One way or another? An international comparison of expatriate performance management in multinational companies

Correspondence Fabian Jintae Froese, University of Goettingen, Platz der Goettinger Sieben 5, 37073 Goettingen, Germany. Email: fabian.froese@gmail.com Abstract Due to the high costs and strategic importance of expatriate assignments, expatriate performance management (EPM) plays an increasingly important role for multinational enterprises (MNEs). However, research on EPM is still in its infancy. Drawing from the convergence/divergence debate in international human resource management, this study investigates and compares EPM strategies and practices across MNEs from three different country clusters to better understand whether EPM practices tend to converge, diverge, or crossverge (i.e., show aspects of both). Results from surveying 132 Anglo-Saxon, Germanic, and Japanese MNEs reveal prominent differences (divergence) at the EPM strategic level such that Japanese MNEs tend to pursue more ethnocentric staffing strategies and design EPM systems specifically tailored to expatriates. On the practice level, we found both commonalities and differences between Japanese and Anglo-Saxon and Germanic MNEs, pointing toward crossvergence. Theoretical and practical implications of our results are discussed.

. Research based on institutionalist theory suggests that country context strongly affects HRM practices and points toward country-specific differences (divergence) (Björkman, Fey, & Park, 2007;Doellgast & Marsden, 2019;Welch, 1994). In fact, research posits that MNE HRM practices are particularly sensitive to the institutional environment in the host country of MNEs' subsidiaries (Rosenzweig & Nohria, 1994). However, there is some disagreement on the topic as it is argued that particular MNEs fuel trends of HRM practices convergence. MNEs are considered carriers of globalization which spread managerial knowledge and techniques internationally through the dissemination of best practices (Evans, Pucik, & Björkman, 2011). Research on MNEs, therefore, frequently points toward convergence of HRM (Von Glinow, Drost, & Teagarden, 2002), often following dominant United States (US) HRM practices. However, we lack a large-scale and focused investigation of these effects across countries for expatriate performance management (EPM). Therefore, we collected and analyzed data from 132 MNEs, originated in three country clusters, to answer the questions whether and to what extent EPM converges, diverges, or crossverges in MNEs.
By doing so, our research contributes to the literature in three main ways. First, we further advance the convergence/divergence/ crossvergence debate by addressing an under-researched HRM practicethat is, EPM. Focusing on novel practices is important, as there is strong agreement that the institutional environment affects each HRM practice differently (Farndale et al., 2017;Rosenzweig & Nohria, 1994). Prior research investigating EPM differences across countries has relied either on data of expatriates' experiences with such practices on the individual level (Suutari & Brewster, 2001;Suutari & Tahvanainen, 2002) or small sample sizes/qualitative data (Shih et al., 2005;Tahvanainen, 2000).
Hence, our study is the first large-scale investigation of country differences in EPM of MNEs on firm level, allowing for a systematic analysis of commonalities and differences of EPM practices across country clusters.
Second, we extend research on the debate by addressing if and on what level the differentiation across countries takes place. Prior comparative HRM research has mostly looked at the practice level-that is, single practices or bundles of practices (Doellgast & Marsden, 2019;Farndale et al., 2017). In turn, international human resource management (IHRM) research of MNEs has focused on more strategic questions that affect the performance management of expatriates, such as expatriate staffing strategies (Bebenroth & Froese, 2020;Belderbos & Heijltjes, 2005;Gaur, Delios, & Singh, 2007). Conceptual research suggests EPM to be a means of expatriate strategy execution and thus strategic questions need to be considered when looking at EPM (Fenwick, De Cieri, & Welch, 1999). However, it has been argued that there is a lack of integration of comparative HRM and IHRM research (Brewster, Mayrhofer, & Smale, 2016;Kaufman, 2016), which has also led to a lack of integrative research of the strategic and practice levels of MNEs' HRM practices. Integrating the two streams of research and investigating EPM on the strategic and practice levels, our study develops a more fine-grained understanding of the mechanisms of convergence, divergence, and crossvergence across different levels of HRM practices.
Third, we provide an empirical contribution by developing a comprehensive overview of current EPM practices in three different country clusters. Prior studies on performance management have mainly been conducted with regard to American (Gregersen, Hite, & Black, 1996;Martin & Bartol, 2003), Australian and Singaporean (Fee, McGrath-Champ, & Yang, 2011;Woods, 2003), and Finnish firms (Suutari & Brewster, 2001;Suutari & Tahvanainen, 2002;Tahvanainen, 2000). Our study, in contrast, is based on MNE data from three different country clusters in order to allow for a larger scale comparison. Typically, the US, Germany, and Japan have been used in cross-country comparison studies as their indigenous HRM systems differ greatly and reflect their specific business and cultural histories (Pudelko, 2006). However, HRM practices differ not only across countries but also across market economies (Farndale et al., 2017). The variety of capitalism literature posits that different types of market economies have developed specific interplaying institutions that are more similar within than between market economies (Hall & Soskice, 2001;Jackson & Deeg, 2008;Witt et al., 2018). Similarly, cultural research has shown that some countries have similar values and can be grouped in country clusters (House, Hanges, Javidan, Dorfman, & Gupta, 2004). To develop a more holistic picture, building on the relevance of the indigenous HRM systems of German, Japanese, and US firms, we therefore collected data from MNCs originated in different countries in the respective three types of market economies in our study: the Anglo-Saxon, liberal market economy cluster (US and United Kingdom [UK]) Germanic, coordinated market economy cluster (Germany, Austria, and Switzerland) and Confucian Asian, highly coordinated economy cluster (Japan) (House et al., 2004;Witt et al., 2018). Comparing these allows us to draw novel insights in under-researched contexts regarding EPM in MNEs.

