FDI motives and city location preferences in the automotive and commercial banking industries

Abstract

Purpose

Cities remain an understudied unit of analysis for understanding the motives of multinational enterprises’ (MNE) foreign direct investment (FDI), with subnational locations in International Business (IB) research to date predominantly captured via the phenomenon of agglomeration. As regional integration projects, such as the European Union and to a lesser degree NAFTA, increasingly reduce the importance of national institutional environments, this paper argues regional and subnational levels become more important for studying MNE location choice. This paper aims to evaluate the explanatory contribution of regional and subnational levels of analysis to understanding MNE location choice.

Design/methodology/approach

A qualitative deductive bottom-up multiple-case study research design is adopted to study the city location choices and FDI motives of six automotive and six commercial banking companies. These purposefully sampled manufacturing and service MNEs have different home countries and regional orientations. Data on their foreign investments across the extended Triad of Europe, North America and Asia-Pacific were collected for the time period of 2000–2021.

Findings

Findings suggest that different classes of city tend to attract specific types of FDI and that these patterns might vary across sectors and be influenced by the regional strategic orientations of MNEs. Industry-specific findings reveal the importance of related and support industries and partners in a city location for the automotive MNEs, while the commercial banks seek investment opportunities in cities that allow acquisition targets that have an attractive customer based and will improve their local market knowledge.

Originality/value

The findings provide evidence in support of MNEs in manufacturing and service industries perceiving the attractiveness of three city types in different ways across the Triad regions.

Citation

Danes, D., van Eijck, P., Lindeque, J.P., Meyer, M.A. and Peter, M.K. (2023), "FDI motives and city location preferences in the automotive and commercial banking industries", Competitiveness Review, Vol. 33 No. 3, pp. 602-626. https://doi.org/10.1108/CR-03-2022-0040

Introduction

This paper integrates the two concepts that most actively seek to rescale the level of analysis for understanding multinational enterprises’ (MNEs) geographical activities, namely, cities as locations (Belderbos et al., 2017; Castellani et al., 2022; Goerzen et al., 2013; McDonald et al., 2018) and regionalization (Rosa et al., 2020; Rugman and Collinson, 2004; Rugman and Verbeke, 2004, 2008). Adopting a multiple-case study research design, the international location choices of six manufacturing MNEs in the automotive industry and six service MNEs in commercial banking are analyzed for their location preferences to answer the following research question:

How do patterns of foreign direct investment vary across city types and macro regions for manufacturing and service industry multinational enterprises?

The findings provide evidence in support of MNEs in manufacturing and service industries perceiving the attractiveness of three city types in different ways across macro regions. In contrast to previous studies, this paper restricts the number of MNEs to study their city/regional location preferences between 2000 and 2021 for a comprehensive set of investment transactions and motives. Theoretically, the paper advances our understanding of city location and investment motive preferences by reconciling the early work of Hymer (1972) with contemporary research on cities and MNEs.

Subnational locations in international business research

Location is a central conceptual category in the study of international business and the multinational enterprise (MNE) (Rugman et al., 2011). As the focal unit of analysis has shifted from the country to the MNE and increasingly MNE subsidiaries over the first 50 years of international business research (Rugman et al., 2011), Beugelsdijk and Mudambi (2013: 1) noted that “location (L) in international business (IB) is almost always conceptualized and operationalized at the country level of analysis.”

Country as the defining location category has been extensively used in studies of internationalization strategies/location choice (Dunning, 2000; Johanson and Vahlne, 1977; Vernon, 1966) and conceptualizing location advantages in the home (Rugman and Verbeke, 2004, 2008) and host countries (Dunning, 1998) in terms of country-specific advantages (CSA). While both home and host country advantages contribute to the firm-specific advantages of the MNE, it is the choice of host country location(s) that offer the MNE discretion in choice of where to operate (Schotter and Beamish, 2013). Clear evidence has been provided that “(host) country” is a significant explanatory dimension of firm performance (Chan et al., 2008). Importantly, Rasciute and Downward (2017) provide evidence that host country attributes differ in their importance depending on industry and firm characteristics. Nachum and Wymbs (2005) similarly argue that there is evidence that different firms can assign different value to the same location attributes.

Exceptions to this country level of analysis-focused approach do however exist and include the work on agglomeration (Donnelly and Manolova, 2020; Hutzschenreuter et al., 2020; Nielsen et al., 2017), clusters (Birkinshaw and Hood, 2000; Donnelly and Manolova, 2020; Kim and Aguilera, 2016; Nielsen et al., 2017; Rugman and Verbeke, 2003) and the regionalization literature (Donnelly and Manolova, 2020; Nielsen et al., 2017; Rugman and Verbeke, 2004). Clusters and agglomeration processes are arguable the subnational spatial categories that have received most interest with respect to MNE investment decisions, even though global cities have been receiving growing interest, see Table 1 (Belderbos et al., 2017, 2020; Blevins et al., 2016; Chakravarty et al., 2021; Goerzen et al., 2013; Hutzschenreuter et al., 2020; McDonald et al., 2018). Evidence suggests subnational locations in different countries now appear to compete directly with global city locations for foreign direct investment (FDI) rather than being a necessary second step in the decision-making process (Chakravarty et al., 2021).

Motives for location choice in international business research

Dunning’s (1998) four FDI motives of resource seeking, market seeking, efficiency seeking and strategic asset seeking FDI have been broadly adopted in the location choice literature, including studies on cities as locations (Chen and Yeh, 2012; Kim and Aguilera, 2016). The nature of these FDI motives is argued to have changed under the growing importance of knowledge and cooperation with stakeholders under conditions of increasing global economic integration (Dunning, 1998). Thus, each of the four FDI motives is (today) argued to include influences from the 1970s, but also to reflect the location characteristics MNEs value, when viewed from the interorganizational network (Kim and Aguilera, 2016) and knowledge-based (Hutzschenreuter et al., 2020; Kim and Aguilera, 2016; Nielsen et al., 2017) views of the MNE.

Resource seeking investment thus would still include a consideration of the cost and quality of inputs, the quality of infrastructure, the regulatory environment and investment incentives, but now also covers assessments of the possibility to improve the quality of inputs and their processing locally, as well as the availability of local partners for knowledge exploitation and risk sharing (Dunning, 1998). Market seeking investments are argued to be attracted by the breadth, depth and accessibility of national markets, but now also by access to regional markets (North American Free Trade Area [NAFTA]/European Union [EU]), educated and skilled labor for selling advanced products, the presence of related and support industries and access to customers in knowledge-intensive sectors (Dunning, 1998). Dunning (1998) therefore argues that in addition to the manufacturing related efficiency seeking motives related to costs, agglomeration efficiencies and investment incentives, MNEs now also seek investment locations that bring efficiencies to knowledge-intensive work such as research and development (R&D), advanced factors of production provided through an educated work force and the possibility to benefit from the economies associated with knowledge-intensive specialized spatial clusters. Finally, strategic asset seeking investments include knowledge-related assets and markets to enhance and protect MNE firm-specific advantages (FSA), a relatedly supportive institutional environment, the exploitation of geographically dispersed knowledge-intensive FSAs, opportunities to access local knowledge-based CSAs and exploit differences across countries (Dunning, 1998). Such strategic asset may be either location or non-location bound (Sutherland et al., 2020). These four motives may be pursued independently or more likely in different combinations depending on the needs of the MNE and characteristics of different locations.

