How do past mode choices influence subsequent entry?

How do past mode choices influence subsequent entry? A study on the boundary conditions of preferred entry modes of retail firms
Bernhard Swoboda *, Stefan Elsner, Edith Olejnik
University of Trier, Department of Marketing and Retailing, Universitaetsring 15, D-54286 Trier, Germany
 A R T I C L E I N F O

Article history:
Received 15 October 2013
Received in revised form 18 August 2014 Accepted 17 October 2014
Available online 15 November 2014
Keywords:
Entry mode Institutional theory International retailing
Knowledge-based theory
A B S T R A C T
This study analyzes whether the preferred entry modes, i.e., the foreign market entry modes that have been most frequently used in the past, influence a retail firm’s subsequent mode choices. We discuss the limitations of this relationship by highlighting the external and internal factors that determine the effects of preferred modes on later entry decisions. To provide insight into these issues, we refer to institutional- and knowledge-based reasoing and use a data set that includes 309 market entries by the 30 leading retailers between the years 1960 and 2008. The results indicate that preferred entry modes show strong explanatory power with regard to the subsequent choice of full- and shared-control modes in entering new country markets. Although this relationship is diminished by the external institutional environment (e.g., political distance), firm-specific capabilities, e.g., international experience and internationalization speed, reinforce the use of preferred entry modes.
© 2014 Elsevier Ltd. All rights reserved.

  Introduction 

The choice of a foreign market entry mode, i.e., an organiza- tional arrangement that makes the entry of firm resources into a foreign country possible (Root, 1987), is a crucial decision that has been extensively analyzed in the literature (e.g., Brouthers & Hennart, 2007; Morschett, Schramm-Klein, & Swoboda, 2010). Most often, it has been assumed that the entry mode choice follows the deliberate efforts of firms to enhance their competiveness, efficiency, and control over critical resources. However, Anderson and Gatignon (1986) state that some firms may behave in this manner, whereas others simply rely on past decisions. In general, past decisions are important to firms’ subsequent behavior. Benito, Petersen, and Welch (2009) argue that past entry decisions may determine subsequent mode choices in new country markets. Hence, this study addresses the effects of preferred entry modes, i.e., modes that have been predominantly used in the past, on later mode choices. We particularly emphasize the contextual limita- tions, or the boundary conditions, of this relationship.
The contextual factors that influence the extent to which firms canemploytheirpreferredentrymodeareparticularlyimportantfor retail firms. Many well-known retailers have aggressively and
* Corresponding author. Tel.: +49 651 201 3050; fax: +49 651 201 4165.
E-mail address: b.swoboda@uni-trier.de (B. Swoboda).
extensively internationalized, especially during the last two decades (Huang & Sternquist, 2007). With every entry into a new country, grocery retailers must commit tremendous resources to build a store network and establish a local presence (Goldman, 2001). Hence, theymostoftenusefull-controlmodes(wholly owned subsidiaries, acquisitions) rather than shared-control modes (joint ventures, franchising, licensing) when entering foreign countries (Gielens & Dekimpe, 2001; Gamble, 2010). They furthermore tend to have preferences for certain entry modes that can be easily reproduced. The Metro Group, the fourth-largest retailer in the world, claims their preference for full-control modes on its homepage: ‘‘Our corporate principles for international expansion areemergingcountries, firstmoveradvantages, andorganic growth’’ (METRO AG, 2013). It can be assumed that many retailers have a preferred entry mode, but we argue that the influence of this preferred mode on later mode choices is bounded by contextual factors. For example, the French retailer Carrefour prefers full- control modes but will adapt its entry method due to local regulations (Bell, Lal, & Salmon, 2004; the same is true of British Tesco in Thailand, Samiee, Yip, & Luk, 2004). Other retailers are forced to use full-control modes due to local governments’ intent of increasingforeign directinvestments(Grewal & Dharwadkar, 2002). Therefore, we aim to analyze whether preferred entry modes predict subsequent mode choices. Moreover, we examine how external environmental pressures and internal capabilities moderate the relationship between preferred entry modes and

http://dx.doi.org/10.1016/j.ibusrev.2014.10.008 0969-5931/© 2014 Elsevier Ltd. All rights reserved.

