Archetypes of Driver Combinations Leading to Foreign Market Exit: An Investigation into European Grocery Retailing

Existing research into retailers’ foreign market exits has uncovered a number of drivers that lead multinational retailers to divest from certain countries. While scholars have investigated these drivers in isolation from one another, combinations of drivers that affect divestment decisions simultaneously remain under-researched despite scholars having indicated their importance and having called for a detailed, more holistic analysis. In this study, using a case study approach and qualitative content analysis of a wide variety of publicly available contemporary documents from different perspectives and experts, we investigate the drivers leading to all 32 country exits of the 50 largest grocery retailers in Europe in the 5-year period between 2014 and 2018. In line with previous research, the study shows the most frequent exit drivers are a low performance of the subsidiary, a low performance of the parent company, and a strategic refocus of the parent company. However, we demonstrate that for most exits, combinations of multiple interrelated drivers at the subsidiary level, the host-country level, and the parent level have a joint influence on retailers’ decisions to exit foreign markets. We also show that exits often include both failure-related drivers and strategy-related drivers. Furthermore, using the configurational approach, we identify exemplary combinations of market exit drivers that occur frequently and propose five archetypes of such combinations that suffice to explain all market exits in the dataset. For future research, we propose extending our typology through the application of a similar approach to different contexts and to use quantitative research based on the qualitative findings to generate more generalizable results.


Introduction
The internationalization of companies is now viewed not as a process of uninterrupted progression and linear expansion but as a process of constant transformation, characterized by periods of both growth and retrenchment (e.g., Burt, 1991;Swoboda & Schwarz, 2006;van de Ven & Poole, 1995;Wrigley & Currah, 2003). International exit strategies are regarded as important and natural features of companies' international expansion paths (e.g., Berry, 2009;Palmer & Quinn, 2007). Consequently, scholars have intensively investigated foreign market exits over the past decades, with an emphasis on the divestment activities of manufacturing companies (for overviews see, e.g., Arte & Larimo, 2019;Coudounaris, 2017;Schmid & Morschett, 2020;Trąpczyński, 2016).
Retail-specific divestment research, however, has received much less attention. This research gap is remarkable, particularly because many large retail divestments have occurred in the past decades, most of which generated strong public attention. Between 2005 and 2020, the ten largest international retailers in the world exited 43 markets (EDGE Retail Insight, 2020). More specifically, the grocery retail sector in Europe recently witnessed several high-profile grocers exit several markets. For instance, the French retailer Carrefour exited Greece and ceased all its operations in the Balkans between 2015 and 2018. The Finnish S Group exited Latvia and Lithuania in 2017, and the British icon Marks and Spencer closed its department stores in 13 countries in Europe as a part of a global restructuring effort between 2016 and 2017.
Scholars have called for investigating international retail divestment separately from international manufacturing divestment , because retail differs from other sectors and its internationalization and de-internationalization follow a different logic (Burt et al., 2003;. Burt et al. (2002, p. 195) argued: "The nature of retailing and the practice, importance and meaning of its internationalization are for example vastly different to an oil producer or a water company". For example, manufacturing MNCs have a multitude of motives to enter foreign countries (Dunning, 1988), whereas a retailer's main motive to open subsidiaries in a foreign country is market-seeking. Moreover, the implications of exit decisions differ between retail companies and manufacturing companies. If retailers exit foreign markets, it inherently ends their ability to sell products to these markets, whereas manufacturing companies can continue to serve markets without a local manufacturing presence (McDermott, 2012). Consequently, retail-specific foreign divestment drivers may differ from drivers identified in manufacturing-based divestment literature. This becomes evident when comparing literature reviews on manufacturing-based international divestment in general (e.g., Arte & Larimo, 2019;Coudounaris, 2017;Schmid & Morschett, 2020;Trąpczyński, 2016) with those on retail-specific international divestment (e.g., Alexander & Quinn, 2002;. Divestment research is characterized by studies that investigate drivers which affect foreign divestment decisions. Many studies highlight that divestment is driven by a combination of drivers simultaneously, rather than by individual drivers (e.g., Aklamanu, 2015;Burt et al., 2018;Yoder et al., 2016). While this assumption is mentioned frequently, it is rarely investigated further. Despite recent calls to investigate combinations of drivers and their joint effect on foreign divestment decisions (e.g., Berry, 2009;Burt et al., 2018;Schmid & Morschett, 2020), there are no studies-with the exception of Burt et al. (2018)-that employ this strategy.
In studies that focus on retail, scholars classify foreign retail exits and their drivers based on different criteria (e.g., Alexander & Quinn, 2002;Burt et al., 2002;Mellahi et al., 2002;Palmer, 2004). These typologies, however, usually focus on one or just a few drivers and do not sufficiently consider the intricate combinations of exit drivers from different levels that may trigger divestment decisions.
Our investigation comprises two major parts. First, we qualitatively analyze all 32 market exits that occurred amongst the 50 largest grocery retailers in Europe between 2014 and 2018 and investigate the drivers behind these exits. To do so, we take a case study approach and apply qualitative content analysis (Krippendorff, 2018;Schreier, 2012). This approach offers new insights into the type of drivers that trigger exit decisions, the frequency of their occurrence, and their combinations. We focus on grocery retailing because it has outstanding relevance in the overall retail sector. Six of the ten largest retailers in the world are grocery retailers (Deloitte, 2020). In Europe, grocery retail sales accounted for 45% of total retail sales in 2019 (EDGE Retail Insight, 2020). By choosing a specific sector, we delimit our dataset with sector-related and geographical case boundaries. This allows for a "highly visible and numerically manageable set of actions to be observed and analyzed" (Burt et al., 2018, p. 179). Second, using the configurational approach (Macharzina & Engelhard, 1991;Meyer et al., 1993;Miller, 1986), we propose five novel archetypes of exit driver combinations. The configurational approach highlights that organizational phenomena should be investigated based on multidimensional combinations of variables that dynamically interact (Miller et al., 1984). Rather than forming an infinite number of combinations, numerous dimensions of strategies, structures, etc., the variables are often grouped into a limited number of frequently occurring combinations referred to as "archetypes" (or "configurations" or "gestalts" Cerrato et al., 2016;Lim et al., 2006;Macharzina & Engelhard, 1991;Meyer et al., 1993;Roth, 1992). Archetypes may be developed conceptually or empirically (Cerrato et al., 2016;Meyer et al., 1993) and be established at various levels of analysis, "depicting patterns common across individuals, groups, departments, organizations, or networks of organizations" (Meyer et al., 1993(Meyer et al., , p. 1175. In our case, we developed archetypes empirically, representing similar and reoccurring multidimensional combinations of exit drivers at the organizational level. Building archetypes is not a novel approach in the international business literature. For instance, Cerrato et al. (2016) developed multidimensional archetypes of internationalizing SMEs, Lim et al. (2006) developed a typology of international marketing strategies of MNCs, and Roth (1992) developed archetypes for the basic configuration and coordination patterns of medium-sized firms in global industries.
Using this qualitative approach enables us to develop a full picture of each market exit, including the "wider environmental setting and institutional context" surrounding exit decisions (Burt et al., 2018). This method allows us to understand the nature of configurations and the importance of the different drivers within these configurations. Such insights cannot be obtained in the usual quantitative studies that are based on large secondary databases. Furthermore, we apply our qualitative approach to a large set of market exits. In the past, scholars have called for studies that cover more retailers than the few that were repeatedly investigated Cairns et al., 2008;Etgar & Rachman-Moore, 2007). Due to the non-selectiveness of our dataset, we cover all market exits of the most important grocery retailers in Europe, including the less prominent exits. Hence, we deliver a balanced picture of the phenomena.
In our study, we found five distinct multidimensional archetypes, which include simultaneous and interrelated drivers at the subsidiary, the host-country, and the parent-company level. The presence of multidimensional archetypes within our set of retailers' market exits supports the value of the configurational approach in helping researchers develop a better understanding of these phenomena, and thus helps extend theory on the subject. In line with the findings of Burt et al. (2018), who recently showed that different configurations of exit drivers exist over time, we investigate different configurations that co-exist simultaneously.
The rest of this paper is structured as follows. The second section provides a literature review on research about drivers that lead to foreign market exit, with a focus on the field's past efforts to characterize and classify retailers' international divestments and their drivers. In the third section, the research approach is introduced, providing an overview of the dataset and the applied methodology. The fourth section illustrates the individual exit drivers (and their combinations) found in the analysis. This is followed by a derivation and discussion of the five archetypes of driver combinations found in the data with a focus on the typologies discussed previously. The paper ends with a discussion and a conclusion composed of theoretical and empirical contributions, managerial implications, limitations, and implications for further research.