Cities and multinational enterprise location choice

The studies in Table 1 contribute to the integration of international business scholarship with the literature on global cities (Sassen, 2001, 2005) and the world city network (Taylor, 2001). Cities are defined as “spatially organized socio-economic systems [that are] places and sites rather than actors” (Friedmann, 1995, p. 22). Importantly, cities are not defined by “natural boundaries,” rather the boundaries of cities are “wholly the artifacts of the cities at their nuclei” (Jacobs, 1984, p. 45). All cities are thus embedded in local city regions subnationally, which can vary in size and importance according to their integration into the economy of the city (Lorenzen et al., 2020), while at the same time some cities play significant roles in the global economic system, called world or global cities (Friedmann and Wolff, 1982; Hymer, 1972; Sassen 2001).

Chakravarty et al. (2021) suggest that cities can be divided into a demographic (with sociological and environmental impacts) and a functional perspective (i.e. global cities). Cities are featured by local and international linkages that enable both material and information flows that accumulate in and around cities; however, much more is known about the nodes (cities) than the flows (Beaverstock et al., 2000), such as high wages and taxes (Nielsen et al., 2017). The availability of city attribute data has allowed the ranking of cities into a hierarchy in terms of their role in command and control of the global economy (Beaverstock et al., 2000), with world/global cities dominating this system. The link between MNEs and global cities made in the ranking by Beaverstock et al. (1999) of global cities, based on the office location choices of major, advanced producer-service firms (accounting, advertising, finance and legal firms), is increasingly being advanced, see Table 1.

The papers in Table 1 represent the ten most recent empirical research projects focused on cities as subnational locations for international business activities. These projects can be described in terms of their research focus, the type of city included in the study, the heterogeneity/homogeneity of the analyzed cities, the degree to which there was an industry focus and finally, the number of national context that were addressed.

None of the research in Table 1 focused on a single city, but Nachum and Wymbs’ (2005) study on agglomeration and mergers and acquisitions covers two global cities (New York and London) and is the only study addressing multiple cities that studies only one city type. Five studies focus either on FDI location choices and motives in a broader sense (Belderbos et al., 2020; Goerzen et al., 2013; McDonald et al., 2018), cities as locations for regional headquarters (HQ) (Belderbos et al., 2017) or locations for R&D investments or HQ investments (Castellani et al., 2022). These studies seek to understand location choices for a heterogeneous set of city types, but they differ in the degree to which these cities are in a single or multiple national contexts. Goerzen et al. (2013) address different types of world cities, their metropolitan areas or other peripheral cities as location for Japanese MNEs across multiple national contexts. McDonald et al. (2018) seek to understand the preference for peripheral versus core cities in China. Belderbos et al.’s (2017) study of regional HQ location choices focus on different types of global cities across multiple countries, and Belderbos et al. (2020) study the preference of MNEs for different types of global cities across multiple countries and for different investment motives. Castellani et al.’s (2022) research on metropolitan locations in the USA for R&D investments or for HQ investments also covers different city types. Importantly, all five studies, except that of Goerzen et al. (2013) that addresses manufacturing industries, focus on both manufacturing and service industries.

Two papers covered multiple city types with heterogenous characteristics as locations for both manufacturing and service MNEs across multiple countries: the first to understand the role of institutional change and entry mode preferences in host countries (Blevins et al., 2016) and the second to understand the influence of the institution of friendship cities on outward investments from China (Zhang et al., 2020). Chen and Yeh (2012) seek to understand the pace of FDI into different types of cities in China, by studying the location choices of Taiwanese manufacturing MNEs. Importantly, a macro regional framing has to date not been taken, reflecting the challenges related to data collection and capturing different motives for MNE location choices at the city level.

Hymer (1972) is in many ways a forerunner of the contemporary research on international business and city locations, which today actively draws on the broader world urban hierarchies’ literature (Beaverstock et al., 2000; Friedmann, 1986, 1995; Friedmann and Wolff, 1982). Hymer (1972) argued that the separation of different functions and their geographic dispersal has the potential to be most complete in the multidivisional firm. A central assumption in this literature is that the same location can be evaluated differently by firms in terms of attractiveness as a location choice because of the unique capabilities of each individual firm to exploit location characteristics (Nachum and Wymbs, 2005). Goerzen et al. (2013) also argue that it is not only leading global cities that matter to MNEs and that different city locations will be attractive as locations depending on the functions MNEs are seeking to transfer internationally. The attractiveness of cities as locations has also been shown to change over time as firms gain experience in a host country (Chen and Yeh, 2012).

There is arguably much in Hymer’s (1972) argument that anticipates later world or global cities arguments (Friedmann and Wolff, 1982; Sassen, 2001). Hymer (1972) conceptualizes business activity into level 1 top management decision-making, level 2 coordination of operational activities and level 3 management of day-to-day operations. Drawing on location theory, it can be expected that different locations will be more suited/attractive for locating each of these levels of MNE activities. Hymer (1972) does not address issues around differences between manufacturing (his focus) and service firms, which are added in the conceptualization of FDI motives for different categories of cities below, a key assumption in the regionalization literature (Rugman and Verbeke, 2008).

Development of working propositions

This paper seeks to observe and understand different investment motives across multiple city types and multiple countries by taking such a regional focus, but will make its own trade off by focusing on a single manufacturing industry (automotive) and single services industry (commercial banking). The advantage of this approach is that the patterns of investment motives across three different city types can be observed by industry type and macro region. Next, the paper turns to interpreting Hymer (1972) for contemporary city location choice, seeking to integrate recent research on cities as three different types/levels of location for international business activities, captured by the four FDI motives.