| EPM: Practices
Research on performance management dates back to the early 1920s but significantly grew in the late 1980s (Lindenholm, 2000), now being a mature field (see DeNisi & Murphy, 2017;DeNisi & Smith, 2014 for reviews). Performance management is referred to as the combination of HRM best practices and the extension of the often-criticized concept of performance appraisals (Schleicher et al., 2018). It can be defined as the "continuous process of identifying, measuring, and developing the performance of individuals and teams and aligning performance with the strategic goals of the organization" (Anguinis, 2013, p. 3) and is implemented as a tool to improve individual performance, often based on the assumption that improved individual performance will improve organizational performance as well (DeNisi & Smith, 2014).
More recently, practitioners and academics have coined the performance management system term to highlight the processual character and the relevance of consistency of practices (Schleicher et al., 2018). There are various models for performance management systems, typically related to a process of goal setting and performance appraisal, which is eventually linked to training and development and performance-based pay. These practices are also commonly implemented in MNEs (Shih et al., 2005;Tahvanainen, 2000). More recently, providing continuous feedback above and beyond the appraisal meetings has received increasing attention Kossek, Huang, Piszczek, Fleenor, & Ruderman, 2017).
This research highlights the need to be in close contact with employees and to provide frequent feedback and monitoring to overcome the shortcomings of annual performance appraisal meetings.
Accounting for these developments, we refer to EPM as a process involving practices of goal setting, continuous feedback, performance appraisal, training and development, and performance-based bay and will review the extant literature on those practices in the expatriate context in the following sections.

| Expatriate goal setting
Goal setting is an important aspect of performance management as it provides the baseline for performance measurement and appraisal. Particularly, in EPM, where defining success and outcomes has proven to be difficult (Harrison & Shaffer, 2005), defining and agreeing on goals is essential.
Goal-setting theory (Locke & Locke, 1967), an important theoretical construct underlying performance management, suggests that goals have a motivational aspect and can be used to direct employees' behavior. From this point of view, goal setting is crucial for expatriates to provide clarity about expectations associated with an international assignment (Martin & Bartol, 2003). Accordingly, Suutari and Tahvanainen (2002) found that set performance goals can increase the perceived efficiency of performance management. There is, however, mixed evidence of whether or not goal setting takes place during expatriation (Suutari & Tahvanainen, 2002;Tahvanainen, 2000). Prior research has highlighted that both "hard" goals (e.g., those based on corporate figures) (Fee et al., 2011) and "soft" goals (e.g., leadership) play an important role during assignments (Gregersen et al., 1996;Suutari & Tahvanainen, 2002). Most frequently, goals are set by the expatriate and the host country manager (Shih et al., 2005).

| Expatriate continuous feedback
Not much research has been conducted regarding the specific meaning of feedback for expatriates as it is commonly seen as a part of the performance appraisal process (Cascio, 2006). However, feedback is important as it provides the expatriate an opportunity to improve (Murphy & Cleveland, 1995). It is also argued that frequent feedback can outperform annual performance appraisals and help expatriates to understand where they stand in terms of goal achievement. How feedback is provided differs across organizations and ranges from one-way, top-down to open, trustful feedback Shih et al., 2005). Feedback is most commonly given by the host country manager (Tahvanainen, 2000), but varies across types of assignments. Furthermore, research refers to the implementation of instruments such as 360 feedback systems, which involve various parties to provide feedback to the expatriate (Woods, 2003). Interestingly, research highlights that satisfaction with feedback varies across different types of expatriates and many expatriates claim that they do not receive sufficient feedback (Tahvanainen, 2000).

| Expatriate performance appraisal
Performance appraisal is an important component of performance management, as it encompasses the evaluation of the expatriate's performance and goal achievement. In fact, it is the basic practice and origin of our understanding of managing performance (Lindenholm, 2000). As in many domestic companies, MNEs have started including instruments, such as multisource or 360 feedback and rating that help to improve the quality of feedback and objectivity of ratings (Kossek et al., 2017).
Such feedback systems include different actors above and beyond the line manager. Gregersen et al. (1996) suggested that up to 10 people can potentially be involved in the process of rating expatriates, with an average of 2.7 raters participating in appraisals. Furthermore, research suggests appraisals are mostly carried out by supervisors, either from the foreign or from the home country, and less frequently by other executives or managers (Suutari & Brewster, 2001). Most frequently, the supervisor in the host country is responsible for the appraisal, particularly when the expatriate stays longer term and has a managerial role (Suutari & Tahvanainen, 2002), although multisource ratings are also increasingly being implemented (Kossek et al., 2017). Annual appraisals used to be the most common approach (Gregersen et al., 1996), with organizations now moving to bi-annual appraisals (Shih et al., 2005) or more frequent assessments.
In terms of the specificities of expatriate performance appraisals, research indicates a few other interesting findings. For instance, there is evidence that the location of the supervisor and the involvement of the expatriate in the appraisal process do not increase the perceived effectiveness of the performance management system (Suutari & Tahvanainen, 2002). Furthermore, the effect of frequency of evaluation is inconsistent across studies. While some suggest higher frequency to increase efficiency and accuracy (Suutari & Tahvanainen, 2002), others do not support this notion (Martin & Bartol, 2003).

