across cases. Second, multiple data collection methods are used as in line with the post positivist view.

(22)

22

Secondly, to increase internal validity, data has to be viewed from different angles to find similarities and differences between cases (Yin, 2009). This is done to support the possibility of a plausible causation between the variables and the results. Hence, measures like the one mentioned before need to be adapted. Furthermore, pattern matching should be used to find similarities within cases of MNE that possess the same regional strategies as shown in table 2 (Gibbert et al., 2008; Hyde, 2000). Consequently, argumentation of possible causal relationships of location choices can be supported.

Thirdly, the external validity is linked to the problem of how findings from multiple-case studies can be generalized; this is the biggest problem for a multiple case study approach (Yin, 2009). The reason for this is that the sample size is small, which limits the possibility of generalizability of a population in the most positive sense. Therefore, other measures must be taken to enhance this, such as cross case analysis being applied. As can be seen in table 2/3, there are three cases which each include theoretically sampled firms. Subsequently, a cross case analysis can be performed to enhance the external validity (Gibbert et al., 2008). As a consequence, both within and cross case analysis increase the analytical generalizability by replicating multiple cases (Yin, 2009).

Lastly, to increase reliability, or otherwise called creating a lack of random errors. This is done by increasing transparency and replication (Yin, 2009). As mentioned before, replication is done by following the research procedures. Furthermore, transparency is created by describing each case in great detail. With the use of the following strategy: ‘preferring internal and constructs validity over external.’ (Yin, 2009). As this research is part of a larger group project with each participant’s own sector, ultimately, this results in a greater external validity across the different sectors. It is critical that each one concentrates on the trustworthiness (construct validity) and avoids interpreting spurious relationships as genuine relationships (internal validity) (Yin, 2009). Moreover, a case study protocol must be followed to carefully document and clarify the taken steps. So this can be filled in to a database which enhances the replication possibilities (Gibbert et al. 2008; Hyde, 2000).

3.2.2 Case Selection and Sample

As seen in the research question, this research is sector specific. The conceptual framework shows that the banking: commercial service sector is related to global financial centers (Sassen, 1999). This provides an example that the city is already as a location choice for the

(23)

23

MNE active in the commercial banking sector. Thus, the choice of this sector contributes to the theoretical development (Eisenhardt & Graebner, 2007).

Table 2: Sample and regional orientation

The sample is theoretically drawn, as the choices fit with extending and excluding particular variables in the constructs that are investigated (Eisenhardt 1989b; Eisenhardt & Graebner, 2007). To increase the theoretical development, three cases studies are executed in which two units of analysis are embedded. Within a case, two commercial banks with the same regional strategy are chosen (Rugman & Verbeke, 2004), which is possible through the availability of secondary data for the year 2012. This knowledge in combination with the literature guides the attention and reduces the influence of the environment, which makes it more manageable for incremental theory building (Shepherd & Sutcliffe, 2011). Moreover, the categorization of the samples increases the replication possibility to assess the working proposition and stimulate external validity (Hyde, 2000; Yin, 2009). Therefore, a sample size of six would seem sufficient as the number of rival explanations seem of minor importance (Yin, 2009).

(24)

24

Case 1 Case 2 Case 3

Banco Santander Barclays Crédit Agricole Société Générale HSBC Holdings Standard Chartered

Home Country Spain UK France France UK UK

Home Region Europe Europe Europe Europe Europe Europe

Rank in Global Fortune 500 (2012, based on revenues)

44 121 58 67 53 448

Total Revenues (Million $)

117,408 68,95 105,156 98,464 110,141 24,488

Sector Banks: Commercial and Savings

Regional orientation Bi-Regional Bi-Regional Home Region Home Region Host Region Host Region

Triad Region Focus Europe Europe Europe Europe Asia Asia

Commercial services Payment methods, insurance Santander and global private banking.

Corporate banking, international banking and personal banking

French and international retail banking

French banks, international retail banking, financial services.

Commercial banking: Business banking from cash management to commercial mortgages, corporate banking, international services

Personal banking, Business banking, the private bank, Islamic banking and online banking.

Description Banco Santander started off in 1857 with a right to trade with the outside world from the port of Santander with LA. Most important pillars of BS are retail focus, capital discipline and financial strength, prudence in risk and efficiency. Moreover, a merger in 1999 between BS and the BCP resulted in the largest financial institution of Spain and LA.

