Does It Pay Off? Understanding Subjective Employment Mobility of European Physicians in Germany

This article investigates the subjective employment mobility, defined as migrants’ evaluation of their employment situation before and after migration, of European physicians in Germany. Analyzing different dimensions of occupation (e.g., income, working conditions, use of skills, career opportunities) of physicians who migrated to Germany from within the European Union (i.e., EU physicians), it examines which factors influence physicians’ perception of whether migration worsened or improved their employment situation. I argue that the original reasons to migrate (e.g., economic, career-related, or family reasons) and other migration-related factors (e.g., language skills), as well as characteristics of the occupation (e.g., the hierarchical structure), must be considered to understand subjective employment mobility. The analyses are based on original survey data collected among EU physicians in Germany (N = 1,058). Results from OLS regressions show that physicians’ original reasons for migration largely matched their subjective employment mobility, suggesting that migration for career reasons and a perceived improvement of use of skills and career opportunities are positively linked while migration for economic reasons positively affected physicians’ perception of income and working conditions. Physicians aiming for the highest position perceived their overall employment situation as worse compared to before migration, and the origin region mattered, particularly for physicians from EU Eastern member-states, who were more likely to perceive an improvement in their employment situation. Results further inform understandings of labor-related migration of high-skilled professionals by identifying obstacles and conducive conditions at migration for a group that is often assumed not to face barriers in using migration for professional advancement.


Introduction
The European Union (EU) grants its citizens free movement rights, enabling them to be mobile beyond national borders without formal restrictions, and actively encourages mobility by adopting policies to enhance intra-EU migration, especially for education and workrelated migration (e.g., Art. 20, Treaty on the Functioning of the European Union, European Union 2012; EU Regulation No 492/2011, European Parliament and Council of the European Union 2011). EU mobility grants, the Erasmus program, and the European Job Mobility Portal (EURES) are only a few examples of how free movement within the European Union is promoted and facilitated (see also EU Regulation No 2016/589, European Parliament and Council of the European Union 2016). The implicit promise behind free mobility within the European Union is the development of an appropriately skilled, fair, and competitive regional labor market that benefits EU citizens and enables them to improve their occupational situation through mobility (European Commission 2014). In particular, high-skilled migrants are assumed to move more freely within the European Union and to face fewer barriers to working across EU member-states (Favell and Recchi 2011, 60).
Migrants' motivations to move within the European Union reflect the importance of labor-related mobility and show how opportunities for work-related mobility are, indeed, used by EU citizens (Bartolini, Gropas and Triandafyllidou 2017;Castro-Martín and Cortina 2015;Favell 2008). Despite increased diversity in reasons for intra-EU migration (Santacreu, Baldoni and Albert 2009), economic and career-related migration reasons, such as better job prospects or hopes for upward social mobility, remain significant, especially for high-skilled migrants (Favell 2008). However, the question of whether intra-EU migrants do, indeed, improve their occupational situations through migration still stands.
Previous research on the effects of geographic mobility on labor-related mobility often compares migrants' last position in the origin country to the first position in the destination country and applies mainly occupational or class schemes as an analytical tool (e.g., Chiswick, Lee and Miller 2003;Simón, Ramos and Sanromá 2014). Despite its importance, this approach lacks two main elements. First, it often analyzes only the change of job position or income and, thus, does not account for other dimensions of the professional setting which might be (more) important for migrant professionals, such as working conditions or career prospects. Second, it does not consider migrants' subjective evaluations of their own employment mobility.
With respect to the first main element, I argue that in addition to taking the (potential) change of migrants' occupational position into account, it is crucial to analyze different dimensions of the occupational situation, such as income, working conditions, use of skills, and career prospects, to fully understand labor-related mobility. These dimensions, especially working conditions, use of skills, and career prospects, are difficult to quantitatively measure on the individual level and have, thus, not been sufficiently scrutinized in past research on migrants' labor-related mobility (e.g., Attias-Donfut and Dimova 2011;Chiswick, Lee and Miller 2005;Stanek and Ramos 2013). Analyzing different dimensions of the employment situation, however, allows us to grasp the complexity of labor-related mobility, which entails more than just change in a job position. Furthermore, a migrant might, for example, evaluate their situation regarding income as better in the destination country but their career opportunities as worse, highlighting the need to analyze different dimensions of migrants' labor-related mobility separately. This complexity of labor-related mobility also becomes apparent when looking at why people migrate (Becker and Teney 2020;Klein 2016). It is, furthermore, important to consider the profession-specific restrictions and opportunities, as well as the organization of the professional context that shapes labor-related mobility differently for different professions (Bauder 2015;Iredale 2001;Nohl et al. 2014). Thus, we must zoom-in on particular occupational groups to better understand the dynamics of labor-related mobility through migration and to go beyond analyses of the broad categories of occupational change examined before (e.g., occupational position, income).