subsequent mode choices. This research offers three contributions to the existing literature. First, we extend knowledge on the mode choices of retail firms, which is important because retailers differ from manufacturing firms in many ways but have rarely been analyzed in previous entry mode research (e.g., Sanchez-Peinado, Pla-Barber, & He´ bert, 2007; Zhao, Luo, & Suh, 2004). Second, answering Meyer, Estrin, Bhaumik, and Peng’s (2009) call for research in this area, we contextualize the relationship between preferred entry modes and subsequent mode choice by investi- gating the moderating effects of external institutional environ- ments and internal capabilities. Third, following the call for more research on retailing issues that considers institutional theory (Huang & Sternquist, 2007) we combine institutional- and knowledge-based reasoning to analyze the relationships. For managers, this study offers insights into the effects and limits of preferred entry modes. Many managers that have used the expression ‘‘we have always done it this way’’ at some point may better reflect the effects and boundaries of these managerial decisions.  Preferred entry modes and subsequent modechoices
The modes that have been most frequently used in the past, i.e., preferred entry modes, may determine subsequent entry mode choices (Benito et al., 2009). Organizations persist in continuing the same type of activity over time because search rules change slowly and remain conditioned by previously attempted solutions (Cyert & March, 1963; Miller & Friesen, 1980). Such decisions are not independent of one another; they are inseparable from the results of previous decision making (Forest & Mehier, 2001). Hence, organizations are shaped by historical factors that limit the range of options available to their decision makers. Furthermore, the theory of organizational inertia (Romanelli & Tushman, 1986; Tallman & Shenkar, 1994) posits that successful firms institution- alize established activity patterns so that the likelihood of change becomes remote. Empirical support for the effects of past decisions on subsequent mode choices is provided only for initial entry modes chosen in sequential entries into one specific county (Chang & Rosenzweig, 2001), for the effect of the number of prior years of full ownership on later initial entries (Padmanabhan & Cho, 1999), and for the effect of preferred modes in the same host country (Yiu & Makino, 2002) as well as across Western countries (Lu, 2002). Little is known about this behavior for service firms in general or for retailers specifically. Moreover, the boundary conditions of preferred entry modes can extend our knowledge of the limitations of choosing the preferred mode repeatedly. Therefore, we will discuss the theoretical roots of preferred entry modes and then elaborate on the specifics for retail firms.
Knowledge-based and institutionaltheory
Mode learning and mode habitualization are common approaches that are used to explain how preferred modes develop and how they affect subsequent mode choice (Benito et al., 2009). The term mode learning can be traced back to the knowledge- based theory of firms. This perspective recognizes that knowledge is the most strategically important resource in a firm (Kogut & Zander, 1993). In entry mode research, this perspective is relevant because firms need to consider the compatibility between the firm’s existing knowledge and the knowledge that the firm will require to be successful in the new market. In this context, both firm-specific and context-specific knowledge (Madhok, 1997; Zander & Kogut, 1995) explain the effect of prior decisions on subsequent entry modes. Chang and Rosenzweig (2001) state that
firms learn from prior mode experience and reduce risk by using established modes. Building on Chang (1995), the authors demonstrate that business lines that initially entered a country through a given mode of entry will, all other things being equal, choose the same entry mode again. Padmanabhan and Cho (1999) emphasize this behavior by referring to decision-specific experi- ence. Positive mode experiences lead to growing knowledge and confidence in the continued use of the same mode, whereas negative mode experiences lead to a bias against using the same mode (e.g., Welch, Benito, & Petersen, 2007). Knowledge-based reasoning is valuable not only because it can reveal the effects of past modes on subsequent mode choices but also because it can be used to determine the limitations of this effect.
The term mode habitualization is based on institutional theory, which considers organizations to be social actors who are embedded in internal and external environments that consist of structures, standards, and practices that have been established in the past by other social actors, organizations, and institutions (DiMaggio & Powell, 1983; Meyer & Rowan, 1977). These institutions impose isomorphic pressures on organizations to assimilate. The organizations must respond to these pressures to gain legitimacy. During their international expansion, firms mainly face two important pressures: they need to achieve conformity with the environment of the host country and with the organizational practices within the MNC. When entering a foreign market, the external pressure to conform to the environment is evident. There are external regulative pressures such as political- legal requirements as well as external normative pressures such as informal norms, values, and beliefs. Although these factors may lead to the adaptation of established decisions, the internal pressure encourages consistency. Internal cognitive pressures include firms’ habitual behaviors and repeated actions that are developed over the course of history (Berger & Luckmann, 1967). These pressures yield organizational processes that result in institutionalized decisions characterized by stability and resis- tance to change. Hence, habitualization refers to processes in which actions that are repeated become taken-for-granted patterns or routines that can be easily reproduced (Berger & Luckmann, 1967). The two primary mechanisms that facilitate habitualization are imprinting and bypassing (Grewal & Dhar- wadkar, 2002). Both mechanisms lead to ‘‘programmed actions’’ (Berger & Luckmann, 1967, p. 75) or ‘‘common responses to similar situations’’ (Mead, 1934, p. 263). ‘‘Imprinting refers to the maintenance of structures and processes that were codified in the early years of the organization’s existence and have become sacrosanct because of reification and symbolism. Bypassing captures the process by which institutional expectations are defined in the larger cultural context, which reduces the need for well-articulated structures and processes.’’ (Grewal & Dharwadkar, 2002, p. 88). According to the literature, preferred entry modes are connected with strategic routines, organizational norms, or ceremonial artifacts (Meyer & Rowan, 1977). Lu (2002) also refers to the imprinting concept (Stinchcombe, 1965) and concludes that firms use a preferred mode rather than searching for alternative entry modes. Yiu and Makino (2002) refer to historical norms, arguing that if a firm attains legitimacy in host countries using a particular mode, the MNC will incorporate this successful experi- ence into its cognitive structure. It seems that international retailers gain cognitive legitimacy through the imprinting of repeated mode choices that become strategic routines. Hence, habitualization supersedes the conscious awareness of decision makers and the entire firm, particularly in the long run (Grewal & Dharwadkar, 2002). In practice, the effects are described as follows: ‘‘. . . managers [. . .] justify actions with the claim that ‘we have always done it this way’, ‘everybody does it this way’ or ‘that’s just the way things are done around here’’’ (Oliver, 1997, p. 699–700).

Thus, such mode choices are closely connected to the managerial and organizational levels, as Porter (1990, p. 580–581) notes: ‘‘Past approaches become institutionalized in procedures and manage- ment school and personnel are trained in one mode of behavior.’’ In sum, the institutional duality can explain the effects of preferred modes on subsequent mode choice by analyzing the effects of internal and external pressures on this relationship.
The retailcontext
International retailers differ from manufacturers — not only regarding management, marketing, and financial issues (Dawson, 1994). We will first address the retail-specific entry mode choice and then consider contextual limitations of preferred modes. We acknowledge imprinting as the main mechanism that leads to mode habitualization by retailers, although the mechanisms are not the subject of this study.
Retailers, grocery retailers in particular are considered multi- domestic as they adapt to local consumer needs and build local supply chains (Bianchi & Ostale, 2006; Gamble, 2010), while non- food retailers, fashion retailers in particular are considered global, replicating retail format elements unchanged abroad (Jonsson & Foss, 2011; Swoboda, Elsner, & Morschett, 2012). Retailers cannot simply export their products. They need to enter foreign countries using full- or shared-control modes. Many retailers prefer full- control modes because they want to retain full control over the foreign activities (Gamble, 2010). Others often use partners in their international expansion to benefit from their local knowledge (Huang & Sternquist, 2007). Whichever mode they choose enormous investments into the infrastructure force grocery retailers to pursue economies of scale and to obtain the required return on investment quickly. Therefore, these firms can be expected to select their markets and entry modes carefully. One would expect that every entry mode choice is a new deliberate decision. However, there are retailers who seem to have a routine for international expansion, i.e., a preferred entry mode. For example, Metro prefers organic growth, as does Carrefour (Bell et al., 2004). Also, successful retailers transfer their preferred retail formats unchanged from the home country to a host country (Gielens & Dekimpe, 2001). In contrast to manufacturers that can usually choose between several products for international sales, retailers usually have one or a few formats (being their products). It can be observed that they repeatedly use the same entry mode, which may be related to the preferred format and which has been habitualized over the course of the firm’s history. Thus, it is theoretically intriguing to examine whether retailers deliberately choose each entry mode every time or whether they rely on preferred modes.
Previous research indicates the existence of preferred entry modes and, in some instances, potential antecedents of this habitualized behavior. For example, Huang and Sternquist (2007) report that The Body Shop had a preference for franchising from the beginning of its international activities. The retail format (i.e., cosmetic stores) and early positive experiences with the entry mode may be antecedents of this behavior. Negative mode experiences and a bias against the previously used mode are reported for Marks & Spencer (Alexander & Quinn, 2002) and for fashion retailers (Doherty, 2000).
Imprinting would suggest that retailers such as The Body Shop have evaluated entry mode options and decided on the processes and structures of the international expansion deliber- ately in the beginning. Since The Body Shop focuses on one retail format (i.e., cosmetic stores) it is likely that the same entry mode will be used in multiple entries because a fit with the retail format has been established. Diversified retailers such as Marks & Spencer or Metro have several retail formats to choose from.