Literature Review
Research about international divestment drivers of MNCs has been an increasingly important area of interest in international business literature since the mid-1970s, mainly focusing on manufacturing firms. While research coverage on international divestment still lags behind research on international entry and expansion (Tan & Sousa, 2019), recent literature reviews (e.g., Arte & Larimo, 2019;Coudounaris, 2017;Schmid & Morschett, 2020;Trąpczyński, 2016) show many existing studies on international divestment drivers. Schmid and Morschett (2020), for example, detected 283 studies investigating international divestment drivers. Furthermore, their meta-analysis showed significant effects for ten variables that have been analyzed frequently in previous studies: subsidiary establishment via acquisition, subsidiary entry mode via wholly-owned subsidiary, subsidiary ownership level, subsidiary product unrelatedness, subsidiary financial performance, parent-firm financial performance, parent-firm R&D intensity, parent-firm advertising intensity,

3
Archetypes of Driver Combinations Leading to Foreign Market… host-country specific international experience of the parent firm, and host-country economic growth.
In the retail-specific literature, research on divestment drivers began later and progress has been slower (Burt et al., 2018). Early studies provided insights into the volumes and patterns of retail divestment (e.g., Alexander et al., 2005;Burt et al., 2004;Godley & Fletcher, 2001). Later, the emergence of qualitative case study research contributed to the understanding of international retail divestment by investigating market exits of select prominent retailers. Some of these, namely Tesco, Marks and Spencer, and The Home Depot, were investigated particularly often . While a few of these studies investigate the process of divestment and the effects on retailers, most of the studies examine drivers that lead to divestment.
Several authors indicated that divestment decisions are often the result of multiple interdependent drivers (e.g., Aklamanu, 2015;Burt et al., 2003;Cairns et al., 2010;Mellahi et al., 2002;Swoboda & Schwarz, 2006). While many studies have detected multiple exit drivers (e.g., Aklamanu, 2015;Yoder et al., 2016), there is a paucity of studies investigating combinations of drivers. The recent study of Burt et al. (2018) is a first but valuable step to address this void. By qualitatively investigating the main rationale for European grocery retailers' exit from East Asia over a 30-year period, Burt et al. (2018) pointed to the existence of three broad phases of retail exit over time, each characterized by different dominating drivers that lead to market exit. In the "pioneer adjustment" phase, retailers mainly exit after realizing the difficulties related to scaling their activities and benefitting from first mover advantages. In a second phase, called "resistance and market consolidation", different forms of resistance across the whole region (e.g., re-regulation) prevent retailers' expansion. In the last phase, "reconfiguring the global firm", the exit is driven by pressures from outside the region (e.g., firm-level reassessment of global activities). The main drivers of each phase differ among retailers, revealing the dynamic nature of retail exits in this region.
While Burt et al.'s (2018) study provides initial evidence for the existence of retail divestment driver combinations, it leaves research gaps. First, the researchers found retail divestments' main drivers to be highly dynamic across time. They adopted a long-term perspective on the highly dynamic Asian retail market, which experienced a financial crisis during the period of analysis. Whether the results of such a dynamic setting are transferable to other, less volatile settings (e.g., the European grocery retail sector) is unclear. Second, Burt et al. (2018) linked each combination of drivers to a specific period of time. However, they did not address the question of whether different combinations co-exist simultaneously in one phase. Third, although they examined combinations of multiple drivers, their focus lay, in each phase, on the dominant narrative for the exits rather than on the interrelations of the drivers in each combination and their joint effect on exit decisions.
There have been several attempts to classify retail divestments and their respective drivers in existing literature. Retail divestments are differentiated from one another by their tactical or strategic nature. While the divestment of foreign operations can be viewed as a reactive tactical response by retailers to difficulties they encounter in foreign markets (for example, when retailers exit markets in which they perform poorly), an exit can also be a strategic action responding to a broader change 1 3 of circumstances unrelated to the specific country being exited (Alexander & Quinn, 2002;Cairns et al., 2010). Cairns et al. (2010) differentiated between two types of strategic divestments. In "positive restructuring" divestments, retailers re-allocate their resources when they observe changing conditions within the international marketplace: that is, they proactively redirect their resources to markets offering higher potential in the long term. Researchers who have examined positive restructuring efforts of retailers have found that they transfer resources from divested markets to markets in which they enjoy a stronger market position (Burt et al., 2002;Cairns et al., 2010;Coe & Wrigley, 2017) or to markets with higher expected returns (Alexander et al., 2004;Berry, 2009;Cairns et al., 2010;Palmer & Quinn, 2007). On the other hand, "corporate crisis" divestments arise from difficulties at the corporate level that require retailers to refocus resources and capabilities on these issues and often result in selling operations in foreign markets to generate liquidity for domestic operations (Burt et al., 2002;Cairns et al., 2008).
Failure-related divestments can be divided into those triggered by drivers external to the company and those triggered by drivers internal to the company (Burt et al., 2002). Burt et al. (2003), working from Benito (1997), subdivides these failures into four broad categories: market failure (e.g., divestment triggered by unexpected negative macroeconomic or political changes in the host country) and competitive failure (e.g., divestment triggered by an unexpected inability to compete with local retailers) are argued to be external drivers; organizational failure (e.g., divestment triggered by a mismatch between the business of the retailer and the host-country operations) and business failure (e.g., divestment triggered by weak operations of the retailer in the home country) are seen as internal drivers. Mellahi et al. (2002) argued that it is often a mixture of internal and external factors that lead to divestment. Swoboda and Schwarz (2006) differentiated between exits mainly driven by (1) the internal environment, (2) the external environment, and (3) both the external and internal environment. Thus, both Mellahi et al. (2002) and Swoboda and Schwarz (2006) acknowledged that these two rationales are not mutually exclusive. Gersch and Franz (2018) followed a similar line of thought by categorizing divestment drivers into push factors (i.e., host market factors driving the retailer to divest through, for example, harmful government regulations or aggressive competitive behavior) and pull factors (i.e., home-market factors driving the retailer to divest through, for example, stakeholder pressures to reallocate resources). Aklamanu (2015) focused on external drivers and investigated failure in international markets from an institutional perspective. He discussed the importance of three types of institutional pressures (i.e., regulative, normative, and cognitive pressures) that can lead to failure when retailers are unable to cope with these pressures. Similarly,  suggested five different categories of pressures retailers might encounter in foreign markets. These are government re-regulation, competition from domestic retailers, pressures from suppliers, consumer movements, and collective action by workers.
In the general foreign subsidiary divestment literature, divestment drivers are often categorized into three levels: (1) drivers at the parent-company level (e.g., financial performance of the parent company), (2) drivers at the host-country level (e.g., political and economic risk in the host country), and (3) drivers at the subsidiary level (e.g., financial performance of the subsidiary) (e.g., Berry, 2009;Tan & Sousa, 2020).
Most of the above-presented rationales are rooted in the resource-based view and related to the idiosyncratic resource base of retailers. Thereby, the international portfolios of firms and the specific characteristics of the international businesses within these portfolios (e.g., the financial performance of a business or the economic development of a host country where the business is located) can be viewed as resources that firms have developed over time and whose optimal use and constant optimization should be the objective of their strategies (Belderbos & Zou, 2009).
Overall, the literature discussed above has advanced the conceptualization of foreign retail divestments and their drivers. Most importantly, it has helped identify and classify different types of divestment drivers. However, it does not sufficiently appreciate the multiple interrelated factors from different levels that drive divestments. For example, divestments might be classified as "business failure" in the typology proposed by Burt et al. (2003) when a retailer under pressure at home needs to refocus strategically. However, the reasons for the retailer to divest from certain markets while continuing its activities in other foreign markets may still be the result of "market failure" or "competitive failure" in these markets. In fact, Burt et al. (2003) suggest that these categories should be used as a "simple explanatory device" and that divestment decisions "may be affected by several of these factors at the same time" (Burt et al., 2003, p. 363). This demonstrates the difficulty of classifying divestments into one-dimensional typologies and indicates the need for a more robust classification of foreign retail divestments by interrogating different types of divestments and the complex patterns of drivers that characterize them.