Each of the levels of MNE functions is associated with different location characteristics, investment motives (Dunning, 1998), which in this paper are argued to be associated with different degrees and types of urban location. Level 1 top management decision-making, the coordination of the internationally dispersed MNE subsidiaries, is expected to be concentrated in cities featured by easy access to “capital markets, the media, and the government” (Hymer, 1972), in other words, global cities such as New York, London or Tokyo (Sassen, 2001), to name the most exemplary cities. These cities create an agglomeration of creativity, learning and innovation which is supported by the cities’ distinct, formal and informal institutional features (O'Neill, 2003). Level 2 coordination of operational activities, are expected to be located where the MNE can gain access to educated employees, connect to communication networks, participate in information exchange, and is argued to most likely be found in large cities featured by agglomeration of firms from different industries (Hymer, 1972). Here, the agglomeration is featured by the benefits of diversity. Level 3 management of day-to-day operations, are argued to be pulled to locations featured by basic factor endowments, labor and markets. These activities are expected to be located in what we now refer to as less developed and emerging economies (Hymer, 1972). The MNE is expected:

[…] because of its power to command capital and technology and its ability to rationalize their use on a global scale, will probably spread production more evenly over the world’s surface. (Hymer, 1972, p. 908)

Thus, MNE top management decision-making is expected to be located in tier 1 cities or global cities (Sassen, 2001), featured by global connectivity, advanced producer services and a cosmopolitan environment, high information and material flow, highly skilled workforce, a large and sophisticated market, presence of expatriates and links with other global cities, financial markets and highly ranked universities (Goerzen et al., 2013). There is a broad range of differences among cities (Goerzen et al., 2013), including their “global capacity” or capacity to exert global control (Sassen, 2001). Global cities enable MNE top management teams (TMT) to participate in both internal and external information exchange and corporate and global control processes, through access to information exchange and drawing on the large number of local experts (Goerzen et al., 2013), and additionally to provide access to capital markets, the international transportation network and influential decision-makers outside of the firm (Beaverstock et al., 1999; Friedmann, 1986). Reflecting the argument that global cities create value from the “flows through them,” and not because of what “stays within” the city (Sassen, 2005), although more static features such as population size may also be attractive as a market. These cities are usually located in developed economies and featured by a highly skilled and specialized pool of potential employees (Iammarino and McCann, 2013) and their development is driven by the global economy more than the national economy (Nachum and Wymbs, 2005):

WP1.

MNE investments in tier 1 cities will predominantly be featured by strategic asset and market seeking motives.

MNE coordination of operational activities are expected to be associated with tier 2 cities that can supply the skilled white-collar employees and communication networks required for this function. Tier 2 cities are featured by large markets, strategic locations, specialized economies, the presence of industry partners, less concentration and sometimes these cities are the capitals of the relevant countries. In many ways, these activities are going to be found in cities that are distinguished by the absence of key features of global cities. Thus, cities that would be attractive for investment to enable the coordination of operational activities will be in large cities that are well connected, but that do not have those features required for the execution of TMT strategic decision-making. Markusen et al. (1999, p. 3) describe tier 2 cities as “spatially distinct [metropolitan] areas of economic activity where a specialized set of trade-oriented industries take root and flourishes” and may be associated with MNE investment motives for innovation, market seeking or cost reduction. Tier 2 cities have a large population, their economy is significant but the material and immaterial flows are significantly smaller than tier 1 cities (Beaverstock et al., 1999; Hymer, 1972). The second tier of cities, because of their large populations and expected higher levels of urbanization growth, are expected to be attractive markets for MNEs (Dobbs et al., 2012). Although these cities are not global cities, they are still important cities within a country or a region, from where MNEs can serve an entire nation or region (Beaverstock et al., 1999). It is argued that they can be expected to have higher national or regional connectivity compared to tier 1 cities and be ideal locations for regional headquarters, their development is expected to be dependent on their home country (Friedmann, 1986) and region. Tier 2 cities are far more specialized than global cities with agglomerations of many suppliers and business partners for different industries (Hymer, 1972) and contain highly specific assets such as plants and facilities, which represent tangible location-bound (LB) FSAs that are attractive for MNEs:

WP2.

MNE investments in tier 2 cities will predominantly be featured by strategic asset, efficiency and market seeking motives.

All other cities are classified as tier 3 cities and are argued to be associated with investments related to the management of day-to-day operations of the MNE subsidiaries. Such tier 3 cities are featured by cheap labor, resources, local embeddedness, strategic positioning (proximity), infrastructure and available land. These cities are particularly attractive because of different factor endowments, such as cheap labor, land or the presence of natural resources, but the presence of key transport and communications infrastructure is a precondition for MNEs to internationalize to these cities. Tier 3 cities may have a strategic geographic position within a region or a country that allows the internationalizing MNE have better and cheaper access to a region and will therefore be attractive to MNEs seeking to make their operations more efficient. Cities are expected to contain tangible FSAs, such as plants, factories and manufacturing facilities. Tier 3 cities are therefore expected to be especially attractive for manufacturing MNEs seeking economies of scale and low transport costs (Cantwell and Piscitello, 2002):

WP3.

MNE investments in tier 3 cities will predominantly be featured by resource and efficiency seeking motives.

All cities are featured by an associated city region. These city regions are important enablers of any agglomeration advantages of a city and are expected to be important in the location decision of MNEs (Arregle et al., 2009; Goerzen et al., 2013). These represent LB city-specific advantages that MNEs recombine with their (partially) internationally transferable FSAs that will exhibit industry specificity. Industry and firm characteristics have also been argued to function as boundaries for institutional influences across investment locations (Donnelly and Manolova, 2020), while more advanced infrastructure available in a city will determine its attractiveness to manufacturing and service MNEs (Nielsen et al., 2017). The importance of FSA specificity is clearly associated with the differences between manufacturing and service MNEs in terms of city location preferences:

WP4.

The patterns of investment motives across the city categories will differ between manufacturing and service MNEs.

A growing literature is questioning the degree to which different city types in developing, emerging and developed economies will be perceived to be the same in terms of the have city-specific advantages of interest to firms (Krätke, 2014). Krätke (2014) critiques the focus of the global cities literature for its focus on finance and advanced producer services, instead looking to indicators of connectivity related to global value chains and global production networks to reveal multisectorial networks of connectivity between cities in different global regions, emphasizing the role of MNEs. Critically, Krätke (2014) advances the argument that the patterns of connectivity across cities globally will be strongly affected by the sectorial specialization of a city, resulting in city-specific advantages of a given city location. While Krätke (2014) is not able to provide detailed specifics on these patterns, recent work by Jedwab et al. (2022) in a policy research working paper for the World Bank allows greater insights in this regard. Jedwab et al. (2022, 2) in their study, which covers 65 countries and sectorial compositions for 7,000 agglomerations, show that “for a given population size and a given level of urban economic development, cities dramatically vary in their sectoral composition across the world.” These differences include the degree to which a city produces tradable manufactured products and/or finance, insurance, real estate and business services (FIRE) or has a significant nontradable urban economy (Jedwab et al., 2022). Their publication suggests these patterns differ significantly over the macro global regions, which reflects a growing stream of research that calls into question if the North American and European centric Global Cities and World Cities Networks literatures can capture the unique urban experiences of developing and emerging economies and critically the theory of convergence of developed and emerging economy cities (Ghadge, 2019; Robinson, 2002). Shatkin (2007) argumentatively and Jedwab et al. (2022) in their mapping of producer, consumption and neutral (balanced) city types across the world, provide clear arguments and evidence for expecting differences in the attractiveness of the city categories in this study across the extended Triad regions. The work of Schotter and Beamish (2013) on location shunning suggests another mechanism that at the firm level could lead to location choices that lead to differences in city category preferences for the same FDI motive across the extended Triad regions. While Blevins et al. (2016) also suggest that regional economic integration is important for understanding the relevance of cities as locations for FDI. The different degrees of progress with economic integration across the Triad macro regions suggests that there could be different outcomes in terms of how the focal MNEs invest across city types in the extended Triad regions. Arguably, the EU is the most advanced integration project, followed by the NAFTA, the Association of Southeast Asian Nations and Asia-Pacific Economic Co-operation in Asia-Pacific (Warleigh-Lack, 2015).