They may have based their initial mode decision on an elaborated decision-making process and a set of variables such as the format, previous mode experiences, home country location etc. However, once a mode has been evaluated and selected it will be used repeatedly and becomes a routine, a taken-for- granted pattern. The establishment of a preferred mode has advantages for retailers because it reduces complexity and disburdens management’s limited information-processing ca- pacities. Preferred entry modes can simplify task complexity and therefore serve as facilitator for the retailer’s international expansion. Instead of deciding every aspect of an international entry over and over again (e.g., country, timing, format, entry mode) retailers can focus on decisions regarding format elements that are visible and important to local customers such as assortment and price (Huang & Sternquist, 2007; Swoboda & Elsner, 2013). Applying a routine simplifies the decision making process and facilitates quick expansion to obtain the required return on investment. For example, the discount giant ALDI has entered countries like Australia or US using full-control modes to be able to open up to 100 stores per year in those countries.
Hence, we analyze whether retailers have a habitualized mode but in particular, we want to research whether retailers deviate from that preferred mode under certain circumstances. Research- ing the boundary conditions of the link between preferred modes and subsequent mode choices is of paramount importance for retailers. In addition to high resource commitments, retailers have two specific characteristics. First, they are strongly influenced by external institutional pressures and react sensitively to foreign market challenges because their business depends on the local circumstances and consumption culture (Huang & Sternquist, 2007; Evans & Mavondo, 2002). Grocery retailers in particular must take into account various regulations (e.g., regarding store locations, opening hours) and norms (e.g., direct contact, high transaction frequencies with local customers). Difficulty charac- terizing the environment ex-ante leads to external uncertainty and insecure returns on investment. Because of these barriers, retailers may refrain from entering a market or may change their preferred mode. Hence, external environmental factors may limit the influence of preferred entry modes on subsequent decisions. Second, internationalization knowledge is important for most retailers. Knowledge-generating internal capabilities, such as international experience and internationalization speed, may reinforce mode habitualization because retailers develop patterns and routines for their entry mode decisions to reduce the complexity of operations abroad and facilitate international expansion.
The following chapter first addresses the rationale for the relationship between preferred entry modes and subsequent entry mode choice (see Fig. 1). Then, a rationale is developed for the influence of external institutional pressures and internal interna- tionalization knowledge on this relationship.
1. Hypotheses
Preferred modes and subsequent modechoices
The following rationale refers to internal institutions and acknowledges that different routes may lead firms to habitualize modes and preserve related strategic routines. One way in which internal institutions influence entry mode choice is through organizational imprinting. Imprinting is a process of institution- alization that creates reality for a firm within its own internal environments (Berger & Luckmann, 1967; Zucker, 1977). That is, once an entry mode has been chosen and implemented, the likelihood of considering and using alternatives in future decisions decreases. As the frequency of adopting a particular

Fig. 1. Research model. mode grows over time, it becomes a ‘‘taken for granted’’ (Zucker, 1977) approach that is difficult or unlikely to change (Mezias, 1990).

As mentioned, conceptual and case study research highlights the existence of preferred modes in retail firms and discusses various mechanisms, such as mode learning, business formats, and strategic norms. Previous studies refer to retailers’ persistence in continuing the same type of activity over time because search rules adapt slowly and remain conditioned by previously attempted solutions or strategic norms/routines. Thus, imprinting seems to be relevant to retailers’ entry mode choices (Huang & Sternquist, 2007). Therefore, this study posits that the entry mode a retailer choses in its early market entries becomes habitualized via imprinting which results in a preference for the application of the same entry mode in subsequent market entries.
Hypothesis 1. The greater the frequency of adoption of an entry mode in an international retailer’s earlier entries, the greater the retailer’s propensity to use that preferred entry mode in its subse- quent entries in other country markets. 
Political-legal requirements are external institutional pressures and regulative factors (Berry, Guille´ n, & Zhou, 2010) that have a tremendous impact on the retailing industry. For example, European governments have formulated strict regulations on land planning (e.g., PPG6 in England), pricing (e.g., Loi Galland in France), store opening hours (e.g., Ladenschlussgesetz in Germany), and store size requirements (e.g., Loi Raffarin in France) that have stimulated many European retailers to expand into other countries and slowed foreign retailers’ expansion (e.g., Wal-Mart in Europe, Huang & Sternquist, 2007). Another group of factors related to the external institutional environment includes norma- tive pressures that emphasize moral beliefs and internalized obligations as the basis of social meaning and the social order (Scott & Christensen, 1995). These pressures are directly relevant to retailers because of the direct customer contact (Evans & Mavondo, 2002). As international retailers mostly transfer formats from their home countries and evaluate external environments prior to market entry, the political and cultural distances between their home and host markets are particularly relevant (Huang & Sternquist, 2007; Kostova & Zaheer, 1999). Retailers may maintain an entry mode that has been frequently used in the past until regulative and normative distances force them to deviate from such an entry mode.
Political distance is defined as the extent of the differences between the governmental and political institutions in the home and host countries (Salomon & Wu, 2012). Political distance refers to the nature of the regulative environment to which firms must conform by adapting to the local expectations and the country’s governance structures (Meyer, 2001). Regulations are a double- edged sword. Although the governance infrastructure reduces firms’ uncertainty, regulations can also act as barriers for firms. Thus, substantial political distance makes it more difficult for firms to use strategic routines abroad and generates a greater probability that a preferred entry mode will need to be changed for the firm to assimilate and gain legitimacy in the host country.
Studies show that regulations largely determine retail firms’ mode choice (Grewal & Dharwadkar, 2002; Huang & Sternquist, 2007). Political distance represents a particularly crucial factor in the internationalization of retail firms because retailers evaluate the regulations of a country before entry and may perceive these regulations as a barrier to gaining legitimacy while using the preferred entry mode. If the regulations in a country impede the use of the preferred entry mode, the retailer may decide not to enter the country and may focus on other countries. Or, if the country is highly attractive, the retailer may decide to adapt its mode of entry. Consequently, entry into a country with significant political distance increases the likelihood that retailers will deviate from their preferred entry modes. This reasoning is supported by real-life examples such as the French retail giant Carrefour’s transition to shared-control modes due to local regulations (Bell et al., 2004), as did British Tesco in Thailand (Samiee et al., 2004).
Hypothesis 2. The effect international retailers’ preferred entry modes have on subsequent mode choices will be weaker as the political distance between the host and home countries increases.
Cultural distance is the extent to which socio-cultural factors differ between the home and host countries (e.g., Evans & Mavondo, 2002; House, Hanges, Javidan, Dorfman, & Gupta, 2007). Cultural distance refers to the normative institutional environment to which firms must conform. Although it seems intuitive that firms would use shared-control modes to enter countries with significant cultural distance to reduce risk, empirical studies are ambiguous regarding the effect of cultural distance on mode choice (e.g., Morschett et al., 2010; Tihanyi, Griffith, & Russell, 2005). Cultural distance is found a relevant moderator of the propensity to use prior entry modes for second