With respect to the second main element that previous research lacks and in line with migration scholars' focus on subjective perspectives on migration experiences (e.g., Favell, Feldblum and Smith 2007), I argue that migrants' subjective perspectives on labor-related mobility (i.e., how they evaluate change in their employment situation through migration) should be taken into account by researchers interested in migrants' employment mobility. Migrants might, for example, have achieved, on paper, a leading position but, nevertheless, perceive their career advancement as slower than in their origin country (Becker 2019). Such assessment of their employment mobility might influence migrants' intentions to stay, move further, or return to their origin country. 1 The importance of subjective perceptions to 1 Explaining return migration is complex (de Haas, Fokkema, and Fassi Fihri 2015). Previous research on return migration has shown how it can be influenced by a variety of factors, such as social ties and socio-cultural integration (Snel, Faber, and Engbersen 2015), not finding a secure job position (Labrianidis and Vogiatzis 2013), or migrants' subjective perspectives on their situations (Schiele 2021) or evaluation of their personal work environment (Cieslik 2011). Although these studies do not analyze subjective employment mobility's effects on return migration, we could assume that subjective assessments of migrants' employment situation might help shape intentions to stay, return, or migrate further. more comprehensively understand employment mobility is supported, on the one hand, by research on subjective social mobility in general, which argues that a person's positioning in a social class is not determined solely by objective factors such as occupational position (e.g., Attias-Donfut and Wolff 2001;Heath, de Graaf and Li 2010). On the other hand, scholars focusing on migrants' "subjective understandings of occupational mobility" emphasize the importance of analyzing experiences of career opportunities and success to understand migrants' overall occupational situation (Moroşanu et al. 2021(Moroşanu et al. , 1789. Thus, the implications of geographic mobility within the European Union for employment mobility can only be fully understood if migrants' subjective evaluations are considered. This article aims to contribute to these lines of research by investigating migrants' subjective employment mobility, understood as their perception of the change in their employment situation before and after migration. More precisely, I analyze different dimensions of labor-related mobility and consider subjective evaluations of the change in migrants' professional situation through migration. In addition to an empirical assessment of subjective employment mobility, I offer an analytical framework to examine subjective employment mobility more broadly. Such an analytical approach is, so far, missing, as previous research on the relationship between geographic and labor-related mobility has mainly focused on objective employment mobility (e.g., Chiswick, Lee and Miller 2003;Simón, Ramos and Sanromá 2014). Empirically, I focus on one professional group of intra-EU migrants in one destination country: EU medical doctors in Germany. These migrant professionals are an intriguing example of high-skilled, intra-European migration, as the medical profession has one of the highest labor demands within the European Union (van der Ende, Walsh and Ziminiene 2014; for Germany: Bundesärztekammer 2017) and physicians' migration has implications for population health and health care provision, particularly in origin countries which could be negatively impacted by "brain drain" (e.g., Glinos 2015;Schultz and Rijks 2014;Żuk, Żuk and Lisiewicz-Jakubaszko 2019). Furthermore, physicians are among the most mobile high-skilled professions within the European Union (European Commission 2020). Germany is a popular destination for physicians considering migration within the European Union (European Commission 2020), 2 and the numbers and share of EU physicians in Germany have significantly increased over the last 20 years (Becker 2020, 5f.).
As my sample consists of only physicians who worked in their profession, this article focuses on employment mobility 3 and addresses the following research questions: Do EU physicians in Germany perceive an improvement in their employment situation compared to before migration (i.e., subjective employment mobility) and, if so, regarding which dimensions of their employment situation? How is the perceived change in their employment situation linked to their original reasons for migrating, and how might subjective employment mobility be further influenced by migration-and occupation-related factors?
In what follows, after a brief discussion of the situation for EU physicians in Germany, I sketch the literature on migration and labor-related mobility, elaborate on three sets of variables to explain physicians' employment mobility (migration reasons, other migration variables, and occupational factors), and derive my assumptions. Then, I introduce my data, which are based on an original survey of EU physicians in Germany. The OLS regression results show that physicians who migrated to Germany for career-related reasons perceived an improvement concerning their use of skills and career opportunities, while those for whom economic reasons were important migration drivers perceived their situation concerning personal income and working conditions as better than before. These results suggest that the physicians surveyed achieved their original migration intentions. Physicians from Southern and especially Eastern Europe were, compared to physicians from Northern and Western EU member-states, more likely to perceive an improvement regarding working conditions and, for Eastern EU physicians, also partially regarding income and use of skills. Glass-ceiling effects for those physicians who completed their specialist training but were not (yet) in leading positions can be assumed, as these physicians perceived their employment situation as worse compared to before migration. These results enhance understandings of labor-related high-skilled migration and high-skilled migration within the European Union, as they indicate barriers even for a group of migrants who migration scholars have long assumed do not face obstacles to migrating for professional advancement (Favell, Feldblum and Smith 2007;Ryan and Mulholland 2014). In this way, the analysis shows limits to the idea of an integrated European labor market based on mobile EU citizens not facing discrimination due to their nationality. The conclusion discusses the main findings and contributions, as well as this article's shortcomings, and addresses avenues for further research.