| Expatriate training and development
Performance appraisals are often used as the basis for making training and promotion decisions. If designed correctly, they allow for a relatively objective way to determine employees' needs and rewards, and thus are an important instrument for leading and developing employees (Anguinis, 2013). Looking at training related to expatriation, research has frequently highlighted the importance of predeparture and cross-cultural training for expatriate success.
Furthermore, assigning mentors to support expatriates is believed to improve performance (Tung & Varma, 2008). Yet, only 28% of companies relate performance appraisal to training during the assignment (Martin & Bartol, 2003); furthermore, expatriates experience a lack of training and mentoring abroad (Fee et al., 2011;Schuster, Ambrosius, & Bader, 2017;Shih et al., 2005). This indicates a decoupling of the performance appraisal process and training. In turn, it is highlighted that the relevance of performance management for career development has a positive influence on perceived performance management success; additionally, the vast majority of US companies base promotion decisions on appraisal results (Martin & Bartol, 2003).

| Expatriate performance-based pay
Research on performance-based pay thus far is univocal. It shows a great majority of expatriates are on performance-based bonus systems linked to the outcome of the performance appraisal (Gregersen et al., 1996). Likewise, in Martin and Bartol's (2003) study, 81% of companies used appraisal results to determine performance, 47% for bonus allocation, and 9% for profit sharing, indicating a strong connection between appraisal results and pay. Both studies, however, were conducted in the US context, thus mirroring the importance of performance-based pay there (Pudelko, 2006). Shih et al. (2005) also indicate that a clear link between performance and total compensation exists yet find one important exception: in Japanese organizations, the link is not clear to expatriates, as pay is mostly based on seniority.

| Strategic level
Both domestic and international literature suggests that there is a strategic level to EPM (Anguinis, 2013;Fenwick et al., 1999). Interestingly, these strategic considerations have not received much attention in EPM literature since the main focus so far has been on EPM practices (Gregersen et al., 1996;Suutari & Brewster, 2001). One EPM aspect that needs particular attention is staffing strategy, as managing expatriates' performance relates to more general questions of control and performance in MNEs (Fenwick et al., 1999). Another, more strategic aspect to consider is the specific overarching design elements of EPM. Prior research indicates performance management of expatriates is different from performance management of domestic employees (Fenwick et al., 1999;Gregersen et al., 1996;Harrison & Shaffer, 2005). It is therefore important to consider the specificities of expatriation when designing an EPM system. We will review the expatriate staffing strategy literature, including the consideration of the local environment in EPM and the specific design of EPM for expatriates, in more detail below.

| Expatriate staffing strategy
It is long evident that companies differ in the degree as to which they rely on expatriates to manage and control their operations (Perlmutter, 1969). Consequently, the expatriate staffing strategy links to goals and performance expectations and should thus be considered a part of the EPM system (Fenwick et al., 1999). Fenwick et al. (1999) linked the expatriate staffing strategy of an MNE with EPM and referred to expatriates as a way of cultural control as opposed to bureaucratic control through systems standardization. Typically, IHRM research has investigated whether MNEs from different countries of origin vary along their types of control and, relatedly, their expatriate staffing strategies (Ferner et al., 2001;Gaur et al., 2007;Pudelko & Tenzer, 2013). For instance, Anglo-Saxon MNEs have been shown to prefer bureaucratic rather than social control of their foreign subsidiaries and implement formal, worldwide policies for performance appraisals and remuneration of managers (Ferner, 2000;Ferner & Quintanilla, 1998). Furthermore, they tend to employ tight control through formal budget-setting and review processes (Edwards, Ferner, & Sisson, 1996;Ferner et al., 2001) and thus prefer formal over social control. Consequently, it has been concluded that Anglo-Saxon MNEs are less ethnocentric than Germanic and Japanese MNEs and rely less on assigning expatriates to ensure control and standards in their foreign subsidiaries.

| Specific design of performance management for expatriates
While it has been argued that EPM should differ from local employee performance management (Fenwick et al., 1999;Martin & Bartol, 2003;Tahvanainen, 2000), not much research has been conducted to address whether and how it differs on more strategic design elements. Rather, prior research indicates EPM is often a duplicate of performance management in the organizational headquarters without much adaptation for expatriates. Initial evidence suggests despite the specifics of expatriate assignments, most MNEs have not developed performance management structures and appraisal forms specifically for expatriates; instead they use one standardized system for all employees (Shih et al., 2005). This has been shown to be negatively correlated with the accuracy of the expatriate performance assessment (Gregersen et al., 1996). However, we do not know much about how it is designed yet and to which extend firms develop specific policies and systems for expatriates.