Approach and Methodology
To identify patterns of retail divestment drivers, we took an embedded singlecase study approach (Yin, 2017). We thereby study the case of foreign divestment within the grocery retailing industry in Europe with its relatively homogenous context and consider the 32 market exits as embedded subunits of this case. We used qualitative content analysis on a large number of documents including retailers' exit narratives (Krippendorff, 2018;Schreier, 2012) to identify the characteristics and context of foreign retail exits and investigate the drivers of each market exit. According to Schreier (2012), content analysis allows authors to engage in some degree of interpretation to arrive at the meaning of the investigated data. It is a method that can be used in multiple ways and has gained popularity for analyses in management, international business research (Gaur & Kumar, 2018), and strategic management (e.g., Osborne et al., 2001;Palmer et al., 1997). We applied extensive data-based triangulation to confirm our findings.

Dataset Construction
We looked at the 50 largest grocery retailers in Europe (incl. Russia), based on their 2018 revenues in Europe, and identified all of their market exits in Europe in the 5 years from 2014 to 2018. 1 To this end, we used two large retail databases, namely the 2020 Edge Retail Insight Database (formerly Planet Retail) and the 2019 LZ Retailytics database. While different forms of international retail divestment exist (Alexander & Quinn, 2002;Burt et al., 2002), from the closure of single stores to the complete termination of foreign activities, the focus of this paper is on "market exit", defined by Burt et al. (2003, p. 359) as the "total withdrawal of a firm from an operational presence in a foreign market". Even though market exits are heterogeneous among each other-leaving a country can mean divesting one store or hundreds of stores-they share a common denominator: the full termination of a retailer's presence in a country. Opening our investigation to other types of divestment (e.g., store reduction) would not allow for this common denominator and would add unnecessary heterogeneity to the analysis. By beginning the time window no earlier than 2014, we avoid the potential for our results to be strongly influenced by the effects of the global economic crisis [see Pederzoli and Kuppelwieser (2015) for an overview on retailers' internationalization behavior during the crisis].
Amongst the 50 largest grocery retailers in Europe, we identified 32 market exits undertaken by twelve different grocery retailers between 2014 and 2018 (see Appendix 1). Two retailers exited multiple countries within this period (13 exits by Marks and Spencer; six exits by Carrefour). Three retailers exited two countries and the remaining seven retailers exited one country each. Twenty-seven retailers had no market exits amongst their international operations, while eleven retailers had no international operations. We cross-checked our findings with a range of other sources (e.g., press articles, annual reports) to avoid missing any market exits. We are confident that our dataset contains all exits of the 50 largest grocery retailers in the given period.

Building the Characteristics and Context of Foreign Retail Exits
For each retailer's market exit, we developed a full picture of its characteristics (see Appendix 1 for a selection of characteristics). We conducted an extensive inquiry based on a range of international sources, including narratives on the retailers' exits. These sources include, inter alia, home-and host-country-based retail trade press, international retail trade press, retailers' corporate communications, market research reports, business biographies, retailers' websites and annual reports, and large public databases such as the World Bank. Retail exits are usually discussed extensively in the trade press of the home and host country, including statements by company officials, but also through the additional and often more critical perspectives of journalists. Market entries and exits are also critical events mentioned in the annual reports 1 3 Archetypes of Driver Combinations Leading to Foreign Market… of the retailers. By analyzing these sources, we were able to follow each operation from its initial entry until the market exit, enabling us to consider the exits within the broader context of the retailers' activities.

Coding Drivers of Market Exits
Next, we investigated the drivers that led to each market exit by content analyzing the previously mentioned sources. These sources provide commentaries relating to the reasons behind the retailers' exit decisions. Investigating publicly available sources (vs. interview-based research) is a common practice in the foreign divestment field (Burt et al., , 2018Finnegan et al., 2019), since divestment is a highly sensitive topic (Alexander et al., 2005;Benito & Welch, 1997;Burt et al., 2003;Cairns et al., 2008;Palmer, 2004). Therefore, divestment rationales are often either erased from corporate memory altogether or, if recollected by individuals who made exit decisions, "retro-fitted to suit post-event re-evaluations" (Burt et al., 2018, p. 179), thus only offering a subjective and distorted picture (Burt et al., 2018;Finnegan et al., 2019). Other authors, such as Golden (1992) and Huber and Power (1985) have pointed out the limits of interpreting retrospective reports and warned about the risks of informants inaccurately recalling events because of memory lapses, inappropriate rationalization, or a tendency to present themselves or their organizations favorably. Content analyzing a wide variety of contemporary documents from different perspectives and experts, on the contrary, offers a more accurate picture of the exit and is unobtrusive and relatively devoid of both researchers' demand biases and informants' recall biases (Gaur & Kumar, 2018).
Our data analysis included two main steps: (1) Following Gioia et al. (2013), we began with the first-order analysis. Using the qualitative data analysis computer software package NVivo, we content analyzed and coded 689 documents and identified the drivers (first-order codes) leading to each market exit. 2 We used an inductive, data-driven coding approach without limiting ourselves in the number of codes assigned, which ensured that the drivers would emerge from the data during the analysis. Two authors collectively coded about 150 documents to align their understanding of the drivers, to design an initial coding-frame with drivers at an optimal level of differentiation, and to set decision rules for the coding. The main coding was then conducted separately by two authors. Following this step, we compared the coded drivers (first-order codes) for each of the 32 market exits and discussed them with the third author.
In particular, we examined the overarching story of each market exit, identified key drivers that led to the exit, and disregarded drivers that played a less relevant role. We thereby considered the actual context of each market exit, including the context on the parent-specific level. This, in turn, allowed us to understand that certain drivers prevail over others, rather than solely counting the existing drivers for each market exit in an overly simplistic, accounting-like manner. Disagreements between the authors concerning the drivers for each market exit were discussed and solved. However, percentages of agreement between the coders were high (90%) across all market exits, demonstrating high inter-coder reliability (Neuendorf, 2017). This first step generated 34 drivers for market exits (first-order codes) that are illustrated in Appendix 2.
(2) We then engaged in the second-order analysis (Gioia et al., 2013), scrutinizing the market exits and their different drivers (first-order codes) several times, seeking common themes between the drivers that allowed us to regroup the drivers into second-order constructs. This regrouping provided a more aggregated level of representation that was needed for the later development of archetypes, since the first-order codes were too detailed for this purpose. This second step created nine exit drivers (second-order constructs) that are illustrated in Appendix 2 and in Table 2.

Triangulation of Exit Narratives
Various authors have identified the risk of overemphasizing the "official corporate line", which might be distorted toward positive rationales for divestment rather than offering an objective image of the actual events (Alexander et al., 2005;Gersch & Franz, 2018;Palmer, 2004;Palmer & Quinn, 2007). Considering this issue carefully, we adopted several measures to minimize it. First, while we used companyauthored documents to investigate the official corporate line, non-company-authored documents dominate the analysis. Second, we were careful to consult a large variety of press-related documents from the home-, host-and sometimes third-country perspectives for each market exit to ensure an unbiased perception of these exits and their drivers. Third, we included analyst views from leading trade press [e.g., The Grocer, Libre Service Actualités (LSA), Lebensmittelzeitung (LZ)] on exits. For the less prominent market exits, where these sources were lacking, we contacted analysts from commercial retail intelligence providers (e.g., EDGE Retail Insight), asking them to comment on exit rationales from their perspective. Fourth and finally, where possible, we verified the coded exit driver narratives with hard data. For example, we researched the actual market position of subsidiaries in host countries to verify the "lack of scaling perspectives" narrative. To follow up on the "bad economic situation in the host country" narrative, we researched the nominal GDP development of host countries during the 5 years prior to exit. For the "low performance of the parent company" narrative, we looked into the development of retailers' net income during the 5 years prior to exit. Finally, to verify the "strategic refocus of the parent company" narrative, we checked, for each market exit, the retailers' other worldwide market exits 5 years prior and up to 5 years after the investigated market exit to determine whether the exit truly was part of a broad refocus strategy. We changed the coding if the data clearly contradicted the narrative we had previously coded. For example, for three market exits, we decoded the strategic refocus narrative due to clear evidence against this corporate narrative explanation. As a result of the above-mentioned measures, we believe we have drawn an accurate picture of the drivers of each market exit.