Distinguishing the regional nature of manufacturing and services firms for different city types by macro region, the study adds more nuance to the findings of studies such as that by Mehlsen and Wernicke (2016) on the liability of foreignness and MNE preferences for global cities, which do not consider the context within which investment location choices are being made, nor the associated investment motives:

WP5a.

The patterns of investment motives across the city categories will differ by extended Triad region.

International business research on cities as locations (Belderbos et al., 2017, 2020; Blevins et al., 2016; Castellani et al., 2022; McDonald et al., 2018) recognizes the need to focus on both manufacturing and service industry investment decisions. At the same time, the regionalization literature (Rosa et al., 2020; Rugman and Verbeke, 2004) has shown clearly that the macro regional investments of MNEs are affected by the differences in the nature of the FSAs of manufacturing and service MNEs. Additionally, Krätke (2014, p. 125) has argued that “an inter-urban network formed by global manufacturing firms will differ from one formed by global firms providing advanced producer services.” Krätke (2014) advances the argument that, because of an overemphasis on service industries in previous research, the different patterns for manufacturing firms in terms of activities across cities and their role in connecting cities globally has been underappreciated. Based on this consistent emphasis on and evidence for differences between manufacturing and service MNEs in all three of the domains of research on which this paper draws, our final proposition argues that across the macro regions, we would expect to see a difference in the choice of city types for investments by manufacturing and service MNEs:

WP5b.

The patterns of investment in cities across the extended Triad will differ between manufacturing and service.

Having established the main expectations for how MNE location choices across the three different types of cities will differ, for manufacturing and service MNEs and across the Triad regions, the paper now provides an account of the research design adopted.

Research design

A qualitative multiple-case study design (Eisenhardt, 1989; Yin, 2003) is adopted for this study. The design draws on a strong theoretical foundation for developing working propositions that serve to draw attention to the central themes of the study (Yin, 2003). The design allows for credible theoretically embedded incremental theory development (Ridder et al., 2012) that extends existing theory (Eisenhardt and Graebner, 2007), specifically with respect to the degree to which FDI location motives can be associated with specific types of cities. The research design described here is presented in more detail in a supplementary online material file (to meet the word limit of the journal).

Theoretical case sampling strategy

This study theoretically samples (Eisenhardt, 1989; Eisenhardt and Graebner, 2007) multiple cases (MNEs) with embedded units of analysis (city location choices and investment motives), sampling six manufacturing and six service sector MNEs and analyzing their city location choices for the period 2000–2021, improving external validity of the study (Yin, 2003). This design reflects the evidence in the literature that location choice at the subnational level needs to be understood in relation to both industry characteristics (manufacturing or services) and firm characteristics (selected MNEs) (Donnelly and Manolova, 2020; Rasciute and Downward, 2017). Finally, the regional nature of MNEs international orientations (Rosa et al., 2020; Rugman and Verbeke, 2004), reflected in the relative importance of regional conditions for MNE subnational location choices (Arregle et al., 2013; Nielsen et al., 2017) is reflected in the MNEs sampled for the study. With these considerations in mind, we explain the sampling of our focal MNEs below (Figure 1).

Six automotive companies and six commercial banking MNEs with different home countries and regional orientations are sampled, reflecting the regional nature of MNEs (Blevins et al., 2016; Rosa et al., 2020; Rugman and Verbeke, 2004). A review of over 20 major management, strategy and international business journals from the CABS Ranking identified 87 studies focusing on the automotive sector and 230 addressing banking between 2012 and 2022. Given the emphasis on generalizing to theory (Yin, 2014) in qualitative studies, this makes the automotive and commercial banking sectors appropriate empirical cases of well-studied manufacturing and service MNEs. Automotive MNEs sampled include Toyota (Home Regional: Asia-Pacific), Mitsubishi (Home Regional: Asia-Pacific), Volkswagen (Home Regional: Europe), Daimler (Global Orientation: North America, Europe and Asia-Pacific), Ford (Home Regional: North America) and General Motors (Home Regional: North America), reflecting Rugman and Collinson (2004), with the exception of the exclusion of Honda and the inclusion of Mitsubishi. Six European commercial banks are included in the study and include HSBC Holdings (Host Regional: Asia-Pacific), Standard Chartered Group (Host Regional: Asia-Pacific and Middle East), Société Générale (Home Regional: Europe), Banco Santander (Bi-regional: Europe and South America), Barclays (Home Regional: Europe) and Crédit Agricole (Home Regional: Europe). It was not possible to select commercial banks from across the three Triad regions in the same manner as for the automotive MNEs, due to data collection difficulties, not least because many of the other potential candidate banks were not only home region oriented, but also had extremely significant home country orientations. However, because of the unique nature of the selected commercial banking MNEs, the regional coverage is similar to that for the Automotive MNEs. Overviews of the MNEs are available on request from the corresponding author in the detailed research design as part of the online supplementary materials for this paper.

Data collection

Multiple sources of data were used in this study (see Table 2), allowing method triangulation (Denzin, 1978) and thus improving the internal validity of the study (Yin, 2003). Data on the investments made by individual firms was collected from three different sources, namely, company annual accounts and websites, newspaper reporting by the Financial Times was collected from the LexisNexis database and investment data from the Orbis and Zephyr M&A databases of Bureau van Dyk. The use of databases and official statistics to identify suitable investment transactions is common in studies of cities and location choice (Belderbos et al., 2017, 2020; McDonald et al., 2018). Multiple years of data from company annual accounts was used to establish the regional orientation of the MNEs, reflecting the norms in the regionalization literature. Sales by region were used, but also assets and number of employees when available, to determine MNE regional orientations (Rugman and Collinson, 2004; Rugman and Verbeke, 2004).

Company annual accounts were used to collect data on FDI decisions that used entry modes such as greenfield investments, while investments by acquisitions, joint ventures or mergers were identified using the Orbis and Zephyr database of mergers and acquisitions provided by Bureau van Dyk. Minority stake investments were not included in the data collected and in the case of the commercial banks, and all the investment banking–related investments were screened out. Newspaper reporting was used to identify further investments not in these two sources opportunistically during their analysis. A list of FDI transactions for each firm was created from these sources, including details such as the year of the deal, the location of the investment (at the city level where possible) and the type of deal (entry mode). The company accounts and websites, the deal editorials from the Zephyr database and the newspaper reporting were then used to collect qualitative data on the motive for the FDI according to Dunning’s (1998) categories. The annual accounts were also used to collect data on the regional strategic orientations (Rugman and Verbeke, 2004) of the individual MNEs, to enable this dimension to be considered in the analysis as needed for this study.