entries into the same country (Zhao et al., 2004). It is plausible that substantial cultural distance makes it difficult for firms to transfer strategic routines abroad and that cultural pressures force firms to deviate from their preferred modes. In contrast, limited cultural distance increases the probability that a firm will choose its preferred entry mode because this mode is legitimate from a cultural perspective (Kostova, 1999).
Retailers with a preferred entry mode may perceive strong pressure to adapt their mode of operation due to the need to serve customers who differ culturally from the customers at home. Accordingly, retailers will prefer not to enter such countries or to mitigate the pressure and to achieve a ‘‘legitimate fit’’ in the host country by adapting their entry modes (Huang & Sternquist, 2007). In sum, a departure from the preferred mode can help retailers to overcome cultural distance and normative impedi- ments and to increase legitimacy. For example, Wal-Mart departed from full-control modes in Argentina (Huang & Sternquist, 2007), and Costco departed from shared-control modes in Japan (Jetro, 2003).
Hypothesis 3. The effect international retailers’ preferred entry modes have on subsequent mode choices will be weaker as the cultural distance between the host and home countries increases.
Moderating effects of internationalizationknowledge
Internationalization knowledge is an important capability (e.g., Madhok, 1997; Newbert, 2007) that may differ across retailers in terms of its nature and scope. Thus, we distinguish between the overall international experience, which begins with the first market entry and refers to the nature of knowledge accumulation (Erramilli, 1991), and the internationalization speed, which refers to the scope of knowledge accumulation (Oviatt & McDougall, 2005; Wagner, 2004). Both the nature and the scope of internationalization knowledge generate important capabilities for retailers. Knowledge helps firms to achieve economies of scale, which are needed to attain a certain critical mass early and to ensure a proper return on investment due to the capital-intensive nature of foreign expansion. When retailers possess international- ization knowledge, the use of preferred modes may be reinforced in subsequent entries because the firms learn from prior market entries and reduce risk by using established entry modes (e.g., Chang & Rosenzweig, 2001).
International experience is the degree to which a firm has been involved in cross-border business activities (e.g., Chang & Rosenzweig, 2001). It poses a firm-specific capability that a firm develops by learning and accumulating internationalization knowledge (Colombo, 1995). A high degree of international experience reinforces the use of a preferred entry mode in subsequent entry decisions. One underlying rationale is that inexperienced firms are uncertain about foreign markets because they lack the experience to develop a particular ‘formula’ for making decisions in new markets (Lu, 2002). One consequence of this uncertainty is that a newly internationalizing firm has no established method of analyzing relevant information, nor has it implemented a heuristic for entry decisions (e.g., Henisz & Delios, 2001). But, with increasing international experience, firms may reinforce their preferred modes. This reasoning suggests that the imprinting effect of prior entry modes becomes more prominent as firms accumulate experience. Lu (2002) found support for this rationale, noting that experience moderates the relationship between manufacturers’ frequency of past entry mode use and their current entry decisions.
In retailing, a similar reinforcing mechanism for international experience is assumed to exist. Retailers are particularly sensitive to knowledge about foreign countries because they have a relatively limited internationalization history, which makes international experience an important capability. International experience enhances retailers’ understanding of and competence in managing the idiosyncrasies of host country environments, and it improves the assessment of the risks and returns associated with the use of preferred modes in subsequent decisions. Highly experienced retailers are more likely to implement a preferred entry mode in subsequent entries after several country openings, exits, and re-entries than they were in earlier internationalization phases in which they lacked experience (Alexander & Quinn, 2002; Doherty, 2000). Thus, it is likely that a retailer’s usage of preferred entry modes will be reinforced over time due to increasing international experience.
Hypothesis 4. The effect international retailers’ preferred entry modes have on subsequent mode choices will be greater when the retailer has more international experience.
Internationalization speed is defined as the number of foreign markets a firm enters within a specific period of time, i.e., how rapidly the firm accumulates foreign market entries (Vermeulen & Barkema, 2002). The ability to internationalize rapidly is the result of learning processes and reinforces the use of a preferred entry mode in subsequent entry decisions. A rapid internationalization pace creates difficulties for firms when they need to adapt to a new business model and hampers the use of new organizational routines (e.g., Hannan & Freeman, 1984). However, a high internationalization speed allows firms to achieve economies of scale by using a particular business concept across countries. Hence, firms must codify their know-how into routines (Morgan, Zou, Vorhies, & Katsikeas, 2003) and act upon them. This rationale suggests that increasing internationalization speed reinforces the use of preferred entry modes. In contrast, internationalizing slowly may increase a firm’s ability to adapt to local markets because of the time available to develop new business model configurations and new organizational routines for local markets.
Retailers, especially grocery retailers, must make substantial investments in store networks and supply chain processes (Swoboda & Elsner, 2013) when entering into a new country. Hence, the ability to enter numerous countries within a short period of time is a valuable capability and a success factor for retailers (e.g., Chan, Finnegan, & Sternquist, 2011; Gielens & Dekimpe, 2007). Consequently, retailers can leverage their accumulated knowledge by transferring a standardized retail format from home, which will allow them to achieve economies of scale and more easily coordinate international operations, among other benefits. Support for this rationale is provided by the two largest fashion retailers. The Spanish company Zara and the Swedish retailer H&M have entered more than 30 countries in the last decade using standardized retail formats and entry modes (Tokatli, 2007).
Hypothesis 5. The effect international retailers’ preferred entry modes have on subsequent mode choices will be greater for firms that have greater internationalization speeds.
1. Empirical studyThe hypotheses were tested using secondary data (e.g., Brouthers & Hennart, 2007). To control for sector influence, we

 Table 1
The world leading grocery retailers (Planet Retail, 2008). Excluded with foreign sales less than 1% such as Kroger (Rank 7), Target (8), Walgreens (11), CVS (14), Supervalu (26), Wesfarmers (27), Sainsbury (28), Migros (30) or only one foreign market such as Sears (16), Safeway (18), Woolworths (24), specific history of expansion Seven&I (10). ued a sample of grocery retailers. The grocery sector is by far the largest retail sector in nearly all economies (Planet Retail, 2011). It attracts different consumer segments and offers a broader range of locally adapted products than specialized retailers, particularly fashion retailers which act more standardized and centralized (Swoboda et al., 2012). Thus studying external boundary condi- tions of the link between preferred modes and mode choice is particular interesting in the grocery context. We collected data on each foreign market entry by the 30 leading retailers in the world. We did not consider smaller grocery retailers because they usually are less or not at all internationalized. Of the 30 leading retailers, those whose foreign sales made up less than one percent of their total sales (e.g., Kroger, Target, and Walgreens) and those that have entered only one foreign market (e.g., Sears, Safeway, and Woolworth) were excluded. This procedure yielded 309 entries by 18 retailers in 82 different countries between the years 1960 and 2008 (see Table 1). We define foreign market entry as an instance in which a retailer begins to operate a store-based business in a country other than its home country for the first time. The variables used in this study relate to the time of entry into a market because Brouthers and Hennart (2007) state that the most effective means of investigating entry mode decisions is to use the time at which decisions are made. Firm data, such as the number of employees, were obtained from annual reports. The years of each market entry and the corresponding entry mode were determined from the Planet Retail database, which is released by the leading retail intelligence provider and includes data from 211 countries (Planet Retail, 2011). Country data were collected from publicly available databases, including the World Bank database, the Globe study, the Macro Data Guide, Nationmaster, and the United