Creating a Typology of Exit Driver Combinations
To create a typology of exit driver combinations, we first deductively attributed the nine exit drivers to the three levels of divestment drivers often found in the literature (e.g., Berry, 2009;Tan & Sousa, 2020): the subsidiary level, the host-country level, and the parent-company level (see top of Table 2). Then, based on the configurational approach (Macharzina & Engelhard, 1991;Meyer et al., 1993;Miller, 1986), we built archetypes of exit driver combinations. The configurational approach postulates that a relatively small number of typical configurations exist and describe most existing combinations (Miller et al., 1984). In our context, configurations (or "archetypes") represent exemplary, reoccurring combinations of exit drivers at different levels that explain retail market exits (Lim et al., 2006).
To detect different archetypes of exit driver combinations, we began by listing all 32 market exits and describing the specific combinations of exit drivers existing on the three levels for each market exit (Ragin, 2009). Next, we searched for repeating patterns in the data through visual inspection of the exits. In a stepwise approach, we then rearranged the market exits within our list to unite similar patterns (see Table 2). This allowed for the creation of five distinct archetypes of exit driver combinations that we discuss in detail in the fourth section.
We did not use quantitative clustering methods, as these were inapplicable due to the nature of the data. The limited number of retailers investigated and the nested structure of the dataset would have created unreliable results (McNeish & Harring, 2017). Instead, through a qualitative approach, we obtained a rich level of knowledge of each market exit (i.e., embedded subunit of the case study)-useful for recognizing patterns and manually building a typology. In fact, pattern identification and typology construction are often combined with qualitative content analysis, especially when researchers are navigating a large number of cases (Schreier, 2012), or, in our case, embedded subunits.

Findings: Archetypes of Exit Driver Combinations
For a better understanding of the retail market exits, one descriptive finding is noteworthy: In most market exits, the divested country operations only represented a small share of the total sales of a retailer. In 27 market exits, the share of sales in the host country compared to the total sales of the retailer was below 0.6%. ICA Group's operation in Norway represented the largest subsidiary, with 13.1% of total sales. Marks and Spencer's divested operations represented a cumulative 0.91% of total sales when it exited 13 markets. Carrefour divested six markets, representing 2.08% of total sales. This suggests that most single market exits are important but not major events for a retail company. However, for some retail companies, several market exits occur within a rather short period (e.g., Marks and Spencer exited 13 markets within 2 years).
Our empirical investigation revealed nine drivers that explain the exit decisions of all 32 international market exits. Figure 1 illustrates the frequency with which these drivers occurred within our market exits.
Of the nine drivers that were found to influence exit decisions, three drivers could be attributed to the subsidiary level, four drivers to the host-country level, and two drivers to the parent-company level.
The most frequent subsidiary-level driver is the "low performance of the subsidiary" (23 exits), which is also the most frequent driver overall. However, other subsidiary-level drivers, namely "issues with subsidiary partners" (seven exits) and "lack of scaling perspectives" for the subsidiary (six exits), also often drive exit decisions.
Two parent level drivers are found in many market exits: "strategic refocus of the parent company" in 21 exits and "low performance of the parent company" in 19 exits. While these findings show the importance of these drivers, their high frequency must be interpreted with caution because the occurrence of a parentlevel driver often causes several exits of the same retailer from different countries, thus making this driver appear in several market exits (e.g., Marks and Spencer exits).
At the level of the host country, the "political situation" played a role in six exits, the "competitive situation" in five exits, and the "negative retail market outlook" in four exits. The "bad economic situation in the host country", however, was only relevant in two exits. It is the least frequent driver across all three levels. This low occurrence may be induced by the overall positive economic climate in Europe in the investigated period.
The drivers presented above, their main rationale leading to market exit, and the exemplary driver-related quotes are exhibited in Table 1.
By applying the previously described methodology, our data revealed five archetypes of multidimensional exit driver combinations, each representing a distinct configuration of drivers that led several retailers to exit a foreign market.  Low performance of the subsidiary Shareholders put pressure on management to divest unprofitable assets (e.g., Christopherson, 2007;Palmer, 2004;Wood et al., 2016) to avoid negative effects on the firm's valuation and to strengthen the balance sheet (Hamilton & Chow, 1993) "ICA Group will leave Norway.
[…] ICA Norway is not viable on its own. For many years the ICA Group has had to make large capital contributions to keep the business afloat." a Per Strömberg, CEO, ICA Group (Dagligvarehandelen, 2015) Lack of scaling perspectives Limited scale and weak market positions lead to high operational costs (Chung et al. 2013;Norbäck et al., 2015) and do not allow the subsidiary to be profitable and competitive Comprises current lack of scale but also lack of prospects to scale subsidiaries in the long term (e.g., because of challenges in securing adequate real estate or too-high investments) "We decided to renew our international business strategy. As part of this process, we concluded that the market share of Prisma stores in Latvia and Lithuania is not high enough for profitable business in the near future." b Jorma Vehviläinen, CEO, S Group (Sputnik News, 2017) Issues with subsidiary partners Conflicts of interest and diverging strategies between partners lead to instability, a possible partnership termination, and divestment (e.g., Burt et al., 2003;Etgar & Rachman-Moore, 2007;Gaur & Lu, 2007;Hennart et al., 1998) Partner (e.g., in IJV, Franchising) is unable to continue activities in the host country (e.g., because of bankruptcy) Leclerc and Conad ended their partnership in Italy because of conflict of interest when they became part of competing alliances. The two retailers had "profound differences with regard to the future organization of the partnership and its strategic orientation" (PGC 2014) "Greece's Marinopoulos Group, which holds the franchise rights for the Carrefour supermarkets in Bulgaria, has filed for bankruptcy protection" (SeeNews, 2016)

Host-country level drivers
Bad economic situation in the host country Economic or financial turbulence negatively affects a company's performance (Belderbos & Zou, 2009;Burt et al., 2018;Chung et al., 2010) Economic growth of a host country determines its market attractiveness and long-term potential by creating business opportunities and allowing for higher profits (Benito, 1997; Berry, 2013) "The low level of crude oil prices and its follow-up, the dive of the ruble and the decline in Russian purchasing power have influenced the decision of the Finnish company." (Turun Sanomat, 2016) Intermarché (2015) exited Serbia during a time of economic instability in the country. In the year before the exit, Serbia's nominal GDP had decreased by 2% (World Bank, 2020) Political situation in the host country Political or regulative host-government related pressures create uncertainty and risk for companies (Aklamanu, 2015;Burt et al., 2003;Coe & Wrigley, 2017;Gersch & Franz, 2018;Song, 2014) Countries with weak institutions do not offer a stable legislative environment to companies (Henisz & Delios, 2004) Interstate conflicts cause problems for companies (e.g., nonpayment of creditors, damage of physical assets, sanctions) (Al Khattab et al., 2007;Burmester, 2000) "No firm in Latvia is protected from unreasonable insolvency proceedings due to an unclear interpretation of the bankruptcy law […] the decision to stop investing in the Latvian market was also accelerated by the insolvency lawsuit filed against the company in Latvia in 2012." c Emil Stefanov, CEO, Palink (Lyrtas, 2014) Negative retail market outlook Industry growth rates are an indicator for future market opportunities (Alexander et al., 2005;Belderbos, 2003;Hennart et al., 1998) "The rationale of the [exit] decision is that the Danish market is considered saturated so that it does not offer satisfactory prospects for growth." Press release, Metro  Competitive situation in the host country Markets characterized by high competitive intensity are less attractive (Aklamanu, 2015;Gersch & Franz, 2018) "The retail sector has invested heavily in both countries [Latvia & Lithuania] in recent years, and as a result, the square footage of retail space in relation to population and consumer purchasing power is already quite high."

2017)
Parent-company level drivers Strategic refocus of the parent company Resources gained from divestments are required for projects with higher expected returns (Alexander et al., 2004;Berry, 2009;Cairns et al., 2010;Palmer & Quinn, 2007) Resources gained from divestments are required for existing business that is under pressure (Burt et al., 2002;Cairns et al., 2010;Coe & Wrigley, 2017; "We strictly need to align our country portfolio with earnings and growth targets and focus on those regions in which we are in a position to extend our market shares." Olaf Koch, CEO, Metro Group (2014) Archetypes of Driver Combinations Leading to Foreign Market… Low performance of the parent company Weak parent companies cannot support their foreign subsidiaries and may need to divest foreign operations to generate liquidity for domestic operations (Burt et al., 2002;Cairns et al., 2008Cairns et al., , 2010McDermott, 2012) Prior to its restructuring effort initiated by Steve Rowe, Marks and Spencer's net income decreased by 79% within 5 years (Marks & Spencer, 2014 Table 2 presents the market exits, their respective drivers, and the corresponding archetypes. In the following, each archetype is described in detail.