| Local environment consideration in EPM
Adding to the question of the overall design of EPM, Martin and Bartol (2003) suggested that performance appraisals need to account for local conditions and thus the particular challenges expatriates face abroad (Black, Gregersen, Mendenhall, & Stroh, 1999). For instance, it is a well-known fact that expatriates face specific challenges in their host countries such as their challenges to establish effective working relationships with host country nationals (Bader, 2017;Bader, Froese, Achteresch, & Behrens, 2017;Toh & DeNisi, 2007; and therefore the extended time they need to adjust to the host country before they are able to fully perform (Logger & Vinke, 1995)-a consideration that should be accounted for in EPM. Research suggests it takes up to 6 months for expatriates to adjust to the local work environment (Logger & Vinke, 1995), varying based on the cultural differences between the home and host countries. Therefore, Martin and Bartol (2003) argued that "expatriate performance appraisal systems encompass many issues not normally addressed by domestic systems (…) related to divergent cultures, legal and political factors, different criteria, and varying environments" (p. 117). Hence, there is a need to account for these specificities throughout the process. Figure 1 provides an overview of the conceptual model of the EPM system used in our study. In the following sections, we will develop hypotheses regarding the convergence, divergence, and crossvergence of the strategic and practice levels of the EPM system. Researchers advocating the convergence view argue on the basis of the universalist paradigm and suggest the process of industrialization and the spread of advanced technology result in globalization, moving all countries toward political and economic systems similar to those of the US, which has been dominating the world economy (Kerr, Dunlop, Harbison, & Myers, 1960;Rowley & Benson, 2002). It is argued that MNEs in particular spread managerial knowledge and techniques internationally and fuel the dissemination of best practices (Brewster et al., 2016). This is because MNEs are forced to adopt common strategies and practices to compete with other MNEs in the global market (Quintanilla & Ferner, 2003). Originally proposed in the 1960s (cf. Kerr et al., 1960), convergence researchers identify the Anglo-Saxon HRM model as the global best practice, which has caused global convergence and a dominance effect (Pudelko & Harzing, 2007). Consequently, researchers have argued that MNEs adopt Anglo-Saxon HRM practices on a global scale, which is referred to as "Anglo-Saxonization" (Ferner et al., 2001;Ferner & Quintanilla, 1998). The global spread of performance management is one example of this process.

| Convergence, divergence, and crossvergence in HRM
The divergence view, in turn, is based on the contextualist paradigm and suggests HRM practices from different countries of origin vary significantly. This paradigm is based on institutional (Kostova, 1999;North, 1990) and neo-institutional theory (DiMaggio & Powell, 2000) arguing for the powerful and constraining impact of institutions on organizations. This is particularly relevant for HRM practices, as they are deeply embedded in legal, political, economic, and sociocultural contexts and national business systems (Ferner, 2000;Ferner, Almond, & Colling, 2005;Scott, 1995). Consequently, research has frequently highlighted differences in HRM practices across countries (Doellgast & Marsden, 2019), but also depending on the country of origin of an MNE (Ferner, 1997(Ferner, , 2000, and on the types of market economies (Farndale et al., 2017). The variety of capitalism literature (Hall & Soskice, 2001;Witt et al., 2018) refers to complex interactions of institutions that have developed in similar ways in different types of capitalism that shape similarities (convergence) within types of capitalism and dissimilarities (divergence) across them.
Trends of crossvergence have also been observed. Crossvergence refers to perspectives that balance the convergence-divergence debate and indicate that a combination of factors affects HRM practices (Al Ariss & Sidani, 2016) where, for example, Anglo-Saxon best practices are adopted, but they are adjusted to the home country requirements (Ferner et al., 2001;Ferner & Quintanilla, 1998).
Accordingly, there are opposing views and inconsistent findings as to the extent of convergence versus divergence. Examples of explanations for these inconsistencies have been found in the degree to which particular HRM practices are contingent upon the institutional F I G U R E 1 Expatriate performance management system: strategic and practice levels environment (Farndale et al., 2017;Rosenzweig & Nohria, 1994) and in the lack of integration of IHRM and comparative research (Brewster et al., 2016;Kaufman, 2016). In the following, we extend these findings by developing hypotheses regarding the convergence, divergence, or crossvergence of EPM in MNEs, accounting for the strategic and practice levels.