Nations World Urbanization Prospects.
Measurement ofvariables
In this study, entry mode choice was measured as a binary variable with full-control modes coded as 1 (including wholly- owned subsidiaries and acquisitions) and shared-control modes coded as 0 (including joint ventures, franchising/licensing, and minority stakes). This procedure is consistent with studies that acknowledge control as a crucial factor (e.g., Canabal & White, 2008) and takes account of the importance of control in retail internationalization. As will be shown in the results section, the
results are stable even if acquisitions and franchising/licensing are excluded from the estimations, as it could be argued that acquisitions and franchising may represent distinct modes that do not fit neatly on a high control – low control continuum. The distribution of the observed entry modes Exhibits 213 full- and 96 shared-control modes, which is in accordance with previous studies on service firms (e.g., Sanchez-Peinado et al., 2007).
The preferred entry mode was measured by the ratio of the
number of full- vs. shared-controlled modes to the total number of entry modes used for the same company in the host countries at the time of entry. This methodology was suggested by Huang and Sternquist (2007) and applied by Lu (2002) and takes into account the fact that full-control modes are the modes that are most frequently used by grocery retail firms (e.g., Gielens & Dekimpe, 2001).
Political distance was measured using the political constraint
index (POLCONV), which is one of the most comprehensive measures of the political environment (Henisz, 2000; Jime´ nez, 2010). To calculate the annual index value for each country from the year of entry, independent institutional actors with veto power were considered (Garcı´a-Canal & Guille´ n, 2008). The index ranges from 0 for governments with no veto power by other institutions (such as dictatorships) to 1 for governments in which several institutions have veto power (such as democracies). To measure the political distance for each market entry, the difference between the political constraint values of the home and host countries was used (Berry et al., 2010; Henisz, 2000).
The cultural distance between the home and host countries was
measured according to the Euclidian distance between the dimensions of the cultural practices (‘as is’ data) from the Globe study (House et al., 2007; Kogut & Singh, 1988). This study employs practical indices because these are particularly suitable for analyzing the influence of culture in an organizational context. Missing data was replaced, e.g., by data of neighboring countries. Moreover, it is important for retailers to measure the distance between their home and host countries because retailers conduct their market entry decisions primarily from their home countries (Ha˚ kanson & Ambos, 2010).
International experience was measured using the number of
years between the first foreign market entry and the observed market entry (e.g., Brouthers, 2002). Because this study investi- gates entry mode choice from the early beginnings, this measure of international experience seems appropriate (Erramilli, 1991) because it reflects how international business operations can be used, developed, and transferred from one market to another.

Internationalization speed was considered as the average
number of market entries per year (according to Vermeulen & Barkema, 2002; Chan et al., 2011 in the retail context). Therefore, the number of foreign markets in which a retailer operates was divided by the number of years that the retailer has operated abroad. Thus, it indicates how rapidly foreign market entries are accumulated in a certain period of time, which differs from the scope of international experience (Erramilli, 1991). Higher values for this measure indicate that on average a greater number of markets have been entered per year.
This study controls for the effects of several variables. Firm size
was controlled for because it is considered a critical resource, particularly in retail entry mode choices. Firm size was measured using the total number of employees expressed in 10,000 persons (e.g., Brouthers, Brouthers, & Werner, 2008). Market size was controlled for using the host country’s GDP expressed in 10bn USD; it is an important indicator of market attractiveness (Henisz & Delios, 2001). Following Johnson and Tellis (2008), openness to FDI (i.e., the degree to which host countries appreciate FDI) was measured by computing the ratio of FDI to GDP. The host country’s degree of industrial development was measured using the percentage of urbanization. To reduce possible bias resulting from unobserved heterogeneity among the 18 retailers that were investigated, firm dummy variables were included to control for higher-order effects. Three firm dummy variables that were significant but did not change the results were not included in the results section due to the complexity of the model. The same procedure was used for the four countries that had more than ten observed market entries. Furthermore, because the dummy approach has its disadvantages, the generalized estimating equations (GEE) approach was employed using robust standard errors to verify the results for biases caused by possible firm and country effects (Liang & Zeger, 1986). 
Before the analysis was conducted, a separate outlier diagnostic was carried out to prevent biases in the analysis that could result from extreme values. Eight observations were excluded because they exceeded the cut-off points of 3.0 for the standardized residuals and 1.0 for the Cook’s distance (Cohen, Cohen, West, & Aiken, 2003). Furthermore, all of the variables were z-standardized prior to the analysis to avoid multicollinearity in models with interaction effects (Aiken & West, 1991). As illustrated in Table 2, all of the correlations are below the threshold of 0.7, and the variance inflation factors (VIFs) are lower than the recommended threshold of ten (Diamantopoulos & Winklhofer, 2001). The results are stable even if the firm and country dummies are included. We additionally examine whether the errors are independent and thus test for autocorrelation using a Durbin-Watson test. The results across the tested models do not exceed or fall below the critical thresholds (between 1.93 and 2.02). Hence, multicollinearity and autocorrelation do not seem to be serious problems in this study. We account for a possible self-selection bias caused by exogenous variables of a preferred entry mode. Following Shaver (1998) and Heckmann (1979) we generate a correction factor by calculating a probit model, with the preferred mode as dependent variable and the above firm-/country-specific independent vari- ables (see Table 3). As the preferred mode is a function of previously entered countries (not of the current entry) we use averages for the external variables of all the previous entries of a retailer. Conform to our theoretic rationale retailers’ diversification was controlled for because undiversified retailers may rely on the same format for international expansion (as their sole capability) and on a particular entry mode that fits this format. This variable (0 = diversified, 1 = not diversified) was obtained from firms’ websites and from the Planet Retail database. Similar sources provide data for the competitive situation of previously entered host countries, i.e., the proportion of modern grocery retailing sales as a share of total grocery spending (scale 1 = 0–20% to 5 = 81–

Table 2
Descriptive statistics, test of multicollinearity and correlation matrix. A binary logistic regression analysis, which is the method most often used for investigating the choice of market entry modes, was performed (Canabal & White, 2008). An additional probit analysis, applied for robustness purposes also based on the generalized estimating equations approach (Liang & Zeger, 1986), almost entirely supports the previous logit analysis

100%; e.g., Goldman, Ramaswami, & Krider, 2000). Based on this model we generated the inverse Mills Ratios for full and shared entry mode (following Shaver’s specification, 1998, p. 581) and included the ratio as a correction factor into all further models for hypothesis testing. Because the impact of the correction factor is insignificant in all models we carefully conclude, that endogeneity of the preferred entry mode does not cause a serious bias to our hypothesis tests.
Table 4 illustrates the results of models 1 to 6, which incorporate the interaction terms in a stepwise manner to demonstrate the stability of the effects. The overall model fit of the final model 6 is reasonably high with a Nagelkerke’s R2 of 0.661. Thus, the data demonstrate a good fit with the proposed conceptual framework. With regard to the asymmetric size of full- and shared-control modes, the classification rate of model 6 is reasonably higher than the proportional chance criterion (PCC) confirming a good predictive power. Model 2 is compared with Model 6 to demonstrate the increase in the explanatory power caused by the moderating effects. The test to compare different degrees of Nagelkerke’s R2 (Aiken &
West, 1991) shows that the Nagelkerke’s R2 values for Model 2 and
for Model 6 differ significantly at p < 0.01 with an f-value of 3.684. In other words, the explanatory power is substantially
enhanced by the moderating effects. Furthermore, rival models, including a model without the acquisition mode and a model without franchising/licensing-type agreements, were estimated.