Archetype 1: Divestment of Struggling Subsidiaries in Unattractive Markets
The market exits in this archetype are characterized by retailers that divest poorly performing subsidiaries from difficult markets. In these market exits, the subsidiaries already exhibit a low financial performance at the time of the decision, indicating existing problems with the operations that require substantial financial and Archetypes of Driver Combinations Leading to Foreign Market… managerial attention. Furthermore, the retailers cannot see ways for improving the performance of these subsidiaries in the future for two reasons. First, they are pessimistic about being able to adequately scale the subsidiaries to enable them to compete successfully against local competition. Second, challenging political and macroeconomic conditions, the competitive situation, or simply an undesirable development of the market (e.g., lower demand) in the host countries are overwhelming and add additional hurdles to a successful continuation of the business.
This archetype was developed from four market exits (ICA Group Norway, REWE Group Latvia, Salling Group United Kingdom, Kesko Group Russia). In the market exit of the ICA Group's from Norway, the retailer faced a poorly performing subsidiary and was unable to solve the underlying operational problems due to unfavorable regulatory decisions.
"The Norwegian operations have been characterized by negative sales and profit growth for a number of years, partly driven by high purchasing prices and logistics costs. After the Norwegian Competition Authority gave notice at the beginning of the year that it did not intend to approve the planned sourcing cooperation with Norgesgruppen, we began looking more actively at alternative solutions." Annual Report, ICA Group (2015) In another market exit, Palink, the REWE Group joint venture in Latvia, found itself operating in a difficult retail market. The CEO of the Salling Group, a competitor and the market leader in Latvia at that time, commented on the exit of Palink: "The fact that Palink is leaving Latvia only confirms the complexity of the retail sector in that country, with relatively low profits. The market is highly competitive, consumer purchasing power is low." Valdis Turlais, CEO, Salling Group (Lyrtas, 2014) Because of both a lack of scale that did not allow them to compete successfully in such a difficult industry and the financial harm caused by an unjustified insolvency lawsuit, the REWE group and its JV partners decided to sell their lossmaking supermarket banner Iki to the Latvian retailer Mego and thereby exited the country.
"In Latvia our market share was too small to compete successfully. The decision of our foreign investors to stop investing in the Latvian market was also greatly influenced by the insolvency proceedings initiated in 2012 against our company in Latvia. All these factors led to the decision to leave the Latvian market" Emil Stefanov, CEO, Palink (Alfa, 2014)

Archetype 2: Divestment of Subsidiaries Due to Partnership Issues
The market exits in this archetype are characterized by retailers that exit foreign markets because their international partnerships ends when the host country partner fails in some form, or conflicts between partners emerge.
In the joint venture between Leclerc and Conad, a conflict occurred because the two retailers joined competing alliances. At the parent-company level, Leclerc decided to join a new buying alliance. This was not compatible with continuing its joint operations with Conad in Italy. Overall, the two retailers had "profound differences with regard to the future organization of the partnership and its strategic orientation" (PGC, 2014).
Carrefour's exit from Greece and four other countries in Southeastern Europe (Albania, Bulgaria, Cyprus, and North Macedonia) was an unavoidable consequence of the bankruptcy of its exclusive franchise-partner, the Greek Marinopoulos Group. While the precipitating factor for the exit was the Greek Marinopoulos Group's bankruptcy, previous decisions by Carrefour had possibly aggravated Marinopoulos's crisis and incited its bankruptcy. Because of the economic crisis and related uncertainty in Greece, Carrefour had left the joint venture with Marinopoulos in 2012 (Kathimerini, 2017) and shifted to a franchising model in all five markets. When Marinopoulos filed for bankruptcy in 2015, the local retailer Sklavenitis acquired the remaining assets (Kathimerini, 2017) but had no interest in taking over the franchising agreements with Carrefour. Similarly, Carrefour saw its franchised operations in Slovakia come to an end when its franchise-partner Retail Value Group was no longer able to sustain operations of its Carrefour outlets due to severe financial difficulties (Aktuality, 2018).

Archetype 3: Divestment of Struggling Subsidiaries by Parents in a Corporate Crisis
This archetype is characterized by exits undertaken by retailers that faced negative circumstances at the corporate level, primarily because they were under significant pressure in their domestic markets. This pressure led them to abandon struggling international subsidiaries and to focus their strength on more urgent matters. In total, 15 market exits (Delhaize Group Bosnia and Herzegovina, Delhaize Group Bulgaria, and all 13 Marks and Spencer exits) can be classified under this archetype.
Marks and Spencer faced significant difficulties at the corporate level, mainly driven by a challenging home-market environment with tough competition from food discounters, but also further aggravated by their international operations' weak performance. Under strong pressure to turn the business around, the retailer, with the arrival of its new CEO Steve Rowe in 2016 (The Grocer, 2016), decided to exit poorly performing international operations, most of which were mainly wholly owned operations. Well-performing operations, mainly those operated through a franchise system, were not divested. […]. Our review has looked at the performance and potential of each of our international markets. Internationally, we propose to cease trading in ten loss making owned markets, but intend to continue to develop our presence through our strong franchise partners." Steve Rowe, CEO, Marks and Spencer (2016) The supermarket chain Delhaize was also confronted with considerable pressures in its home market Belgium, where the retailer faced fierce competition, and witnessed its net income decrease by 65% in the 5 years prior to exiting Bulgaria and Bosnia and Herzegovina (Delhaize . Both exits of Delhaize are in fact part of a broader strategic refocus of the retailer, which manifested itself as an almost total withdrawal from Southeastern Europe (except for Serbia). Shortly before the exit from Bulgaria and Bosnia and Herzegovina, the retailer had already left Albania (2013) and Montenegro (2013) (Reuters, 2014). Interestingly, for Delhaize, as well, the arrival of a new CEO was important in triggering the refocus strategy.

"We have now completed a forensic review of our estate both in the UK and in our international markets. Our international business consists of two parts: an owned business that is loss making in a number of markets and a profitable franchise business
"With the arrival of Frans [Muller], the Executive Committee took the opportunity to review our strategy and develop a new strategic framework." Annual Report, Delhaize  Like Marks and Spencer, Delhaize, under significant corporate pressure, decided to review their international portfolio and chose to divest only the poorly performing operations (e.g., Bulgaria), while continuing to rely on their stronger international operations.
"We want to focus on the most profitable markets such as Serbia, Romania or Greece. But this was not the case in Bulgaria." Company Statement, in L'Echo (2014)

Archetype 4: Divestment of Struggling Subsidiaries in Unattractive Markets by Parents in a Corporate Crisis
In this archetype, exiting retailers were confronted with troubles at all three levels: the corporate, the subsidiary, and the host-country level. This archetype was developed from four market exits (S Group Latvia, S Group Lithuania, Metro Group Denmark, and X5 Retail Group Ukraine). It is particularly well illustrated in S Group's exit from Latvia. Scrutinizing the narrative for this exit reveals issues at all levels. First, fierce competition caused pressure in the domestic market. (2017) S Group's annual reports revealed its dismal financial performance in the 5 years before the exit from Latvia, with an average return on sales of just 0.16% (S Group, , 2017a.

"Discontinuing operations in Latvia and Lithuania will allow the retailer to focus on its home market of Finland. There, its main competitor Kesko is investing heavily, also bolstered by divesting of overseas markets and non-core businesses." Daniel Johansson, Retail Analyst, LZ Retailytics
Second, its Latvian subsidiary was performing poorly and lacked scale. Third, the grocery retail sector in Latvia was becoming more competitive as local retailers were investing heavily. Fourth, S Group's Latvian operations were harmed by the authorities' intervention that closed one of its stores for months because of "alleged" security issues with the building.

"During our revision, we understood that the Latvian and Lithuanian Prisma market share is not high enough to ensure the profitability of operations and that obtaining enough additional volume through organic growth would be very challenging in the next few years. […] Competitors have invested heavily in both countries in recent years
, with the result that the square footage of trade relative to population and consumer purchasing power is already quite high. In addition, our operating conditions in Latvia were significantly weakened by the closure of Prisma Deglava last year and the subsequent termination of the lease." 3 Jorma Vehviläinen, CEO, S Group (2017a, b)

Archetype 5: Proactive Divestment of Subsidiaries to Optimize Resource Allocation in the International Portfolio
This archetype describes market exits that are undertaken because a retailer shifts its resources to other markets that it identifies as being more lucrative in the long term. Both market exits of Intermarché (Serbia and Bosnia and Herzegovina) build this archetype. These market exits were part of a shifting focus from Southeastern Europe towards Western European markets, where the retailer had more promising expansion opportunities, and were unrelated to performance (LSA, 2015). Intermarché had already started this refocus effort in 2012 by exiting Romania (Advertiser Serbia, 2015).