| Strategic level of EPM
Due to the strong connection between strategic decisions and core values, we suggest strategic level variables diverge across countries.
Supporting this notion, research has frequently highlighted that MNEs from different countries differ tremendously in terms of the strategic importance of expatriates. To which extent MNEs rely on expatriates being sent is a core decision that has been related to managerial and cultural values (Perlmutter, 1969). Consequently, it was highlighted that US and UK MNEs rely less on expatriates and are more likely to employ a polycentric (i.e., staff their key positions with host country nationals) rather than an ethnocentric policy (i.e., staff their key positions with expatriates) to manage their international operations (Ferner & Quintanilla, 1998). In turn, it is argued that Germanic MNEs are more reliant on long-term strategic planning, coupled with the use of informal personal control and feedback methods (Ferner et al., 2001). Therefore, they tend to use more ethnocentric staffing policies and prefer social over bureaucratic control (Ferner et al., 2001). Japanese MNEs tend to globally integrate and export their management philosophies and HRM practices, such as long-term orientation and teamwork, to their foreign subsidiaries (Beechler & Yang, 1994;Grill, Maharjan, & Sekiguchi, 2016), as well as strongly pursue an ethnocentric staffing policy Gaur et al., 2007;Oki, 2020;Sekiguchi, Froese, & Iguchi, 2016). In this approach, decision-making is centralized at headquarters, and key positions at home and in subsidiaries tend to be filled with expatriates from Japan. Therefore, Ferner (1997, p. 27) argues that Japanese MNEs "rely much less than Anglo-Saxon MNEs on formal systems and more on faceto-face informal assessmentone reason that they are so expatriateintensive." In fact, Japanese MNEs have been shown to have much higher expatriate rates than German ones (Tungli & Peiperl, 2009). Furthermore, we expect there are country differences regarding the other strategic level aspects, as these are core decisions that shape the overall EPM system. For instance, as Japanese MNEs heavily rely on expatriates, they focus on sending and managing high numbers of expatriates abroad. Furthermore, due to their predominant ethnocentrism, Japanese MNEs place particularly high importance on assigning expatriates to ensure control and standards in their foreign subsidiaries (Caligiuri & Stroh, 1995). Consequently, they need to design a specific performance management system that accounts for their critical role in subsidiary control. Anglo-Saxon MNEs, in turn, are likely to have standardized systems of performance management and rely on polycentric staffing (Ferner & Quintanilla, 1998;Kopp, 1994). Therefore, they are less likely to place emphasis on an expatriate-specific design. Finally, Germanic MNEs typically choose an ethnocentric approach as well, although they rely less on expatriates than Japanese MNEs but more than Anglo-Saxon ones (Tungli & Peiperl, 2009). Therefore, it is reasonable to conclude that they differ in their approach to EPM design as well. Building on these considerations, we hypothesize: Hypothesis 1. EPM diverges on the strategic level across the three country clusters.

| Practice level of EPM
Historically, performance management was developed in the Anglo-Saxon institutional environment and focuses on linking individual performance with pay and rewards. This has been highlighted to reflect the Anglo-Saxon culture of individualism and performance orientation (Pudelko, 2006). Due to the dominance of Anglo-Saxon HRM (Pudelko & Harzing, 2007), companies across the globe started to follow these best practices in the 1990s and early 2000s and, even in Japan, many companies began implementing performance-based HRM practices (Sekiguchi, 2013). However, as Japan's HRM system has a longstanding tradition of seniority-based wages, life-time employment, and company-based unions , over time, there is evidence of implementation of these practices, yet with a country-specific focus. In consequence, novel performance management practices emerge where HRM practices emphasize performance at work but do not link it to pay and promotion (Nakamura, 2006;Shih et al., 2005). Similarly, Festing et al. (2012) suggested in a comparative study that performance management in MNEs shares some practices but develops a country-specific profile overall. To explain such crossvergence, the meaning of institutionalization and adjustment of practices over time is relevant (Oliver, 1991). Sekiguchi (2013) exemplified that performance management was a typical management fashion in Japan that got institutionalized but adjusted to the local context over time. Accordingly, some aspects, such as goal setting, seem to have developed as best practices while others, such as performance-based pay, seem to be contradictory to the core values in the country. Consequently, we expect that specific profiles have emerged that share some, but are different regarding other, practices. We therefore propose: Hypothesis 2. EPM crossverges on the practice level and countries develop specific profiles that share some of the practices, but differ regarding others.

| Data collection and sample
We collected data from three country clusters (Witt et al., 2018): Anglo-Saxon, liberal market economies (UK and US), Germanic, coordinated market economies (Germany, Austria, and German-speaking parts of Switzerland), and Asian, highly coordinated economy (Japan). We chose those clusters following reasoning in previous studies on IHRM (Pudelko & Harzing, 2007;Tungli & Peiperl, 2009) that the country clusters have significantly different cultures (Hofstede, 2001) and business systems (Witt et al., 2018). We approached MNEs, as they are likely to make use of more sophisticated HRM practices (Tungli & Peiperl, 2009).
We collected our data via an online questionnaire. The original questionnaire was developed in English. We then translated it into German and Japanese using the translation-back translation method to ensure translation equivalence (Brislin, 1976). Respondents could choose to fill out the survey in English or their home country language.
Since we were interested in EPM, and because respondents needed to have oversight of the worldwide EPM process, we invited respondents working in the global mobility or IHRM departments of the respective MNEs. As there is no publicly available directory of such people, we had to identify potential respondents via an intensive screening of personal and social networks. Overall, we invited 523 companies in the Anglo-Saxon, 484 in the Japanese, and 796 in the Germanic cluster.
The response rate was 7%. The final sample consists of 132 MNEs (40 Anglo-Saxon, 42 Germanic, and 50 Japanese), which is comparable with prior research on expatriate management in MNEs (Tungli & Peiperl, 2009). Participating MNEs employed, on average, 1,235 expatriates and 36% of MNEs were manufacturing companies.

| Measures
If not indicated otherwise, we used measures from existing literature, having respondents answer on a 7-point Likert scale. If necessary, we adapted the text to the expatriation context.