Model 1 Model 2 Model 3 Model 4 Model 5 Model 6 Beta Exp(B) Beta Exp(B) Beta Exp(B) Beta Exp(B) Beta Exp(B) Beta Exp(B) Independent variable Preferred entry mode 2.664*** (14.357) 2.948*** (19.070) 3.084*** (21.840) 3.090*** (21.969) 3.170*** (23.808) 3.451*** (31.524) Correction factor for —0.776y (0.460) —0.696ns (0.499) —0.720ns (0.487) —0.759ns (0.468) —0.650ns (0.522) —0.555ns (0.576) self-selection Moderator variables Political distance —0.022ns (0.978) —0.134ns (0.874) —0.141ns (0.869) —0.123ns (0.885) —0.080ns (0.923) Cultural distance 0.808** (2.244) 0.874** (2.369) 0.911** (2.486) 0.923** (2.518) 0.929** (2.533) International experience —0.669** (0.512) —0.690** (0.502) —0.697** (0.498) —0.810** (0.445) —0.740** (0.477) Internationalization speed —0.287ns (0.751) —0.302ns (0.739) —0.298ns (0.743) —0.266ns (0.767) —0.087ns (0.916) Interaction terms Preferred entry mode —0.425* (0.653) —0.428* (0.652) —0.404y (0.668) —0.476* (0.622) × Political distance Preferred entry mode 0.135ns (1.145) —0.177ns (0.838) —0.278ns (0.757) × Cultural distance Preferred entry mode 0.600y (1.823) 0.969* (2.634) × International experience

Preferred entry mode
× Internationalization speed 0.915** (2.497)

Controls Firm size	0.328*	(1.388)	0.487**	(1.628)	0.470**	(1.600)	0.456*	(1.577)	0.472**	(1.603)	0.632**	(1.881)
Market size	0.798*	(2.220)	0.862*	(2.368)	0.910*	(2.485)	0.899*	(2.457)	0.887*	(2.429)	0.964*	(2.622)
Market growth	0.120ns	(1.127)	0.141ns	(1.152)	0.115ns	(1.122)	0.110ns	(1.116)	0.116ns	(1.123)	0.089ns	(1.093)
Openness toward FDI	0.383ns	(1.467)	0.496ns	(1.643)	0.496ns	(1.642)	0.490ns	(1.633)	0.540ns	(1.717)	0.531ns	(1.700)
Industrial development	—0.096ns	(.909)	—0.068ns	(0.934)	—0.033ns	(0.968)	—0.034ns	(0.967)	—0.024ns	(0.977)	0.030ns	(1.030)
Constant term	1.250***	(3.489)	1.413***	(1.109)	1.433***	(4.191)	1.447***	(4.251)	1.394***	(4.029)	1.397***	(4.045)
Model indices
N	301	301	301	301	301	301
Model chi-square	162.5***	177.9***	181.6***	181.7***	183.918***	190.498***
R2 (Cox & Snell)	0.417	0.446	0.453	0.453	0.457	0.469
R2 (Nagelkerke’s)	0.588	0.629	0.638	0.639	0.644	0.661
Correct classification	83.1%	84.1%	85.7%	85.4%	86.3%	86.1%

y p < 0.10.
* p < 0.05.
** p < 0.01.
*** p < 0.001.
ns = not significant.

 Nagelkerke’s R2 increased to 0.715 and 0.671, respectively, and the main effect of a preferred entry mode on the choice of subsequent entry modes and the moderating effect of all constructs remained significant. Thus, we acknowledge the stability of the present solution.
Hypothesis 1 suggests that the more often an entry mode has been used in the past, the higher the propensity to use that same entry mode in subsequent market entries. The positive coefficient is highly significant (Exp(B) = 31.524; p < 0.001). Thus, entry mode choices are particularly dependent on the preferred entry modes. In Hypothesis 2, it is assumed that the aforementioned relationship will be negatively influenced by increasing political distance between the host and home countries. The negative coefficient (Exp(B) = 0.622; p < 0.05) suggests that retailers adapt
their preferred entry modes in countries with a significant political
distance. The moderating effects are displayed in Fig. 2 and shown in typically s-shaped curves. Each of the following curves shows the estimated relationship between the preferred entry mode and the probability to choose a full-control mode for different levels of the moderator. For the case of politically distant countries, the probability of choosing full-control modes decreases when the preferred entry mode represents a full-control mode, and the probability of choosing full-control modes increases when the preferred mode represents a shared-control mode. Thus, it can be concluded that for large political distances, retailers depart from their preferred entry mode.
Hypothesis 3 proposes that a retailer tends to depart from the preferred mode when the cultural distance between the host and home countries is high. While the negative beta coefficient in Table 4 is in the expected direction, it is not significant. Thus, Hypothesis 3 is rejected. To increase the robustness of the results alternative cultural distance measures were tested. First, tests of the models with the ‘as should be’ data form of the GLOBE study do not change the reported results. Second, tests of the models with the cultural distance measures according to Berry et al. (2010) still show insignificant results for the interaction term and also for the direct effect of cultural distance on mode choice. Thus, it cannot be confirmed that the effect of a retailer’s preferred entry mode is affected by cultural distance (Huang & Sternquist, 2007). One reason for this finding may be that cultural distance – in opposite to political distance – does not directly encourage retailers to adapt their preferred modes at the time of market entry but rather influences them later. Consequently, retailers do not feel to be culturally forced to legitimize their entry modes at the time of market entry. Retailers may change their modes or adapt business formats when they are operating in the country because they may perceive the normative pressure more strongly at that time. The direct effect of cultural distance on the choice of full-control modes is positive; that is, retailers prefer full-control modes in distant countries. This is an unusual observation compared to meta- analytical studies in the field of manufacturing firms which have mostly indicated an opposite effect (Magnusson, Baack, Zdravko- vic, Staub, & Amine, 2008; Morschett et al., 2010; Tihanyi et al., 2005; Zhao et al., 2004). However, successful retailers standardize and control important retail format elements in cultural distant countries (Swoboda & Elsner, 2013). This might explain the choice of full control modes but also requires future exploration.
Hypothesis 4 posits that greater international experience results in a greater propensity to employ the preferred entry mode. The positive coefficient is significant (Exp(B) = 2.634; p < 0.05) and supports Hypothesis 4. However, as the effect is
less visible for firms that prefer full-control modes the upper part
of the typically s-shaped curve is shown in Fig. 3 to make the intersections of the curves more visible. The shift of the intersection to the upper right can be explained by the quite strong negative direct effects of international experience on mode choice. The latter demonstrates that experienced retailers tend to increasingly use shared-control modes and to share the risk of foreign expansion. For high international experience, Fig. 3 shows that retailers who prefer full-control modes are more likely to employ a full-control mode in subsequent market entries, whereas retailers who prefer shared-control modes are likely to employ shared-control modes in later market entries. Thus, the preferred entry mode is reinforced with increasing international experience. The direct effect of international experience on the choice of full- control modes is negative, what contradict studies on manufactur- ing firms (Zhao et al., 2004). However, we believe that our observations are plausible, as most retailers start their expansion abroad with full-control modes and use shared-control modes with increasing experience.
Hypothesis 5 suggests that internationalization speed rein- forces the propensity to employ a preferred entry mode, i.e., retailers that internationalize rapidly follow a strategic routine. The results support this assumption (Exp(B) = 2.497; p < 0.01).
With respect to high internationalization speed, Fig. 4 shows that
retailers preferring full-control entry modes are more likely to employ full-control modes in subsequent market entries, whereas retailers preferring shared-control mode are more likely to employ shared-control mode in later market entries. Thus, the usage of preferred entry modes is reinforced with increasing internationalization speed.