Discussion
The previous section offered a detailed explanation and discussion of the five archetypes. To allow for an overarching discussion of the archetypes and the drivers that form these archetypes, the findings are illustrated in the form of a Venn diagram (see Fig. 2). Figure 2 provides a clear picture regarding the importance of the different levels of drivers. Subsidiary-level drivers clearly dominate amongst the market exits-they are relevant in 30 of 32 market exits and are present in four out of five archetypes. Parent-level drivers are the second most important type; they occur in 21 of 32 market exits and are present in three out of five archetypes. Finally, host-country level drivers play a role in only eight market exits and in only two out of five archetypes, making it the least frequent level of drivers in the dataset. This outcome is similar to what is found in manufacturing divestment research. Even though the exact variables investigated in retail and manufacturing divestment differ, in both research streams, subsidiary-level drivers are the most important ones, followed by parentlevel drivers. Finally, while host-country-level drivers are the least prominent drivers in our retail-specific dataset, we find evidence that they do, in fact, play a role in explaining some foreign market exits. This differs from manufacturing divestment, where literature review results demonstrate most host-country-level drivers do not play a significant role in triggering foreign subsidiary divestment (e.g., Arte & Larimo, 2019; Schmid & Morschett, 2020). Figure 2 highlights that exit drivers most frequently occur simultaneously on several levels. Only nine of the market exits in this study exclusively involved drivers from a single level: Seven market exits were triggered solely by subsidiary drivers (archetype 2), while parent-level drivers alone only led to two market exits (archetype 5). Archetypes 1, 3 and 4 are characterized by configurations of drivers on several levels; these archetypes represent 23 out of 32 market exits. Most of these market exits (19) were dyadic-comprising two different levels (archetypes 1 and 3). In four market exits, drivers from all three levels played a triadic-type role in the divestment decision (archetype 4).
Comparing the five archetypes with the typologies from previous retail divestment studies, we find that most market exits in our study can be attributed to several categories within extant typologies.
In the first archetype, exits follow the logic of tactical divestments rather than strategic divestments, since the retailers decide to divest failing operations abroad. On the one hand, drivers at the subsidiary level, i.e., lack of scale and operational underperformance of the units, lead to "competitive failure" in the country and can be seen as an underlying motive for divestment. Financial difficulties at the subsidiary level create pressure for management teams and often lead to divestment decisions (Benito & Welch, 1997;Berry, 2013;Feldmann, 2016;Finnegan et al., 2019).

Number of market exits
No.  Boddewyn (1979) stressed that such decisions are influenced not only by the current but also by the expected future financial performance of foreign operations. On the other hand, drivers at the host-country level, i.e., the local competition authority denying a merger (as for ICA Group) and an insolvency lawsuit (as for REWE Group) lead to "market failure" and act as precipitating circumstances that eventually trigger the exit decision. Hence, following the typology of Burt et al. (2003), we can observe both simultaneous "competitive failure" (i.e., operational underperformance) and "market failure" (i.e., unfavorable political and regulatory conditions) that lead to these market exits. This observation illustrates how subsidiary-level and host-country-level drivers mutually reinforce themselves to trigger exit decisions. This scenario also aligns with Mellahi et al. (2002)'s arguments, highlighting that internal and external factors often simultaneously affect foreign retail divestment. Internal factors are not necessarily related to operational mistakes of the retailer in their international operations (e.g., non-adaption to the host-country culture), as argued by Burt et al. (2003), but can be related to the natural hurdles of international expansion (e.g., scalability of subsidiary operations). Furthermore, external factors regularly do not drive divestment directly; instead, they do so indirectly by reinforcing internal drivers. Several scholars have discussed the interaction of subsidiarylevel and market-level drivers. Berry (2013) indicates that a high market potential can offset the influence of poor performance on retailers' divestment decisions. Instead of divesting the operations, retailers may be willing to invest further financial and managerial resources into poorly performing subsidiaries when the market is attractive. If the market is challenging per se, however, as in the above described market exits, it is difficult for retailers to justify further investments into such subsidiaries. Wood et al. (2016) emphasize that retailers have to protect their reputation towards shareholders and financial institutions by demonstrating a coherent and adequate capital allocation in their international expansion activities.
In the second archetype, exits are predominantly caused by changes in circumstances related to international partnerships. This contrasts starkly to the other archetypes, where more complex and interrelated motives lead to the divestments.
In the third archetype, exits follow the logic of "corporate crisis" divestments as described by Cairns et al. (2010), or "business failure" as described by Burt et al. (2003). The archetype includes exits by two retailers: Marks and Spencer and the Delhaize Group. Both retailers abandoned their poorly performing international operations, as they distracted the retailers' attention from more pressing strategic priorities, i.e., negative developments at the corporate level. Thus, both retailers strategically refocused their international operations: While Marks and Spencer discontinued a large portion of its wholly owned operations, Delhaize almost completely withdrew from Southeastern Europe. The 13 discontinued Marks and Spencer operations represented only a cumulative 0.91% of total sales. However, together with four exited countries outside of Europe (EDGE Retail Insight, 2020), as part of the same refocus effort, Marks and Spencer left 17 of the 47 markets in their international portfolio virtually overnight. For both Marks and Spencer and Delhaize, it was a newly appointed CEO that initiated the refocus strategy. This observation delivers further evidence for the frequently discussed importance of a change in senior management as a precipitating circumstance in prompting foreign divestments Archetypes of Driver Combinations Leading to Foreign Market… (Alexander & Quinn, 2002;Cairns et al., 2010;McDermott, 2012). While the need to refocus on the domestic market is a strategic driver for divestment, the choice of which specific markets are to be exited is a tactical one. In this archetype, parent companies in crisis only divest their poorly performing foreign subsidiaries. Consequently, the forces that drive a retailer to divest from specific markets while continuing its activities in others are the result of "competitive failure", i.e., operational underperformance of the subsidiary (Burt et al., 2003). This demonstrates that the exit decisions in this archetype are driven by a dyadic relationship between homeand subsidiary-level drivers. Authors have developed various arguments to explain this type of dyadic relationship. On one hand, parent companies with financial difficulties are often unable to satisfy their poorly-performing subsidiaries' needs for financial and managerial resources and thus divest them (Torneden, 1975). On the other hand, the tolerance of a parent company towards foreign losses or loss potentials may be lower when faced with difficulties in its home market (Torneden, 1975). Shareholders and financial institutions expect corrective measures, such as exiting the lossmaking markets, to safeguard the financial well-being of the retailer (Palmer & Quinn, 2003).
In the fourth archetype, exits follow a similar underlying logic as in archetype 3. In both archetypes 3 and 4, the exits are driven by retailers experiencing a crisis at the corporate level and feeling the need to refocus at a strategic level. Under these circumstances, they decide to exit certain international markets. What differentiates the fourth archetype from the third is that the exited operations are not only performing poorly but they are also suffering from difficult circumstances at the host-country level such as a difficult economic and/or political situation, a negative retail market development, or an increasingly challenging competitive situation. In archetype 3, we argued that struggling parent companies would not be able or willing to support their weakly performing subsidiaries and they would be willing to support them even less in unattractive markets with no promising future prospects. Following the reasoning of Burt et al. (2003), these exits are the result of not only "business failure" and "competitive failure" but also "market failure". In contrast to archetype 3, where we observed a dyadic combination of drivers, in this archetype, we observe a triadic relationship between drivers on the parent, the subsidiary and the host-country level.  and Doherty and Alexander (2015) have emphasized that retail divestments are often triggered by both conditions in the home country (e.g., deteriorating market conditions and financial pressures in the home country) and multiple host-country and subsidiary-level conditions at the level of the divested operations (e.g., weak subsidiary performance due to competitive pressure and harmful regulations).
In the fifth archetype, exits follow the logic of "positive restructuring" as discussed by Cairns et al. (2010). Retailers leave markets proactively-in this case, divesting operations for which they see only limited prospects relative to other active or potentially new operations. Scholars have argued that well-managed retailers change their international strategy and divest foreign operations to focus on good performers when their resources are not allocated optimally within the international portfolio (Alexander et al., 2005;Berry, 2009;Hamilton & Chow, 1993;Hennart et al., 1998). In this archetype, it is solely the attractiveness of operations unrelated 1 3 to the observed market exits that lead retailers to divestment decisions. Resources gained from the divested operations are used for investments in markets with higher expected returns (Alexander et al., 2004;Berry, 2009;Palmer & Quinn, 2007). By allowing other, more lucrative investments to occur, foreign divestments represent essential strategic means in an internationalization process that is a continuum of retrenchment and growth. This archetype is thus distinguished from archetypes 1 through 4, since in these other archetypes, exits represent corrective actions.
As the preceding discussion shows, most market exits can be explained by several categories within the existing typologies of foreign retail divestment. This confirms the need for a multidimensional typology of foreign market exit drivers.
As mentioned previously, a qualitative approach was used to investigate market exits and to form the archetypes. This approach included developing a full picture of the characteristics for all market exits. Some of these characteristics represent the dynamics of contemporary market exits particularly well and deserve to be discussed in more detail.
Most of the retailers in our study were active in many different countries simultaneously: As previously illustrated, the divested countries often only represented a small share of the retailers' total sales.
In several market exits, a change in senior management at the retailer helped precipitate exit decisions. Marks and Spencer, Intermarché, the Delhaize Group, and the Kesko Group all changed their top management 1 year prior to their exits. Within this category, all but the Kesko Group's exit belong to archetypes characterized by strategic refocus motives (in archetypes 3 and 4). This offers further evidence for the importance of the often-discussed role of leadership changes in prompting strategic reviews that lead to foreign market exits (Cairns et al., 2010;McDermott, 2012).
Retailers often exited multiple countries in the same period. While six exits were "isolated" market exits (archetype 1 consists of only such market exits), twelve exits were part of a regional restructuring wave (in archetypes 2, 3, 4, and 5) and fourteen exits were part of a global restructuring wave (in archetypes 3 and 4). The latter market exits are mainly attributed to Marks and Spencer, which exited from Eastern and Southeastern Europe, Central Europe and, looking beyond our dataset, China, in the same period.
Retailers often viewed exiting several countries simultaneously as the only way forward when considering the many challenges they faced in their home markets (Coe & Wrigley, 2017). Although this type of exit would only have minor effects on retailers' overall sales figures, it would free up significant managerial resources and allow them to focus on their challenges at home (in archetypes 3 and 4) or regions or countries where they had identified more promising growth opportunities (in archetype 5). Furthermore, in the case of a successful operation sale to a third party, the divestment proceedings might ease a struggling retailer's financial situation. However, our data showed that retailers were only able to sell their foreign operations in one third of the market exits.
An interesting divestment driver, visible in archetypes 1 and 4, was grocery retailers' realization that they could not achieve the necessary scale in a country to be competitive there. While this issue only clearly influenced divestment decisions in six of the investigated market exits, it might have played a certain role in other exit decisions. Interestingly, the divested subsidiaries were only among the three biggest retailers in the host country in four of the 32 market exits. Thus, most subsidiaries did not have a leading position in their markets, making it difficult for them to be competitive. McDermott (2012) discussed the importance of achieving scale in international markets for the retailer Carrefour. After the appointment of its new CEO José Luis Duran in 2005, Carrefour initiated a phase of strong refocus, exiting from markets where it failed to achieve a leadership position. Our data shows that this mindset has been growing in popularity within the grocery retail sector. The increasing prominence of a longer phase of consolidation and regional focus in the grocery retail industry has only recently begun to be discussed in more depth (e.g., Burt et al., 2018;Coe & Wrigley, 2017;Palmer & Quinn, 2007;Wood et al., 2016). While this phase originated in the mid-2000s  with Marks and Spencer's large-scale withdrawal from continental Europe, the global economic and financial crisis further strengthened these patterns. An increasing number of multinational retailers were forced to shift their attention away from globalization and to instead concentrate on challenges in their domestic markets (Coe & Wrigley, 2017). In both our findings and those of Burt et al. (2018), we observe that this phase of regional and home-market-refocus exits is still ongoing. Specifically related to the findings of Burt et al. (2018), we show that European grocery retailers, in their quest to "reconfigure the global firm", divest not only from geographically distant markets like Asia but also from operations close to home.