Barclays has a

worldwide presence and possesses 300 years’ experience in banking. As it is present in over 50 countries, its core services are moving, lending, investing and protecting money.

Since 1894, Crédit Agricole is the top financial partner of the French economy as it connects local and regional banks. It is a customer-focused banking model as their retails banks serve financial specialism.

Société Générale is a bank with 150 years’ experience in "fostering business and industrial growth", the focus lies on entrepreneurs. Moreover the company focusses lies in innovation and team spirit.

Started off as in March 1865, as a local bank in Hong Kong Active in 60 countries with 54 million customers. The core strengths of the company are its capital control, cost control and long term relationship building.

In 1969 a merger between Standard bank of British South Africa and the Chartered Bank of India, Australia and China resulted in beginning of Standard Chartered Bank. The aim of this merger was to accommodate growth in trade between Europe, Asia and Africa.

Financial Times 1,984 (36) 3,378 (57) 1,977 (68) 4,901 (40) 2,190 (47) 717 (31)

Search Criteria Banco Santander Barclays PLC Crédit Agricole SA Société Générale HSBC Holdings PLC Standard Chartered Zephyr database and

company history

84 (19) 63 (24) 60 (18) 85 (22) 44 (19) 23 (17)

Annual reports 03'-13' (10) 00'-13' (13) 03'-13' (10) 05'-13' (8) 01'-13' (6) 01'-13' (9)

Table 3: Overview of selected cases + different sources used for the collection of data (2000-2014)

(25)

25

Six MNEs in the West-European banking commercial service sector are identified with different regional strategies. These firms are: HSBC holdings, Standard Chartered Group, Société Générale, Banco Santander, Barclays and Crédit Agricole. The sample includes two home region oriented firms, two host region oriented and two bi-regional firms, and therefore, three cases defined by regional orientation are analyzed (Table 3). As stated by Gibbert et al. (2008), in light of a post-positivist research, case study research must have validity and reliability. To increase the external validity or generalizability, the sample has a similar home region across cases and similar triad region preferences within cases. This is to account for culture distance within the European Union.

3.3 Data Collection

With respect to the data collection, multiple data collection methods are used. Both qualitative and quantitative can have a cooperative influence on the case study (Eisenhardt, 1989b). The use of multiple methodological strategies can attain the complexity of management research (Shah & Corley, 2006). Moreover, it provides a more accurate and analytical generalizable base for extending theory (Shah & Corley, 2006; Yin, 2009). The use of multiple methods refers to the process of triangulation, as it increases validity through the use of multiple methods (Hussein, 2009; Yaesmin & Rahman, 2012). Consequently, a more profound image of regionalization and location choice on a sub national level is given.

First, longitudinal data obtained from annual reports provides the sales, asset and employee percentages of these years across regions. The time frame is 2007 to 2013; this is chosen to provide a broader overview of the regional strategy and the availability of annual reports. By further expanding the regional orientation, this research increases its rigor. Hence, analytic generalization about the different regional strategy of the MNEs is possible (Yin, 2009).

Secondly, qualitative data is used over a time period of 2004 until 2012. More information can be found in table 3 and appendix C. With the use of Zephyr database, companies’ merger and acquisition (M&A) activity is collected. The minimal requirement is 50% as the MNE needs to be in control. In addition, the activity is screened before being taken into account in order to keep out all investment banking related activity. Moreover, if available, data from the annual reports complements the M&A activity with further information and adds Greenfield activity. Hence, a list of location choices is developed. Furthermore, with the use of the LexisNexis database, newspaper articles from the Financial Times London are found to

(26)

26

analyze the location choice decisions by MNEs. These articles are collected from dates between 1-1-2000 and 04-05-2014. In addition, the company’s history from the website and annual reports within this time frame are analyzed to further unfold the motives for location choices of the MNEs. As a result of using multiple data sources, the different motives of the working proposition can be identified and also data triangulation is utilized (Yin, 2009).

Lastly, by providing city data, another theoretical perspective is taken. Therefore, triangulation is also applicable due to the combination of theoretical perspectives (Hussein, 2009; Yin, 2009). To collect data in order to categorize the cities several sources of data are used. First, data of the population is obtained from the OECD database and a city population website that collects population numbers from several city reports. Moreover, the GaWc research database provides data on the global capacity, in terms of global connectives. A more detailed overview of sources is shown in appendix D. The arguments for choosing these variables and sources are explained in the next chapter.