 According to previous observations of market entry mode choices, firm size and market size are significantly and positively linked with the choice of a full control mode. The effects of openness toward FDI, market growth and industrial development are not significant. Three firm dummies are significant, but the results remained unchanged when the dummy variables are omitted. The check for higher-order effects due to the investigated retailers by employing the generalized estimating equations (GEE) approach shows almost the same results with regard to strength, direction and significant level as illustrated.

We have also conducted tests in order to address performance implications of preferred modes. Because most retailers are not obligated to publish any performance data (at least on a country level) we use FDI survival in a country (=1 vs. withdrawal = 0) (1) in the first five years after entry and (2) in general as performance indicators (following Shaver, 1998). As tests for example correla- tions with the preferred mode were calculated, yielding values of
(1) 0.052 (p > 0.10) and (1) 0.123 (p < 0.05). Additionally we tested
a correlation of the survival rate of each firm’s subsidiaries with the strength of a firm’s preference for a certain mode, i.e., the percentage to which the preferred entry mode was used in all
entries over time. The results show coefficients of (1) 0.288 (p > 0.10) and (2) 0.123 (p < 0.05). These results indicate that firms with a higher preference for full-control modes and with a stronger
preference for their preferred mode tend to perform better in terms of FDI survival, at least in the long run.
1. Discussion
This study addresses the effects of preferred entry modes, i.e., modes that have been used most frequently in the past, on subsequent mode choice with a special emphasis on the boundary conditions of this relationship for retail firms. The results strongly indicate the imprinting of retailers’ mode choices, although the extent of the imprinting is limited by contextual factors. These observations provide two major research implications and conclusions for managers.
Concerning our first research aim, we should note that the results suggest that retailers’ preferred entry modes have a strong effect on later entry mode choices. This observation confirms previous research on the relevance of preferred entry modes (e.g., for manufacturer, Yiu & Makino, 2002; Lu, 2002) and extends this research by contributing to our knowledge in two regards. We add to the understanding of mode choice as a dynamic process and provide a more realistic view of entry mode choice over time (e.g., Benito et al., 2009; Brouthers & Hennart, 2007; Canabal & White, 2008). Mode habitualization may be more important for MNC than
cross sectional studies can reflect. We add to the knowledge of the foreign market entry mode choices of service firms, particularly retail firms that are rarely analyzed in previous entry mode research. Preferred modes may constitute their ‘‘rules of the game’’ (Meyer & Peng, 2005) also because preferred modes determine mode choice stronger than all further observed variables (see the direct effect in model 2 in Table 4). Various mechanisms may lead retailers to establish preferred entry modes, such as imprinting and bypassing. Particularly imprinting suggests that retailers evaluate entry mode options and decide on the processes and structures of the international expansion deliberately in the beginning and then learn if the choices fit with their established retail format chosen for international expansion. We believe that this mechanism explains why retailers establish preferred modes but independent of how they establish, these preferred modes strongly influence future mode choices. However, because this study is a first attempt to analyze preferred entry modes in retailing and because we know few about dynamic changes within international retail firms over time (e.g., Jonsson & Foss, 2011) we call for further research on this issue, especially on the antecedents of preferred entry modes.
With respect to our second research aim, we should note that assessing the boundaries of the effects of preferred entry modes on subsequent entry contribute to a deeper understanding of firms’ entry mode choices and to the theoretical explanations of the moderation effects (Meyer et al., 2009). We highlight two conclusions concerning the moderation effects. First, analyzing external influences on the effects preferred modes have on subsequent modes choices we provide a dual institutional explanation of firm behavior (Rosenzweig & Singh, 1991). Managers respond to internal cognitive institutions and to external (regulative and normative) institutions when choosing entry modes. The results contribute to the understanding of interplays and tradeoff decisions in a specific way. External institutional pressures diminish the effect of retailers’ habitualized behavior with regard to its entry mode choice, but surprisingly this can be shown for regulative institutions only (i.e., political distance; cultural distance at best by trend). We may conclude that retailers use a preferred mode until host country regulations force them to depart, but not so cultural norms. But, we may also conclude, that retailers deal with regulatory institutions prior to mode decisions, while normative institutions are not analyzed or considered to a lesser extent prior to mode choice decisions (maybe even generally, Swoboda & Elsner, 2013). However, although we observe frequently used institutional moderators, further ones should be considered to fully understand dual institutional pressures (e.g., retail-specific norms, Huang & Sternquist, 2007). Second, a firm’s internationalization knowledge reinforces its habitualized behav- ior with regard to mode choice, i.e. international experience and internationalization speed. Thus, both variables determine orga- nizational routines, such that experienced retailers as well as those which aim to expand quickly abroad develop a particular ‘formula’ or internal routines in making investment decisions across countries, which, in turn, reduces the ex-ante and ex-post risks of operations abroad. Institutionalized routines concerning pre- ferred entry modes allow rapidly internationalizing grocery retailers to achieve economies of scale, which is vital as they have to adapt their offers, for example. Also internationally experienced retailers rely on their ‘formula’, even when they have expanded first into psychic close and then into psychic distant countries (Swoboda & Elsner, 2013). In sum, this study contributes to our understanding of the contextual limitations of used preferred strategies which occur in two directions, decreasing and reinforcing habitualized behavior.
Managers of retail firms know their firms’ preferred entry modes and may acknowledge that a preferred entry mode is an effective choice. Although we found some indication for positive performance-effects of a preference for a certain mode, there is a danger of missing opportunities when relying too much on habitualized entry-modes. This may happen for example if such routines affect functions, steps, and criteria applied for market selection, which may result in consciously or subconsciously excluding possible target markets because the preferred entry- mode would not be applicable there. Managers must understand that entry modes determine long-term foreign performance (Gielens & Dekimpe, 2007) and that entry mode choice and market selection are distinctive decision processes that drive international performance. Huang and Sternquist (2007) conclude that retail managers must identify institutional pressures to understand the potential impact of those pressures and to strategically manage the external environment. However, this type of management may not be feasible in practice. This study identified factors that lead firms to reconsider their preferred entry modes but also factors that reinforce the use of preferred modes. Thus, managers face tradeoffs, e.g., to expand quickly but at the same time to respond to specific institutional pressures in a host country. Consequently, to manage the interdependencies between capabilities, entry modes, and environmental factors, managers must broaden their understanding of the interactions among these factors and must identify the boundaries of preferred entry modes.