| Strategic level
To collect data about the expatriate staffing strategy, we presented four ordinal scale items that we developed based on prior literature (Caligiuri & Stroh, 1995). Each staffing strategy (ethnocentric, regiocentric, polycentric, and geocentric) formed a category of this variable and was complemented by a short description. We then dummycoded whether the MNE applied an ethnocentric strategy (1) or not (0).
To capture the specific design of EPM for expatriates, we asked respondents to indicate whether performance management of expatriates was different from performance management of other employees in their organization and to indicate whether performance management is "different and specifically designed for expatriates," "the same as for all employees worldwide, regardless of expatriate status," "largely the same for all employees worldwide with minor adjustments for expatriates," "expatriates are treated the same as employees in the host country, but different from employees in the headquarters," or "expatriates are treated the same as employees in the headquarters, but different from employees in the host countries." As our conceptual model refers to the question of whether EPM is specifically designed for expatriates or not, we dummy-coded this item, indicating whether EPM was specifically designed for expatriates (1) or not (0).
To further scrutinize how it was adjusted to expatriate-specific issues and local context, we measured local environment consideration using five items taken from Martin and Bartol (2003), asking for the degree of agreement to the statements. A sample item is "additional time that may be required to complete a task/function in a foreign environment is considered in evaluating the job performance of an expatriate." Cronbach's alpha was .78.

| Practice level
In order to determine EPM practices, we built on our literature review and measured five practices most frequently referred to as goal setting, continuous feedback, performance appraisal, training and development, and performance-based pay (Anguinis, 2013;Cascio, 2006;Gregersen et al., 1996;Locke & Locke, 1967). For each of these practices, respondents were asked to indicate how important the respective practice was in their MNE.
Given the centrality of feedback in prior EPM literature, we added a further measure to gain deeper insight. Therefore, we asked how EPM feedback was communicated and applied a five-item measure taken from Festing et al. (2015), which was slightly reworded to fit the expatriate context. A sample item is "the feedback situation is open and truthful." Cronbach's alpha was .89.
As performance appraisal is the main component of performance management (Lindenholm, 2000), we asked for criteria used and frequency of appraisals. These criteria used for performance appraisal were incorporated by including seven different criteria taken from (Festing et al., 2015). In particular, we asked whether previous appraisal results, expertise, educational background, access to networks and people, empowerment of others, effective teamwork, and result targets met were included. For each criterion, respondents were asked to indicate how important it is in terms of the performance appraisal score. Following Martin and Bartol (2003), to determine frequency of expatriates' performance appraised, we asked respondents how often an appraisal happens, giving them Likert-type options from 1 = less than annually, 2 = annually, 3 = semi-annually, 4 = quarterly, and 5 = more frequently.

| Success indicators
While we did not hypothesize about the success of the EPM system, we did collect data to determine whether the EPM system was considered successful. Therefore, we applied the three-item measure developed by Martin and Bartol (2003) capturing perceived success of expatriate performance management. Respondents were asked to respond on a 7-point Likert scale. A sample item is "our expatriate performance management system is helping to motivate expatriates." Cronbach's alpha is .95. In addition to perceived success, we collected data to estimate the objective success by asking for failure rates of expatriates. Because there is no general agreement about what assignment failure is (for an overview, see Harzing, 2002), we asked respondents to indicate (a) the percentage of expatriates who perform below expectations during their assignment and (b) the percentage of expatriates who return prematurely from their assignment without having completed the job abroad. Following previous research (Tung, 1982;Tungli & Peiperl, 2009), we then categorized these values into low (<10% = 1), medium (≥10% but <20% = 2), and high (≥20% = 3).

| RESULTS
In the following, we present the results of our analysis. To test for similarities and differences, we conducted a series of analysis of covariance (ANCOVA), including our covariates. Moreover, in each step, we applied Levene's test for homogeneity of variances. When the assumption of homogeneity was met, we opted for a Tukey test in our post hoc analysis; when it was violated, we conducted a Games-Howell test, which does not rely on the assumption of homogeneity of variances. When both dependent and independent variables were ordinally scaled, we used a chi-square difference test. Table 1 shows the correlation of the main variables.

| Strategy of EPM
Hypothesis 1 proposed that EPM varies on the strategic level across the three country clusters. Indeed, as outlined in Table 2, there are significant differences between staffing strategies of MNEs, with Japanese companies significantly more often opting for an ethnocentric approach and Anglo-Saxon companies doing so less frequently (X 2 [2, N = 132] = 26.08, p < .001). In a similar vein, we compared whether the performance management system was specifically tailored to expatriates and observed a similar pattern, highlighted in Table 2 (X 2 [2, N = 132] = 56.81, p < .001). While Anglo-Saxon and Germanic MNEs use it less than expected, the vast majority of Japanese MNEs implements an expatriate-specific performance management system. Overall, these findings partially support Hypothesis 1 as we see a clear and distinct profile of Japanese MNEs (higher ethnocentric staffing, higher expatriate specific HRM and higher local environment consideration). However, Anglo-Saxon MNEs are in between Japanese and Germanic MNEs in terms of the local environment considerations but more similar to the Germanic MNEs in terms of the lack of expatriate specific design of EPM.