To collect data to categorize the cities, several sources of mediated data are used to collect data on city population size, whether a city is the capital, or has the presence of a university, a (sea) port or an airport. Three measures from the database of the Globalization and World Cities (GaWC) research network representing the degree of world connectivity of the cities (Taylor, 2001) are used. Selected industry specific measures are also adopted. See Table 3 for variables collected and how they were analyzed. The full categorization and a seperate data file are available online in the online supplementary materials for the paper.

Analytical strategy

The data collected for the individual MNEs were then analyzed to develop an analytical table (Miles and Huberman, 1984, 1994) for each firm capturing the tier of city invested in, the timing of the investment, the entry mode, the FDI motive and exemplary qualitative materials for the analysis. The analytical tables that capture the analytical categories of the working propositions allowing a descriptive matrix (Miles and Huberman, 1984, 1994) to be presented and structure the chain of evidence (Yin, 2003) for within and cross-case analysis (Eisenhardt, 1989; Yin, 2003). These analytical tables have been interpreted by at least two investigators for both types of firms, allowing a limited degree of investigator triangulation (Denzin, 1978). There were three stages to the analysis for developing the analytical tables; first the regional orientation of the MNEs was established for each MNE, next the FDI motive was identified and finally, the cities associated with each investment were categorized into one of the three tiers.

The regional orientation of the MNEs was established using the standard sales, assets, employees measures for the extended Triad and other major regions, primarily using MNE annual accounts (Rugman and Verbeke, 2004).

The FDI motives were identified by coding the qualitative materials collected in relation to each investment decision, similarly to Belderbos et al. (2017), using Dunning’s (1998) four-part typology of FDI motives (resource seeking, efficiency seeking, market seeking and strategic asset seeking). For each investment identified, the qualitative evidence describing the investment by a firm in a given city location was thematically analyzed using Dunning’s (1998) four-part typology of FDI motives. Based on this analysis, a final evaluation of the motives for the investment was assigned (motives were not considered mutually exclusive and so more than one could be applied).

The city tier category was assigned to each city in which an investment was identified, with the full list of city classifications is available in “Supplementary_Online_Materials_Document_1_-_Classification_of_Cities.” It is the most challenging of the analysis processes because the (world/global) city literature does not yet provide a generally accepted methodology for categorizing cities and positioning them into a hierarchy, and with that, one of the main limiting factors is limited availability of data (Beaverstock et al., 2000). The contemporary literature either adopts existing city classifications, as can be seen in the studies of global cities by Belderbos et al. (2017, 2020), Blevins et al. (2016), Nachum and Wymbs (2005) or world cities by Goerzen (2013). Three studies focus on multiple city types without emphasizing the global or world cities literatures. Chen and Yeh (2012) operationalize Chinese city locations with subnational regional data and distinguish between costal and noncostal locations, but otherwise use general economic location attributes. Zhang et al. (2020) use the institutionalized status of foreign cities as friendship cities or not in their study of outward Chinese FDI. While Castellani et al. (2022) do not classify the US metropolitan locations, their study is instead based on official statistics for each city location (available from official records) to describe each location. McDonald et al. (2018) distinguish between core and peripheral cities drawing on government classifications in a manner analogous to Hymer (1972). Goerzen et al.’s (2013) focus on types of world cities, and their metropolitan areas or other peripheral cities take a similar approach. Our paper is different to these previous studies in that we adopt an approach that first identified the city locations and then seeks to categorize them into one of three city types conforming to Hymer’s (1972) attractiveness due to the activities associated with the investment motive. In total, we categorized 218 city locations (see online supplementary materials for the data file).

The classification process we adopted was one of elimination according to the degree to which a city meets criteria related to global capacity (connectivity and control) and national consolidation or importance of the city within its country boundaries (Sassen, 1999) (see Table 2 for the criteria used to establish a city type with the data). Because tier 1 cities have been studied extensively, the cities in this category are typically unanimously accepted by a number of authors (Beaverstock et al., 2000; Friedmann, 1986), as adopted in the IB studies of global and world cities by Goerzen et al. (2013). Thus, the decision rules for the classification of cities in this paper recognize the criteria for tier 1 cities, but emphasis the criteria for distinguishing between tier 2 and 3 cities (see Table 3). Importantly, the classification of a city into a category is the result of a holistic assessment of the different city characteristics. Our goal therefore was to ensure the appropriate categorization in keeping with the classification of Hymer (1972) and the world cities literatures. Hymer’s (1972) criteria and the world cities literature emphasizes measures of global and national/regional connectivity. Practically, the procedure was to complete the necessary data collection for the development of the data set in the file “Supplementary_Online_Materials_Document_2_-_Cities Classification Data,” which included the same variables as specified in Table 3. The decision criteria specified in Table 3 were then applied and a city category applied to each city location. This was necessary because of the large number and diverse types of cities in our data set.

Findings by city type

Tier 1 cities and multinational enterprises location choice motives

The focal automotive MNEs made only a limited number of investments in tier 1 cities during the analyzed period, making it difficult to identify clear patterns in location choices by the firms (see “Supplementary_Online_Materials_Document_3_-_Table_S1a_Investments_in_Tier_1_Cities_by_Automotive_MNEs” for the investments by automotive MNEs in tier 1 cities). Most of the investments are associated with market and strategic asset seeking FDI motives, as was expected. There does seem to be evidence for regional variations, with Mitsubishi and Toyota making market seeking investments in Europe and both market and strategic asset seeking investments in their home region of Asia-Pacific. Interestingly, the investments of the North American and European firms are all in Asia-Pacific in the earlier years studied, while more recently European firms have been far more active in tier 1 cities in their home region. The regional variation again emerges, with both Daimler and Volkswagen making strategic asset and efficiency seeking investments, suggesting that the development of Asian economies still leaves room for efficiency gains in tier 1 city locations for MNEs from developed economies. While in Europe, these firms are pursuing strategic assets, associated with the contemporary transition in the automotive industry. Broadly, though the motives were as expected and there is no clear evidence of specific entry modes being associated with tier 1 cities for the automotive firms.

The commercial banking MNEs studied locate extensively in tier 1 cities during the focal time period. These European MNEs also predominantly invest in their home region tier 1 cities using acquisitions and joint ventures (see “Supplementary_Online_Materials_Document_4_-_Table_S1b_Investments_in_Tier_1_Cities_by_Commercial_Banking_MNEs” for the investments by commercial bank MNEs in tier 1 cities). HSBC makes relatively few investments in tier 1 cities, and if so, this is for efficiency, strategic asset and market seeking motives. Standard Chartered and Barclays make predominantly strategic asset seeking investments in tier 1 cities and Banco Santander, Credit Agricole and Society General make strategic asset and efficiency seeking investments in tier 1 cities in earlier years and more recently also market seeking investments. The lack of market seeking investments in the earlier years is interesting and speaks to the sector-specific characteristic of tier 1 cities for these MNEs at the time. Tier 1 cities associated with financial markets represent key sources of inputs for these MNEs and acquiring or starting joint ventures with other firms in these locations is arguably an example of the MNEs seeking to access the factor market for their industry.