Conclusion
This paper began with the premise that the identification of multidimensional combinations of drivers leading to foreign market exits is essential to truly understanding the phenomenon of foreign divestment. We analyzed the exit drivers of all 32 country exits that occurred amongst the 50 largest grocery retailers in Europe in the 5 years from 2014 to 2018. By doing so, we were able to draw a full picture of contemporary retail divestment in this region. Our findings take the specific and complex context of each individual market exit into account, while still considering the overarching issues across all market exits.

Theoretical and Empirical Contributions
Our paper contributes to literature in several ways. First, we identified nine drivers and their combinations that explain all exit decisions. Empirically, we show that these drivers do not have the same frequency of occurrence. While some drivers affect almost all market exits (e.g., low performance of the subsidiary), others only occur rarely (e.g., bad economic situation in the host country). This finding indicates that some drivers are more relevant than others.
Second, we responded to the explicit calls by different researchers not only to investigate individual drivers of foreign subsidiary divestment but also to consider combinations of drivers that affect exit decisions simultaneously (e.g., Berry, 2009;Burt et al., 2018;Schmid & Morschett, 2020). Thus, we adopt a holistic, multidimensional approach to analyzing individual retail market exits. As a theoretical contribution, our findings confirm the existence and importance of exit driver combinations for each individual exit. Many of our investigated market exits feature a simultaneous occurrence of multiple exit drivers at the subsidiary, the host-country, and the parent-company level. Third, on an aggregated level, our concept of retail divestment driver archetypes arguably represents the first truly multidimensional way of classifying combinations of drivers of retail market exits. While Burt et al. (2018) demonstrated the dominance of different driver combinations over time, we show that a number of distinct types of driver combinations co-exist at the same time. As our main theoretical contribution, we developed an initial typology of exit driver combinations by classifying reoccurring combinations of exit drivers into archetypes. These archetypes were described in detail and reviewed in the discussion section. We propose a limited set of driver configurations that represent all the different types of exits in our database. The market exits in archetype 1 are characterized by retailers divesting struggling subsidiaries in unattractive markets. Archetype 2 includes retailers that exit foreign markets as a result of ending international partnerships. In archetype 3, retailers facing a corporate crisis and experiencing financial difficulties at both the corporate and foreign subsidiary levels divest their struggling subsidiaries in a strategic refocus effort. In archetype 4, drivers at the host-country level are added to reinforce the drivers related to financial difficulties at the subsidiary and parent-company levels; thus, retailers within this archetype are confronted with troubles at all levels. Archetype 5 is characterized by proactive exits by retailers that strategically focus their resources and capabilities on more promising markets. Our typology makes no claim to be exhaustive, as almost no typology does (Miller, 1996). Nevertheless, it can serve as a tool for further investigation of foreign retail exits because familiar patterns can be recognized and categorized more easily without having to explain the drivers behind each exit in detail.
Fourth, we empirically show that some archetypes appear more frequently than others. Archetype 3 represents 15 market exits, archetype 2 represents seven market exits, archetypes 1 and 4 represent four markets exits each, and archetype 5 represents two market exits. This result is partially driven by the nested structure of our dataset in which multiple market exits are sometimes nested within one retailer. However, it must be noted here that all archetypes except archetype 5 (composed of two Intermarché exits) are built of market exits stemming from at least two retailers. Furthermore, nested structures of market exits, as seen in this dataset, represent a real manifestation of the phenomenon of foreign market exits and thus deserve to be investigated properly. This means that all 32 market exits, regardless of whether they belong to the same or different retail companies, must be investigated and, in our case, needed to be considered when building the archetypes. It is imperative, however, that the nested structure is considered when investigating market exits from the perspective of foreign operations in order to avoid distorted findings. In quantitative research, such multiple market exits from the same retailer are often treated in isolation. However, with the deeper understanding of each market exit gained through qualitative research, it was possible to consider the nested relationships in the interpretation of the findings. In this respect, the nested structure also has advantages because it allows parent-company level influences to be separated from subsidiary level influences.
Fifth, we contribute to theory by showing that previous typologies often do not suffice to describe exit types fully since they lack multidimensionality. More specifically, we contribute to the discussion about the importance of strategic motives to foreign divestment, especially in contrast to failure-related motives (e.g., Cairns et al., 2010;Palmer & Quinn, 2007;Sousa & Tan, 2015;Tan & Sousa, 2019). Indeed, the study confirms that exits are often part of a broad strategic refocus of retailers (as in archetypes 3, 4, and 5). However, it also provided evidence that proactive and positive refocus exits, as sometimes discussed in divestment literature, are rather rare (appearing in two market exits from one retailer in archetype 5). In several market exits, retailers attempt to hide their exits under the guise of proactive strategic changes as part of their international strategy to optimize their investments. By using a wide array of sources and investigating the market exits in depth, we were able to illustrate that most strategic exits, even if part of a broader refocus strategy, are reactive responses to challenges at the subsidiary, host-country, and/or parent-company level (as seen in 19 market exits by five retailers in archetypes 3 and 4).
Sixth, as another theoretical contribution, we find evidence of differences between manufacturing firms' and retail firms' divestments. While manufacturing divestment research found that host-country level drivers do not play a significant role overall in triggering MNCs to divest their subsidiaries, we find evidence that host-country level drivers do play a role in explaining certain foreign retail divestment decisions. This might be related to the internationalization motives of manufacturing companies and retailers. As mentioned before, manufacturing MNCs and international retail companies differ in their motives to open subsidiaries in foreign host countries. Retailers' main motive is market seeking, which makes them vulnerable to market changes. We found eight market exits that were strongly influenced by changes regarding the macroeconomic, political, competitive, or retail market environment in the host country. Manufacturing MNCs, in contrast, have a multitude of motives to enter foreign countries, including resource seeking and efficiency seeking (Dunning, 1988). Thus, changes in the host market may not deteriorate their future prospects. This could be a reason why previous research on manufacturing divestment has found little evidence for host-country drivers that trigger foreign divestment (e.g., Arte & Larimo, 2019;Schmid & Morschett, 2020). However, the effects also may be more difficult to measure in manufacturing research since positive and negative effects offset each other due to the multiple motives of these operations (e.g., an increasing income level is positive for the market-seeking motive but negative for the resource-seeking motive). While this question cannot be answered in this study, it would be an interesting issue to explore in the future.