| Practice level
To test Hypothesis 2, we focus on the importance of five distinct performance management-related HR practices and compare them across the three country clusters. Results are displayed in Table 4.
Our data show that the MNEs from the three country clusters did not differ significantly regarding the importance of goal setting (F [2, 123] = 3.024, p > .05). Regarding the importance of continuous feedback, we did find significant differences (F [2, 123] = 3.448, p < .05).
Post hoc analyses revealed that Japanese MNEs make more use of it compared with Germanic ones. As Table 4 depicts, feedback communication was important in all MNEs in our sample, without any significant differences across countries. Subsequently, we investigated differences in the importance of performance appraisals. Indeed, there are significant differences (F [2, 123] = 3.143, p < .05), in particular between Japanese and Germanic MNEs, indicating that Japanese MNEs place more importance on performance appraisals. Going into more detail, we analyzed the number of actors involved in the process and distinguished between home and host country actors. As Table 4 illustrates, the number of actors does not significantly differ. When looking at the frequency of appraisals, we do find significant differences (F [2, 123] = 3.496, p < .05). Japanese MNEs conduct appraisals significantly more often than Germanic ones. Finally, we found interesting differences in terms of the criteria that were used to evaluate performance. Festing et al. (2014;Festing et al., 2015) indicated that these criteria can be grouped into input (previous appraisal results, expertise, educational background, access to networks and people) and output (empowerment of others, effective teamwork, and result targets met) criteria. Regarding input criteria, previous appraisal results and access to networks and people were not significantly dif- Eventually, looking at the meaning of EPM for training and development and performance-based pay as HR practices, we found signifi- p < .10). Post hoc analyses revealed significant differences between Japanese and Anglo-Saxon MNEs, indicating that Anglo-Saxon MNEs more strongly relate appraisals to training and development and performance-based pay; albeit, we need to point out that the differences in performance-based pay should be interpreted with caution.
These findings support Hypothesis 2 in so far as MNEs from all countries share some practices (goal setting, feedback communication) but differ in others and develop country-specific profiles (crossverge). with those of Anglo-Saxon. Table 5 summarizes our results.

| DISCUSSION
Our findings suggest, at the strategic and practice levels of EPM, there is divergence across countries. In particular, Japanese MNEs are different from Anglo-Saxon and Germanic MNEs. That is, because Japanese MNEs tend to pursue ethnocentric staffing strategies Pudelko & Tenzer, 2013) and use expatriate-specific performance management with higher local environment consideration to achieve social control of their foreign subsidiaries with Asian-style leadership, the practice level of their EPM tends to focus on the monitoring of expatriates' roles and behaviors (e.g., continuous feedback, frequent performance appraisal) rather than enhancing their employees' careers through expatriation (Tungli & Peiperl, 2009, p. 166 (Martin & Bartol, 2003). Yet, there were no clear differences between Anglo-Saxon and Germanic MNEs at the strategic and practice levels of EPM. Indeed, although aspects of EPM of Germanic and Anglo-Saxon MNEs were different from Japanese MNEs, our findings suggest the EPM of Anglo-Saxon and Germanic MNEs is more similar than different, as we did not find statistically significant differences.
Second, we also found some practices to be similar across the three country clusters, highlighting trends of crossvergence. In particular, the importance of goal setting, feedback communication, and the number of actors in the appraisal process was similar in all three country clusters. Our results indicate that these have evolved as global best practices that are applied irrespective of cultural and institutional differences (Pudelko & Harzing, 2007).
Finally, the evaluations of the EPM success were similar among Anglo-Saxon, Germanic, and Japanese MNEs. It seems that no system is superior to the other. However, despite rather low premature return and underperformance rates, the self-evaluations of success were not particularly positive, which suggests there is room for improvement of EPM in all three country clusters. Table 6 provides an overview on the main results of our analysis.