There is thus evidence that tier 1 cities provide strategic assets for both manufacturing and service MNEs, but that for the automotive manufacturers, these cities are more likely to represent key markets than for the service sector commercial banks, which can realize efficiencies in these cities due to the unique symbiosis of the sector and tier 1 cities. There is evidence that for the automotive manufacturers there is variation between regions in the motives for investing in tier 1 cities, but this is less the case for the service sector commercial banks, notwithstanding the fact that most of their investments are home region expansions.

Tier 2 cities and multinational enterprises location choice motives

The investment motives of the automotive manufacturers reflect what was expected for tier 2 cities, with evidence of strategic asset, efficiency and market seeking motives (see “Supplementary_Online_Materials_Document_5_-_Table_S2a_Investments_in_Tier_2_Cities_by_Automotive_MNEs” for the investments by automotive MNEs in tier 2 cities). Market and efficiency seeking investments are made across all three regions, strategic asset seeking investments are made by Daimler in North America and by Volkswagen in Europe in earlier years in our data. More contemporary activities show far greater strategic asset seeking by Daimler in tier 2 cities in Europe and by all firms but GM in tier 2 cities in North America; these are highly related to the ongoing transitions in the automotive industry. There is also no clear entry mode preference associated with the investments of any one automotive MNE within or across the regions.

While investments in European tier 2 cities by the commercial bank service MNEs dominate these predominantly home and bi-regional MNEs, there is a much greater degree of investments in second and third regions in tier 2 cities (see “Supplementary_Online_Materials_Document_6_-_Table_S2b_Investments_in_Tier_2_Cities_by_Commercial_Banking_MNEs” for the investments by commercial banking MNEs in tier 2 cities). Although there are some limited instances of efficiency seeking investments by the banks, it is striking that strategic asset and market seeking motives predominate for these service firms in tier 2 cities. The commercial banks predominantly make use of acquisitions in tier 2 cities, and this does not seem to be region specific, although greenfield investments are also common. There does seem to be firm-level variation in the preferences for entry modes, for example, Banco Santander, Society Generale, HSBC and Standard Charter effectively only seem to use acquisitions across the three regions, whereas Credit Agricole uses acquisitions and joint ventures, and Barclays uses acquisitions and greenfield investments. In the case of Credit Agricole there is no regional effect, but Barclays made all the greenfield investments in the Asia-Pacific region for market seeking motives.

The findings speak to the unique role of financial markets in the banking sector, as discussed above, and points to the possibility of sector-specific variations in the FDI motives that can be expected for a given tier of city. There is also clearly a difference in the motives of the automotive manufacturers and the commercial bank service MNEs for investing in tier 2 cities.

Tier 3 cities and multinational enterprises location choice motives

As was expected, tier 3 cities predominantly attract efficiency seeking investments by the automotive manufacturers. Interestingly, there is no significant resource seeking investments and even in tier 3 cities, there are opportunities for strategic asset seeking investments. Predominantly, these are in Europe (see “Supplementary_Online_Materials_Document_7_-_Table_S3a_Investments_in_Tier_3_Cities_by_Automotive_MNEs” for the investments by automotive MNEs in tier 3 cities). There is evidence that individual automotive MNEs have different preferences for entry modes in tier 3 cities as well, with the large Toyota and Volkswagen clearly preferring greenfield investments, whereas the other firms use a broader range of equity entry modes.

Strategic asset seeking remains the dominant motive for commercial banking service MNEs in the tier 3 cities, although market and efficiency seeking motives may accompany these investments (see “Supplementary_Online_Materials_Document_8_-_Table_S3b_Investments_in_Tier_3_Cities_by_Commercial_Banking_MNEs” for the investments by commercial bank MNEs in tier 3 cities). Bi-regional Banco Santander and home region–oriented Credit Agricole and Society Generale predominately invest in European tier 3 cities with no clear pattern in different motives across the three regions. Bi-regional Barclays and HSBC and host region–oriented Standard Charter invest in tier 3 cities across the three regions and in South America or Africa and the Middle East to some degree. Again, there is insufficient evidence to suggest a systematic motive preference within a given region. Interestingly, in tier 3 cities, all the commercial banks make acquisitions, except for one greenfield investment by Barclays in the Asia-Pacific for market seeking reasons.

There are again clear differences in the FDI motives of the automotive manufacturers and commercial banking service MNEs in tier 3 cities, with the automotive MNEs better fitting the expectations for tier 3 cities. The continuing importance of strategic asset seeking for the commercial banks even in tier 3 cities arguably reflects the sector-specific nature of the FSAs needed to provide their services regardless of the tier of city.

Cities and the location choice preferences of manufacturing and service multinational enterprises

The findings of this study suggest that Hymer’s (1972) arguments fit the automotive manufacturers better than the commercial banking service MNEs (see Table 4). In the analyzed period, the automotive MNEs make far fewer investments in tier 1 and 2 cities comparison to the commercial banks. These sectors are arguable at very different stages of overall progress in internationalization. Despite the affinity of finance with globalization arguments, the commercial banks were arguably only beginning to expand internationally during the period 2000–2013, and this pattern continued in the subsequent period 2014–2021, while the automotive MNEs operate in a sector that has long been subject to internationalization and thus already had significant international presence. Firms in the automotive sector make some strategic asset seeking investments in tier 3 cities, regardless of their home region. Overall, the automotive firms clearly are focusing on investments in Europe and the Asia-Pacific, and with no clear home country effects during the years 2000–2013, but more contemporary investments are strongly focused on North America and Europe because these firms pursue strategic assets needed due to the technology transitions currently shaping the sector.

European commercial banks fit the original arguments by Hymer (1972) far less as well. They tend to make more efficiency than market seeking investments in tier 1 cities and their investments are dominated by strategic asset seeking investments across all tiers of cities. There does however seem to also be a clear home county effect for the European commercial banks because the UK banks make far more active investments in the Asia-Pacific region, followed by investments in North America, while admittedly this may be also simple be a case selection effect. The Spanish and French banks have a clear focus on South America and only limited investments in North America. The investments of the banks in Africa speak to a strong home country effect as a result of potentially previous colonial ties.

The far more heterogeneous patterns of investment motives across the city types for both the automotive firms and commercial banks makes clear evidence in support of working propositions 1, 2 and 3. MNEs arguably have a far more nuanced understanding of subnational locations at all three levels of city compared to what theory would accommodate currently. Firms are clearly seeking bundles of location advantages and investing in the city location that offers the optimal bundle of location advantages at all three levels of city type. Working proposition 4 is clearly supported because there are clear sectorial effects in the data and the patterns of investment in city type and specific city locations clearly differs. Finally, working propositions 5a/b also find some support, with the combinations of city type and extended Triad regions providing some evidence of differences, although these are not as clear as could be expected.