3.4 Data analysis

This part discusses the analysis for the different data used, the regionalization data (quantitative data), the categorization of cities (quantitative data) and FDI motives (qualitative data).

3.4.1 Regionalization

First of all, quantitative, longitudinal data of the regionalization between 2007 and 2013 is analyzed by providing a wider picture of the regional orientation within and across the different cases. The different regional strategies are derived from Rugman & Verbeke (2004). The focus lies on the dispersion of sales, assets and employees across regions. Hence, it reveals the regional presence and the possible shift in orientation that occurred.

3.4.2 Categorization of cities

The second part of the analysis continues with processing the different aspects of the city as a location. This can be attributed to distinguishing between features of table 1 and the special institutional factors that a city conveys. The different features of a city that distinguish them from each other are shown before in table 1: global capacity (Sassen, 2001), national consolidation, major institutional equity holding and the presence of a liberal market (Sassen,

(27)

27

Case 1: Bi-regional Case 2: Home region Case 3: Host region

Banco Santander Barclays Crédit Agricole Société Générale HSBC Standard Chartered

Bilbao 3 Beijing 2 Athens 2 Athens 2 Asunción 3 Adelaide 3

Birmingham 3 Bridgetown 3 Brussels 1 Brussels 1 Athens 2 Beijing 2

Buenos Aires 2 Cairo 2 Cairo 2 Bucharest 2 Birmingham 3 Cape Town 2

Casablanca 2 Charlotte 3 Casablanca 2 Budapest 2 Bishkek 3 Dhaka 2

Dallas 2 Delhi 2 Copenhagen 2 Cairo 2 Cairo 2 Dublin 2

Frankfurt 1 Edinburgh 3 Kiev 2 Casablanca 2 Chester 3 Durban 3

Leicester 3 Guersney 3 Limassol 3 Chisinau 3 Dusseldorf 3 Hong Kong 1

Lisbon 2 Hong Kong 1 Lisbon 2 Geneva 3 Fuzhou 3 Hsinchu 3

London 1 Islamabad 3 Lyon 3 Hamburg 2 Hamilton 3 Istanbul 2

Madrid 1 Isle of Man 3 Milan 1 Kiev 2 Istanbul 2 Johannesburg 2

Mexico City 2 Jakarta 2 Novi Sad 3 Milan 1 London 1 Karachi 2

Miami 2 Jersey 3 Paris 1 Moscow 2 Luxembourg 3 London 1

Philadelphia 2 Johannesburg 2 Parma 3 New Delhi 2 Makati 3 Mumbai 2

Poznan 3 Karachi 2 Pordenone 3 Paris 1 Mettawa 3 New York 1

San Juan 3 Lahore 2 Roubaix 3 Podgorica 3 Mexico City 2 Seoul 2

Santiago 2 London 1 Sofia 3 Prague 2 Paris 1 Shanghai 2

Sao Paulo 1 Milan 1 Spezia 3 Saint Denis 3 Singapore 1 Singapore 1

Torino 3 Moscow 2 Tokyo 1 Sao Paulo 1 Taipei 2

Wroclaw 3 Mumbai 2 Split 3

New York 1 Tbilisi 3

North Highlands 3 Toronto 1

Paris 1 Tunis 3

Singapore 1 Wroclaw 3

Wilmington 3

Zaragoza 3

Table 4: list of city's entered between 2000 and 2013 (excluding investment bank related activity)

Source: See appendix D.

(28)

28

1999). In line with these characteristics, variables are identified that are used to operationalize the characteristics shown in Appendix A and B. As can be seen in table 4, the different cities (total of eighty-eight) are distinguished per firm and are categorized in three classes, as is discussed further. More details about the used sources are shown in appendix C.

The first characteristic refers to the global capacity of the city and the global control it can exert (Sassen, 2001). The global cities as identified by (Taylor, 2001) give the possibility to provide a quantitative layer on identifying the capacity of cities. This is done in terms of the global network, as it identifies the global city as a center of knowledge constellation (Taylor, 2001). The measurement identifies the world economy as a network in which firms distribute their services (Taylor, 2001). The interlocking network model used by Taylor (2001) uses the links between directors of corporate boards, and the key decision maker. As the corporate boards possess the most control and are not influenced by local governments (Taylor, 2001) it is interesting to see how many connections corporate board possess in cities. Moreover, the data used connections between advanced service firms such as law, accountancy, advertising, insurance, management consultancy and finance. To broaden this research an extra layer for the connectivity of the global city is added. This attains to the relative centrality of the city in the global network which is based on air traffic (passengers). This type of measurement is common and acceptable in measuring differences in the connectivity of cities (Smith & Timberlake, 2001). The data for global connectivity is derived from the database of the GaWC research network.