1. Limitations and research directions

As previously noted, additional research should be conducted to better understand the boundary conditions of the effects of past entry mode decisions on current ones, also because the current study is not without limitations. We highlight four issues of this nature. First, although we devoted special attention to data collection, the scope of this study is limited. Because only grocery retailers were analyzed, the results may differ for others, i.e., specialized non-food retailers. However, examples of retailers from other industries such as The Body Shop, IKEA, Zara, and H&M emphasize the preference for specific entry modes as well. Second, in addition to addressing the previously mentioned alternative measures of context factors, future research should consider alternative moderators (e.g., external ones such as different distance measures according to Berry et al., 2010, or internal ones such as changes in ownership structure or top management teams) and alternative dependent variables, e.g. equity holdings in wholly owned subsidiaries or franchising contracts, which are particularly important in non-food retailing. Third, the moderators were partly measured with respect to the year of market entry, which may be problematic because grocery retailers must establish new value chains in host markets, which requires time. Thus, because the market entry decision process may begin years prior to entry or may occur during the year of entry, the specific time cannot be generally estimated. Finally, although we paid special attention to correcting for an endogenity/self-selection bias of the preferred entry mode further research should address the antecedents of preferred entry modes by analyzing variables such as the extent to which firm-specific capabilities are transferrable, e.g., preferences for a retail format for international transfers, or the role of inter- organizational mimicry (Lu, 2002). Although in our empirical data we found not clear indication, in theory it is possible that retailers prefer entering countries in which their entry mode is appropriate and refrain from entering further ones; thus country selection and entry mode selection might be entangled decisions.
References
Aiken, L. S., & West, S. G. (1991). Multiple regression: Testing and interpreting interactions.
Newbury Park, CA: Sage.
Alexander, N., & Quinn, B. (2002). International retail divestment. International Journal of Retail and Distribution Management, 30(2), 112–125.

Anderson, E., & Gatignon, H. (1986). Modes of foreign entry: A transaction cost analysis and propositions. Journal of International Business Studies, 17(3), 1–26.
Bell, D. E., Lal, R., & Salmon, W. J. (2004). Globalization of retailing. In J. A. Quelch & R. Deshpande´ (Eds.), The global market: Developing a strategy to manage across borders (pp. 288–312). San Francisco, CA: John Wiley & Sons.
Benito, G. R. G., Petersen, B., & Welch, L. S. (2009). Towards more realistic conceptua- lisations of foreign operation modes. Journal of International Business Studies, 40(9), 1455–1470.
Berger, P. L., & Luckmann, T. (1967). The social construction of reality. London: Penguin.
Berry, H., Guille´ n, M. F., & Zhou, N. (2010). An institutional approach to cross-national distance. Journal of International Business Studies, 41(9), 1460–1480.
Bianchi, C. C., & Ostale, E. (2006). Lessons learned from unsuccessful internationaliza- tion attempts: Examples of multinational retailers in Chile. Journal of Business Research, 59(1), 140–147.
Brouthers, K. D. (2002). Institutional, cultural and transaction cost influence on entry mode choice and performance. Journal of International Business Studies, 33(2), 203–221.
Brouthers, K. D., Brouthers, L. E., & Werner, S. (2008). Resource-based advantages in an international context. Journal of Management, 34(2), 189–215.
Brouthers, K. D., & Hennart, J.-F. (2007). Boundaries of the firm: Insights from interna- tional entry mode research. Journal of Management, 33(3), 395–425.
Canabal, A., & White, G. O. (2008). Entry mode research: Past and future. International
Business Review, 17(3), 267–284.
Chan, P., Finnegan, C., & Sternquist, B. (2011). Country and firm level factors in international retail expansion. European Journal of Marketing, 45(6), 1005–1022.
Chang, S.-J. (1995). International expansion strategy of Japanese firms: Capability building through sequential entry. Academy of Management Journal, 38(2), 383–407.
Chang, S.-J., & Rosenzweig, P. M. (2001). The choice of entry mode in sequential foreign direct investment. Strategic Management Journal, 22(8), 747–776.
Cohen, J., Cohen, P., West, S. G., & Aiken, L. S. (2003). Applied multiple regression/
correlation analysis for the behavioral sciences (3rd ed.). Mahwah: Lauwrence Erlbaum Associates.
Colombo, M. G. (1995). Firm size and cooperation: The determinants of cooperative agreements in information technology industries. International Journal of the Economics of Business, 2(1), 3–28.
Cyert, R. M., & March, J. G. (1963). A behavioral theory of the firm. Englewood Cliffs, NJ: Prentice Hall.
Dawson, J. A. (1994). Internationalization of retailing operations. Journal of Marketing Management, 10(4), 267–282.
Diamantopoulos, A., & Winklhofer, H. M. (2001). Index construction with formative indicators: An alternative to scale development. Journal of Marketing Research, 38(2), 269–277.
DiMaggio, P. J., & Powell, W. W. (1983). The iron cage revisited: Institutional isomor- phism and collective rationality in organizational fields. American Sociological Review, 48(2), 147–160.
Doherty, A. M. (2000). Factors influencing international retailers’ market entry mode strategy: Qualitative evidence from the UK fashion sector. Journal of Marketing Management, 16(1), 223–245.
Erramilli, K. M. (1991). The experience factor in foreign market entry behavior of service firms. Journal of International Business Studies, 22(3), 479–501.
Evans, J., & Mavondo, F. T. (2002). Psychic distance and organizational performance: An empirical examination of international retailing operations. Journal of International Business Studies, 33(3), 515–532.
Forest, J., Mehier, C., & John, R. (2001). Commons and Herbert A. Simon on the concept of rationality. Journal of Economic Issues, 35(2), 591–605.
Gamble, J. (2010). Transferring organizational practices and the dynamics of hybrid- ization: Japanese retail multinationals in China. Journal of Management Studies, 47(4), 705–732.
Garcı´a-Canal, E., & Guille´ n, M. F. (2008). Risk and the strategy of foreign location choice in regulated industries. Strategic Management Journal, 29(10), 1097–1115.
Gielens, K., & Dekimpe, M. G. (2001). Do international entry decisions of retail chains matter in the long run? International Journal of Research in Marketing, 18(3), 235–259.
Gielens, K., & Dekimpe, M. G. (2007). The entry strategy of retail firms into transition economies. Journal of Marketing, 71(2), 196–212.
Goldman, A. (2001). The transfer of retail formats into developing economies: The example of China. Journal of Retailing, 77(2), 221–242.
Goldman, A., Ramaswami, S. N., & Krider, R. E. (2000). Barriers to the advancement of modern food retail formats: Theory and measurement. Journal of Retailing, 78(4), 281–295.
Grewal, R., & Dharwadkar, R. (2002). The role of institutional environment in marketing channels. Journal of Marketing, 66(3), 82–97.
Heckmann, J. J. (1979). Sample selection bias as a specification error. Econometrica,
47(1), 153–162.
Ha˚ kanson, L., & Ambos, B. (2010). The antecedents of psychic distance. Journal of International Management, 16(3), 195–210.
Hannan, M. T., & Freeman, J. (1984). Structural inertia and organizational change.
American Sociological Review, 49(2), 149–164.
Henisz, W. J. (2000). The institutional environment for economic growth. Economics and Politics, 12(1), 1–31.
Henisz, W. J., & Delios, A. (2001). Uncertainty, imitation and plant location: Japanese multinational corporations, 1990–1996. Administrative Science Quarterly, 46(3), 443–475.
House, R. J., Hanges, P. J., Javidan, M., Dorfman, P. W., & Gupta, V. (2007). Culture, leadership, and organizations: The globe study of 62 societies. Thousand Oaks, CA: Sage.