| Theoretical contributions
Our study makes several contributions to the literature. First, our findings advance the understanding of convergence-divergencecrossvergence of HRM practices by detailing current trends in EPM.
Looking at particular practices is important because each practice is affected differently by the institutional context and some are more, others less, constrained (Farndale et al., 2017;Rosenzweig & Nohria, 1994 While we observed divergence in EPM at the strategic level, we saw both convergence and divergence at the practice level, pointing toward crossvergence. MNEs tend to adopt global best practices in order to compete effectively in the global market (Pudelko & Harzing, 2007;Quintanilla & Ferner, 2003) but seem to do so while retaining country-specific profiles in accordance with their home country context. However, they seem less willing or able to modify the strategy level as these are more strongly related to core MNE decision-making and values. Our data show that, in particular, Japanese MNEs keep a distinct profile, whereas the Western institutions seem to differ less on that level. We will discuss this finding in more detail below when looking at the specifics of country profiles. Taken together, our findings suggest IHRM (Doellgast & Marsden, 2019;Farndale et al., 2017) (Chowhan, 2016). Because of this fit, there were no significant differences in the success of EPM. In other words, each EPM system seems to be suitable for the respective home country background.
Third, we extend the country context of EPM studies by conducting a large-scale empirical study of MNEs from three different regions, whereas most prior research has been limited to US (Gregersen et al., 1996;Martin & Bartol, 2003) and Finnish MNEs (Suutari & Brewster, 2001;Suutari & Tahvanainen, 2002;Tahvanainen, 2000). Although performance management has evolved in the US context, and, even though there is a pressure of convergence for MNEs (Pudelko & Harzing, 2007) (Ferner, 1997;Ferner et al., 2001;Froese et al., 2020). Accordingly, due to their ethnocentric staffing strategy, Japanese MNEs apply expatriate-specific performance management that focuses on frequent evaluation and continuous feedback. In contrast, Japanese MNEs may neither want nor be able to successfully implement stricter performance control mechanisms, such as performance-based pay, as their management style typically echoes the importance of seniority (Pudelko, 2006), which is deeply rooted in their national and cultural characteristics, and HRM is largely path-dependent Sekiguchi et al., 2016). This shows how pressures of divergence (the institutional context of the home country) shape performance management systems toward alignment of strategies and practices, which leads to consistent country profiles . However, as we did not find statistically significant differences of Germanic and Anglo-Saxon companies, our research indicates that the role of institutional and sociocultural contexts (DiMaggio & Powell, 2000;Festing, 2012;Scott, 1995) differs and is weaker within Western countries, leading somewhat to convergence to a Western model of EPM. This is in line with prior research that highlighted the influence of Anglo-Saxon practices on German companies (Ferner et al., 1998). In contrast, differences are stronger between Western and Asian countries (e.g., Anglo-Saxon countries/ Germanic countries and Japan) leading to divergence and a distinct Japanese profile.

| Managerial implications
Overall, our study provides important implications for practitioners.
First, our study offers an up-to-date and comprehensive overview on EPM in MNEs. Not only did we include EPM practices but also more strategic aspects of the overall EPM system, which highlights a new aspect of EPM that has not received much attention.

| Limitations and avenues for future research
Our results need to be interpreted in the light of the study's limitations. First, despite the considerable sample size in our study, future research could collect larger data sets, in particular covering more country clusters. However, as the number of MNEs large enough to implement EPM is limited, and prior research on expatriate management practices used a similar sample size (Tungli & Peiperl, 2009), we see this as an acceptable limitation. Although there is sufficient statistical power to derive significant conclusions, it would still be worthwhile to replicate our findings in a larger sample. Second, due to the cross-sectional nature of our data, we cannot draw final conclusions about convergence or divergence of EPM, but focus on the similarities and differences as an outcome. As EPM is a rather young HRM practice, future research can develop longitudinal designs to capture and describe trends of convergence, divergence, and crossvergence over time. Third, we included companies from three developed country clusters. Therefore, identified convergence trends cannot be generalized globally, particularly when looking at the increasing importance of Emerging Market Multinationals (Held & Bader, 2018). International career success is a rather complex and country-dependent phenomenon (Breitenmoser, Bader, & Berg, 2018). Therefore, the measurement of this, that is, managing expatriate performance, needs to account for these differences and could also include the repatriation phase (Breitenmoser & Bader, 2016;Breitenmoser & Bader, 2020;Chiang, van Esch, & Birtch, 2020). Other large and important nations, such as China or India, have significant international assignment activity as well, yet may have unique performance appraisal systems. For instance, state-owned (Chinese) enterprises may adapt foreign policies to overcome distrust (Meyer, Ding, Li, & Zhang, 2014), which may shape the convergence-divergence debate. Moreover, the level of institutional discrimination of expatriates in certain countries plays an important role as well (Bader, Stoermer, Bader, & Schuster, 2018), which may impact what EPM practices may or may not converge in certain countries. Therefore, our choice of country clusters can only be considered as a starting point for further investigation. Future studies could replicate our study with MNEs from more country clusters and other business systems (Witt et al., 2018) and also account for the different levels of economic development, that is, MNEs from Brazil, China, India, or Mexico. In this regard, future research could also try to identify "practice profiles"-clusters or systems of practices and how these differ (or are similar) across countries. Lastly, our data are derived from one source only. As we obtained mostly objective measures about company policies and practices and compared EPM across MNEs from different countries, this concern is mitigated. However, with regard to outcomes of EPM, it would be helpful to incorporate the view of other stakeholders as well, such as expatriates themselves, actors conducting the performance appraisal, general HR managers, and executives in the headquarters and subsidiaries. That way, we could derive a much more holistic picture. Furthermore, a longitudinal approach would be promising.
To understand EPM in more depth, future research could consider the whole cycle of expatriation, considering items such as goal setting and training prior to the expatriate assignment and/or feedback and career development after the expatriation assignment. Furthermore, future research could look into the relationship between particular EPM practices and expatriation success and see how country context alters these relationships. This could also take into account country profiles and systems of practices in more detail. For example, research can investigate whether the EPM practices of MNEs which adopt the typical profile of a country cluster are more successful than others. Another potential avenue for future research could be to investigate how MNEs could successfully implement changes in their EPM. To do so, we would recommend longitudinal studies that measure expatriate performance before and after the change. Despite these limitations, our study has provided intriguing insights into crosscultural EPM and hopefully inspires further research to increase our understanding of EPM.

DATA AVAILABILITY STATEMENT
The data that support the findings of this study are available upon reasonable request from the corresponding author. The data are not publicly available due to privacy restrictions.