The second characteristic refers to the national consolidation (Sassen, 1999) of a city through the growth in population across years (if available). This refers to the importance of the city within its country boundaries; therefore if a city is growing increasingly and is the capital of a country, this provides an indication of the national consolidation of the city. Moreover, the third characteristic indicates the presence of major institutional equity (Sassen, 1999). If the city is a capital of a country or province, this makes the presence of institutional equity probable. Furthermore, the presence of top two hundred universities (Shanghai listing) and/or stock exchange market increase the institutional equity of a city. Lastly the characteristic of the liberal market opportunity of the city is made more specific to this research. A link is drawn from the European Union as home region and its trade agreement with the country in

(29)

29

which the city is embedded. Hence, the trade agreement indicates an important factor of entry for banks and therefore must not be neglected.

As indicated in the final column in appendix A and B, three classes are made in accordance to table 1. The first category cities need to provide all aspects mentioned above with most importantly the presence of top universities; a capital market and connectivity (prop. of 0.5) as this seems to be most importantly overlapping with a possible strategic asset seeking motive. For example, in spite of the relative centrality of Seoul and Miami, due to their lower connectivity (Lower than 0,50 proportion of the highest connectivity), this research doesn’t address them as category one cities. Moreover, category two cities also need to possess certain features of institutional equity and at least one and a half million people in 2010/2011. The minimum population is to distinguish between the more market seeking motive for category one and two on one hand and on the other hand the efficiency seeking motive of category three. An example of an exception is Fuzhou which is categorized in the third category. However this city embeds around four million people but does not possess the necessary equity to attain the second category.

Nodes Subnodes Working Propisition

FDI Motives Strategic Asset seeking 1

Market seeking 2

Efficiency seeking 3

Location Characteristics Category 1 City 1

Category 2 City 1+2

Category 3 City 3

Region Africa & Middle East Region Asia incl. Oceania Region Europe

Region Latin America Region North America

MNEs Banco Santander

Barclays Crédit Agricole Société Générale HSBC

Standard Chartered

(30)

30

3.4.3 FDI motives

The third part is analyzing the qualitative data. As Yin (2009) argues, the case study approach answers the research question of ‘How’ and ‘Why’. Therefore, a more deductive approach is chosen and as a consequence, qualitative analysis must adapt this approach to form particular procedures. The data is coded in line with the coding scheme of table 5. This is to reduce the data in a critical manner; this is deductively driven through the working propositions (Miles & Huberman. 1984). However, the nodes are not only linked to the working propositions, they are also coded as a way to store the data, as a way to manage the data more easily (Miles & Huberman, 1984).

As argued by Yin (1981), coding can establish cross-case comparisons. The qualitative strategy focuses on the following theoretical propositions (Yin, 2009), which provide the guide lines for analyzing the data. Furthermore, pattern matching is used to find similarities within cases of MNEs that possess the same regional orientation (Yin, 2009; Gibbert et al., 2008). Consequently, argumentation of possible causal relationships of location choice can be supported. In line with Miles & Huberman (1984), a descriptive matrix over the time period studied will support the analysis. The descriptive matrix combines the regionalization data (quantitative) with the amount of motives per regions and category of cities; this is done to structure the chain of evidence (Pratt, 2009). This matrix forms the basis for the cross case comparison. In addition, a more explanatory matrix is added in the second part of the tables to provide explanation for the reasons of the location choice motives (Miles & Huberman, 1984). The explanation consists of short aggregation of the arguments found in the articles. Lastly, the results are written down in a consistent way, by separating the categories of cities.

4. Results

The next part discusses the results of this multiple case study. First a categorization is made between different cities as a location choice (Appendix A/B) to further enhance within and cross case analysis. Secondly, three separate cases are done studied to distinguish between the regional orientations. Within each unit of embedded unit of analysis, two tables are presented. The first table shows the different motives and sales across regions and years. The second table provides more details on the different motives for choosing a location. The final part consists out of a cross case analysis to identify differences or similarities across cases.

Regional orientation of European commercial service banks and the importance of cities as a location choice