Managerial Implications
Our study also provides implications for retail practitioners who can learn from the divestment experiences of other retailers (Ajai, 2015;Alexander et al., 2005;Palmer, 2004). However, this should be done with caution because our study does not allow normative conclusions per se, as it includes no analysis of the performance effects of the divestments. Still, divestment in itself can be viewed as an outcome, and our results show drivers of that outcome.
First, grocery retailers can use our overview on recent foreign exit patterns in the European market to better understand the foreign divestment dynamics in their industry. In Appendix 1, they can find extensive information about the 32 market exits and their context. Second, by understanding the drivers that frequently lead multinational retailers to exit markets, retailers can be better prepared for exit drivers that might arise within their own international expansion.
Third, retailers can try to proactively influence certain drivers to keep them from becoming exit drivers in the future. For example, by understanding that many foreign exits are driven by the limited scalability of a foreign subsidiary, retailers can be more attentive to this issue when entering host countries. From this perspective, integrating our findings into retailers' decision-making processes might diminish their probability of having to exit certain markets.
Fourth, retail managers can use the identified archetypes of exit drivers when revising their portfolio of foreign markets and can use it to question the "raison d'être" of their different foreign operations. By applying the archetypes, managers, who are sometimes reluctant to divest from foreign markets even though they would fulfill all criteria to be divested (Nargundkar et al., 1996;Torneden, 1975), can be convinced to do so. By examining the market exits discussed in our contribution, managers might realize that other retailers have divested similar operations and may, consequently, reduce their own barriers to divest their foreign operations.

Limitations
Regarding the generalizability of findings and counterexamples, this case study has the inherent limitations of qualitative research (e.g., Sinkovics et al., 2008;Wood et al., 2016). Analytical generalizations drawn from case study research, no matter how rigorous the methodology, deserve caution. More specifically, the scope of the dataset is a limitation. Because this study focuses on the grocery retail sector in Europe and recent market exits, the findings might not be transferable to other contexts (e.g., different retail sectors, industries, regions, and periods of time). Certain drivers that were found to be relevant may be less relevant in other contexts and vice versa. Furthermore, regarding the legal ownership structure, most retail companies in our dataset are publicly listed corporations. The effect of financial performance at the subsidiary and/or parent-company level may not be the same for non-listed companies that do not face the same pressure to react to short-term results (Palmer & Quinn, 2003).
Our dataset has a nested structure, i.e., multiple market exits are sometimes nested within one retailer or one host country. This structure causes certain retailers to dominate certain archetypes (i.e., Carrefour exits in archetype 2; Marks and Spencer exits in archetype 3). As previously mentioned, this limitation was taken into account in the qualitative approach. However, in the future, it would undoubtedly be beneficial to investigate market exits from a larger number of parent firms to see whether they can be attributed to the archetypes identified in this study.
The archetypes in this study are identified based on 32 market exits of large grocery retailers. Thus, our study demonstrates the existence of certain archetypes of exit driver combinations but does not provide an exhaustive typology. We believe that additional archetypes could be identified when investigating more market exits.
Our qualitative content analysis is based on publicly available sources and does not include interviews with the company actors responsible for the exit decisions or those close to them. We believe that by investigating large and public retailers, and by consulting a wide range of sources and expert opinions, we were able to capture an accurate picture of the nature of the country exits in our dataset. However, publicly available sources do not allow insights into certain internal issues, no matter how much data-triangulation is applied. For example, a retailer might not want to report internal problems leading to failure and might prefer to blame the divestment on bad financial results due to the host country's poor macroeconomic situation. Some authors also criticize annual reports because they are sometimes prepared by public relations specialists rather than the top management team (Abrahamson & Hambrick, 1997) and may be distorted in the attribution of organizational actions and outcomes (Barr et al., 1992;Clapham & Schwenk, 1991). Conducting additional interviews with people involved in the exit decisions would allow a more extensive triangulation of sources and ensure a higher degree of validity and reliability.
Lastly, our study focuses on divested subsidiaries and does not investigate retail subsidiaries that survive. Therefore, the conditions under which similar driver configurations do not lead to divestment could not be analyzed.

Implications for Further Research
We showed that, for most market exits, combinations of drivers rather than single drivers lead to foreign market exit. Thus, this study lends further support to the configurational approach in general (see, e.g., Cerrato et al., 2016;Lim et al., 2006;Miller, 1986Miller, , 1996, but importantly also within divestment research. Especially when considering the classification of strategic divestments vs. failure-related divestments, we see the necessity of a more integrated, multidimensional approach. Our findings are based on a limited set of market exits in Europe that took place between 2014 and 2018 and, thus, depend on this context. It is therefore important to apply a similar approach in different contexts (e.g., different retail sectors, industries, regions, and periods of time) to determine, first, whether the archetypes found in our case-study context also represent market exits in other contexts and second, which additional archetypes can be identified. This research would also help to interrogate the importance of the different archetypes. For example, we only detected two genuine "proactive restructuring" divestments such as those discussed by Cairns et al. (2010). However, this archetype may be found more frequently beyond the scope of our study. For instance, in 2020, the German hypermarket retailer Kaufland ceased its activities in Australia shortly before opening its first stores in the country, following investments of over USD 500 million in real estate. There is no indication that the local situation in Australia had any influence on this decision. Instead, Kaufland argued that it would refocus on the growth opportunities in Europe, where it had a large and attractive investment opportunity at the time (Kaufland, 2020).
Based on the broader scope, it would also be worthwhile to investigate the association of existing archetypes with the specific context; this research would be similar to the study by Burt et al. (2018) that links specific configurations of divestment drivers to different periods of time.
By building archetypes based on similar underlying logic, we demonstrate that not each market exit has to be dismantled with regard to the detailed drivers that lead to exit but that a relatively small number of drivers suffices to identify a specific archetype. This permits an efficient investigation of large sets of market exits based on a smaller set of describing variables. Thus, quantitative research should be applied based on our qualitative findings to analyze their generalizability and robustness. Quantitative retail divestment studies could apply clustering methods to identify multi-dimensional configurations of divestment drivers. Such approaches have proved to be very promising in other areas of international business research (e.g., Hagen et al., 2012;Ronen & Shenkar, 2013). However, in this case, researchers must explicitly address the nested multilevel data-structure in this field to avoid generating distorted findings (Peterson et al., 2012). Quantitative methods would also allow the inclusion of surviving operations for comparison. However, instead of the usual regression-based studies that focus on individual drivers (or simple interactions), our findings indicate that combinations of drivers should be investigated together.
Finally, we see significant potential in conducting further in-depth qualitative research based on more longitudinal datasets. Taking only a few retailers for a multiple-case study and analyzing their exits over a long period would allow researchers to investigate how drivers (and combinations thereof) vary over time within different companies. Burgelman (1994), in his work on the withdrawal of the microchip manufacturer Intel from the semiconductor business, already illustrated that withdrawals usually develop "co-evolutionarily" within companies and thus represent a multilayered, complex research problem which is best investigated holistically.