Discussion

For both the automotive and the commercial banks, there are underlying explanations for the patterns of investment observed. The findings for the automotive sector provide clear evidence of the location choice for these firms being influenced by the presence of industry partners in a given location. These firms tend to invest in locations where they already have a presence or where key suppliers, distributors and collaborators are already present, reflecting the analysis of the sector by Rugman and Collinson (2004). This is supported by Nielsen et al.’s (2017) empirical study, confirming that the higher the concentration of organizations in a particular industry is, the more likely it is that this location is chosen for FDI by firms in the same industry. Kim and Aguilera (2016) believe that the importance of network connections (suppliers and buyers) is much more pronounced in emerging markets where organizations are still in transition. Proximity to specialized resources such as industrial knowledge and labor pooling (Donnelly and Manolova, 2020; Nielsen et al., 2017), human capital, suppliers, collaborators and distributors allows MNEs to reduce communication and transport costs, while also increasing coordination with suppliers (Kim and Aguilera, 2016; Nachum and Wymbs, 2005). In addition, elements such as cheap labor and available land determined MNEs to adopt a mimetic behavior in selecting locations to place their day-to-day operations. Lastly, Donelly and Manolova (2020) propose “industry openness” as a framework condition and as a characteristic of FDI host location. All sampled MNEs have internationalized in clusters because of cheaper, faster and better access to specialized resources (Porter, 2000) and because most of the activities that were positioned in clusters were lower-level operations, the clusters were found predominantly in tier 2 and 3 cities (Hymer, 1972).

The findings for the commercial banks show support for the arguments of Buch (2000) that the most important asset for these MNEs is the possession of local market knowledge, reflected in the extensive use of acquisitions as an entry mode. The preference for acquisitions and a focus on tier 2 and 3 cities also reflects the need for banks to build trust and deliver services close to their customers (DeYoung and Roland, 2001; Kim and Aguilera, 2016). These MNEs would appear to prefer targeting midsized banks for acquisition, which provides them with a certain amount of assets in terms of customers, branches, employees, organizational capabilities or institutional equity (banking licenses and government ties). It is interesting to note that inside the banks’ home region, acquisitions are carried out to gain market access. These investments tend to be in tier 2 and 3 cities and represent a combination of strategic and market seeing investment, and further evidence that investment motives are not exclusive (Dunning, 1998).

The findings for this study suggest that the argument that tier 1 cities provide unique gateway function by offering lower transaction costs (Goerzen et al., 2013) may need more critical engagement in light of the findings for these established MNEs from both a service and manufacturing sector. Furthermore, there is clear support for seeking to more systematically integrating a more nuanced classification of city “types” as originally argued by Hymer (1972) and the very well-established global/world cities literature (Beaverstock et al., 2000; Friedmann, 1986, 1995; Friedmann and Wolff, 1982; Sassen, 2001, 2005). The adoption of this analytical categorization has provided an opportunity to gain new insights into the subnational location choices of established MNEs in two established global sectors. These findings additionally can be explained by the underlying mechanisms that drive value creation in these sectors, as explained above, providing further support for the findings in the study.

Our research therefore provides context to business activity as per Hymer (1972) in FDI, highlights the implementation of various market strategies by industry and preferences by city tier. Most importantly, it highlights a nuanced understanding of MNE location strategies.

Conclusion

Implications for research

The findings of this study suggest that IB theorizing could provide a much richer and fuller account of national and city-level influences on location choice decisions by MNEs. There seem to be some clear effects that would need to be considered if a city level of analysis is to be more systematically integrated into MNE location choice theory, with this study finding evidence of sectorial specific patterns (manufacturing/services), as well as that the overall regional strategic orientation of sectors matters. Furthermore, there is evidence of regional (developed market (DM)/emerging market (EM) markets) variation in the patterns. The data collection for this study also suggests that methodologically future research needs to be carried out in such a manner that longer time frames are addressed to accommodate temporal effects in terms of the stage of the industry internationalization phase. Conceptually, more work shall be invested on conceptualizing different levels of cities. Of course, there are often exceptions to far better understand why some lower tier cities offer attractive strategic investment opportunities. The conceptualization of FDI motives may also not be nuanced enough to allow more in-depth understandings of the FDI motive/national/subnational location interactions. In addition to examining the perspective on strategic decision-making by MNSs, an outcome perspective related to performance and broader economic development on social impact could be additionally explored.

Limitations of the study

The characterization of the cities could make better use of flows data in the categorization process, but the overall initial effort seems a reasonably good fit compared to GaWC methodology typically adopted. In this study, the data on city characteristics and FDI motives for each sector was only coded by one researcher, this makes ensuring intercoder reliability challenging, assessments in future research should improve construct and internal validity by making use of multiple coders.

Figures

Figure 1.

Multiple-case study research design with MNEs by regional orientation

Table 1.

Overview of international business research on cities in 2012–2022

IB research focus See articles index for paper citation	City type				Single city	Multiple city focus		Industry type			Single country	Multicountry
	Global cities	World cities	Peripheral cities	Multiple types		Homogenous	Heterogenous	Manufacturing	Service	Multi-industry
FDI location choice and motives	7	3	6	3, 6, 7			3, 6, 7	3		6, 7	6	3, 7
FDI location choice – MNE regional HQ	5			5			5			5		5
Institutional change and MNE market entry mode	4			4			4			4		4
FDI location choice – MNE R&D and HQ				9			9			9	9
Agglomeration	1					1			1			1
FDI pace				2			2	2			2
Outward FDI				8			8			8		8

Notes:

All the papers identified in a systematic search of 24 top journals publishing research on cities related to international business are indexed in this table to improve its readability and comply with journal style guidance. Where needed a reference is also included in the references for the paper, when cited in-text

Sources: Nachum and Wymbs (2005), Chen and Yeh (2012); Goerzen et al. (2013), Blevins et al. (2016); Belderbos et al. (2017), McDonald et al. (2018); Belderbos et al. (2020), Zhang et al. (2020); Castellani et al. (2022)

Table 2.

Summary of data collection

	Zephyr investments				Company website and reports 2013				Total FT articles from Search 2013	FT Articles Sampled Evaluate FDI Motive 2013
	Total		Sampled		Total	Sampled
Company	2013	2021	2013	2021	2013	2013	2021	Focal investments studied	2013	2021	2013	2021
Automotive MNEs
Mitsubishi	12	40	4	3	4	3	–	10	8,956	Focused Search	13	15
Toyota	58	117	5	11	17	10	–	26	8,907	Focused Search	32	28
General Motors	24	181	4	2	7	4	–	10	10,170	Focused Search	14	17
Ford	36	135	9	9	13	9	–	27	7,862	Focused Search	30	36
Daimler	65	246	5	11	7	7	–	23	5,792	Focused Search	17	41
Volkswagen	38	125	8	6	21	9	–	23	8,765	Focused Search	31	17
Commercial Banking MNEs
Banco Santander	84	557	19	5	10	10	–	34	1,984	Focused Search	36	24
Crédit Agricole	60	363	18	2	10	10	–	30	1,977	Focused Search	68	6
Société Générale	85	1310	22	3	8	8	–	33	4,901	Focused Search	40	7
Barclays	63	761	24	0	13	13	–	27	3,378	Focused Search	57	1
HSBC Holdings	44	1211	19	3	6	6	–	28	2,190	Focused Search	47	19
Standard Chartered	23	127	17	3	9	9	–	29	717	Focused Search	31	18

Table 3.

Summary of variables used to categorize cities into Hymer’s (1972) tiers for automotive and commercial banking foreign direct investment

	Evaluation criteria		Category frequency and descriptive statistics		Source
Variable	Automotive investments	Commercial banking investments
City overall international connectivity	The GaWC ranking of world cities provides an indication of global or world city membership, the higher tier cities are more likely to be tier 1 cities and the lower tier cities are more likely to be tier 2 cities or tier 3 cities		Alpha++ Alpha+ Alpha Alpha− Beta+ Beta Beta− Gamma+ Gamma Gamma− High Sufficiency Sufficiency No category	4 11 13 18 18 5 10 8 1 5 7 13 105	Globalization and World Cities Research Network (www.lboro.com/gawc)
Capital market present	A capital market indicates a tier 1 or 2 city, in combination with high connectivity a tier 1 city is more likely, with low connectivity a tier 2 city		Present: Not present: Unclear/other:	66 57 95	World Stock Exchanges (www.world-stock-exchanges.net)
Capital of country	On the other hand, a tier 2 city must have at least an airport or seaport. If it is a capital of a country, then it is clearly a tier 2 city. A tier III city will never have an airport and will never be a capital, but it may have a seaport	Will be associated with tier 1 or 2 cities, distinguished by connectivity	Present: Not present: Unclear/other:	53 165 0	World Capitals (www.infoplease.com/ipa/A0855603.html)
University present		The presence of top ranked universities is expected to be associated with tier 1 cities	Present: Not present: Unclear/other:	82 136 0	Shanghai Ranking Top University Ranking (www.shanghairanking.com)
Presence of airport		The more central the city in the global air travel network the more likely it is tier 1 or tier 2	Present: Not present: Unclear/other:	66 58 94	The A-Z Group (https://azfreight.com/)
Presence of seaport		The presence of a seaport is expected to be associated with tier 2 and 3 cities	Present: Not present: Unclear/other:	69 142 7	World Port Source (www.worldportsource.com)
Population (2000–2013) (2014–2021)	If the population is larger than 2 million, then the city is a tier 2 city	Tier 2 cities have a population of at least one and a half million people	Min: Mean: Max:	229 3.7 M 35.4 M	City Population OECD (http://stats.oecd.org) Other (www.citypopulation.de)

Note:

See Supplementary_Online_Materials_Document_1_-_Classification_of_Cities and Supplementary_Online_Materials_Document_2_-_Cities Classification Data

Table 4.

Summary of MNE investment motives by region and city Tier 2000–2021

FDI motives		Automotive MNEs						Commercial banking MNEs
Home region		Asia-Pacific		North America		Europe		Europe
Home country		Japan		USA		Germany		Spain	France		UK
Region	City tier	Mitsubishi	Toyota	GM	Ford	Daimler	Volkswagen	Banco Santander	Crédit Agricole	Société Générale	Barclays	HSBC	Standard chartered
Strategic orientation (2000–2013)		Home	Home	Home	Home	Home – Global	Home	Bi-regional (EU/SA)	Home	Home	Home – Bi-regional (EU/NA)5	Host/Bi-regional (AP/EU)6	Host (AP)7
NA	Tier 1	–	–	–	–	–	–	–	–	NM	SAS	SAS	SAS
	Tier 2	SAS/ES	SAS/ES	–	SAS/ES/MS	SAS/NM	SAS	SAS/MS	–	–	–	SAS/MS	–
	Tier 3	–	RS/ES /SAS	–	ES	SAS/RS	ES/RS	–	–	–	SAS	SAS/ES/MS	SAS
EU	Tier 1	MS	MS/SAS	–	–	SAS/NM	SAS/RS	SAS/ES /MS/NM	SAS/ES /MS	SAS/ES/ NM/MS	SAS	SAS/MS/ES	SAS
	Tier 2	SAS	–	MS/NM	ES	RS/SAS	SAS/MS	SAS/MS	SAS/MS/ ES	SAS/MS	SAS/MS	SAS/ES	SAS
	Tier 3	SAS/RS	RS/MS	SAS/ES	SAS/ES	SAS/RS/ES	SAS/ES/RS	SAS/MS	SAS/ES/ MS	SAS/MS	SAS	SAS/ES	–
AP	Tier 1	–	MS/SAS /RS	SAS	–	SAS/ES	SAS/ES	–	?	–	SAS/NM	SAS	SAS/MS
	Tier 2	ES/MS	ES/MS	–	RS/MS/ES	MS/RS/SAS	ES/MS /SAS	–	–	–	SAS/ES/ MS	SAS/MS	SAS/MS
	Tier 3	–	ES	SAS	ES	–	ES	–	–	–	MS	SAS/MS	SAS/MS
SA	Tier 1	–	MS/SAS	–	–	–	–	SAS	–	SAS	–	–	–
	Tier 2	–	–	–	–	–	–	SAS	–	–	SAS/MS	–	–
	Tier3	–	–	–	–	–	–	SAS	–	–	ES	–	–
AM	Tier 1	–	–	–	–	–	–	–	–	–	–	–	–
	Tier 2	–	–	–	SASMS	–	–	–	SAS/MS	SAS/MS	–	SAS/MS	SAS/MS
	Tier 3	–	–	–	–	–	–	–	–	?	–	–	SAS/MS

Notes:

SAS: Strategic asset seeking; MS: Market seeking; ES: Efficiency seeking; RS: Resource seeking;

See Supplementary_Online_Materials_Document_3_-_Table_S1a_Investments_in_Tier_1_Cities_by_Automotive_MNEs, Supplementary_Online_Materials_Document_4_-_Table_S1b_Investments_in_Tier_1_Cities_by_Commercial_Banking_MNEs, Supplementary_Online_Materials_Document_5_-_Table_S2a_Investments_in_Tier_2_Cities_by_Automotive_MNEs, Supplementary_Online_Materials_Document_6_-_Table_S2b_Investments_in_Tier_2_Cities_by_Commercial_Banking_MNEs, Supplementary_Online_Materials_Document_7_-_Table_S3a_Investments_in_Tier_3_Cities_by_Automotive_MNEs, and Supplementary_Online_Materials_Document_8_-_Table_S3b_Investments_in_Tier_3_Cities_by_Commercial_Banking_MNEs for the underlying analytical tables showing detailed information about individual investments by city tier and region for the automotive and commercial banking MNEs

Supplementary material

The supplementary material for this article can be found online.