Geographic Mobility and Its Effects on Migrants' Occupational Situation
Most studies observing the relationship between migrants' geographic and labor-related mobility focus on how migration affects migrants' careers or social mobility (e.g., Attias-Donfut and Dimova 2011;Chiswick, Lee and Miller 2005;Stanek and Ramos 2013). These studies have found that most migrants in Western countries are likely to experience downward mobility at migration and upward mobility thereafter, describing a so-called U-curved relation of labor-related mobility and migration (e.g., Chiswick, Lee and Miller 2003;Rooth and Ekberg 2006;Stanek and Veira 2009;c.f., Simón, Ramos and Sanromá 2014). Many of these studies, however, focus on low-skilled migration (e.g., Verwiebe 2008, 1) and third-country nationals (i.e., non-EU citizens). Systematic analyses of the relationship between geographic and employment mobility of intra-EU migrants remain rare (Recchi 2009, 73), perhaps because of a lack of data, as the sample size of intra-EU migrants in many national surveys is often low (Braun and Arsene 2009, 26). Findings of the few available studies on the relation between intra-EU migrants' geographic and labor-related mobility align with previous work describing the U-curved relationship (Lubbers and Gijberts 2016) and have observed their low mobility rates, in terms of intergenerational social mobility (Braun and Recchi 2008;Favell and Recchi 2011).
We know rather little about high-skilled intra-EU migrants' subjective assessments of their professional situations, particularly with respect to different dimensions of the employment situation, before and after migration. Some studies approach intra-EU migrants' employment situation from a more subjective angle and concentrate on their satisfaction with the current situation (e.g., Donegani and McKay 2018;Pantenburg et al. 2016 on foreign-national physicians in Saxony, Germany), but few have compared intra-EU migrants' employment situation in the destination and origin countries. For instance, Wanner (2019) demonstrates how migration reasons influence integration for recently arrived migrants in Switzerland by looking at whether surveyed migrants stated an improvement or worsening of their overall employment situation. The results indicate that the strongest improvement in migrants' employment situation was seen for those who migrated for professional reasons. Lubbers and Gijberts (2016) observe that migrants' satisfaction with income is higher after migration for recent migrants from Poland and Bulgaria in the Netherlands, a finding which is unsurprising in the context of economic migration driven mainly by financial aspects. Assirelli, Barone and Recchi (2019) compare migrating and non-migrating Italian graduates and show that those who migrated had higher salary and higher satisfaction with career opportunities but also higher perceived risks of over-education. Although these studies do not measure change in migrants' employment situation through migration at the individual level, they inform understandings of the labor-related transition from origin to destination country. To the best of my knowledge, no studies observe subjective employment mobility before and after migration for highskilled intra-EU migrants in general or for physicians in particular.
This article, thus, contributes to research on migrants' employment mobility with a detailed analysis of how migrants perceive the change in their employment situation through migration concerning personal income, working conditions, use of skills, and career opportunities. This decomposition of different dimensions of migrants' professional situation allows not only a more detailed consideration of migration's effect on migrants' job positions (which are difficult to track by objectives means) but also a detailed consideration of subjective assessments, which can be important indicators of satisfaction and intentions to stay, return, or move onward (Cieslik 2011). Focusing on one profession further allows us to better account for profession-specific factors that might shape these perceptions, while observing migrants from different EU regions (in one destination country) allows us to track origin effects.

EU Physicians in Germany
Overall, the situation for EU physicians in Germany is rather favorable. On the one hand, recognition of medical professional credentials obtained in another EU member-state is facilitated by the EU directive 2005/36/EC on the recognition of professional qualifications in regulated professions (European Parliament and Council of the European Union 2005). Furthermore, labor demands in Germany are high, although unevenly distributed, with no clear patterns between East and West Germany (Becker 2020, 26f.), suggesting good job prospects for physicians looking for employment. Labor shortages are particularly apparent in smaller cities and rural areas, especially in Eastern German federal states (Hoesch 2012), and in different specialty areas (general medicine, ophthalmology, neurology, psychiatry, and dermatology) (Kopetsch 2010), although the latter varies across regions (Sachverständigenrat zur Begutachtung der Entwicklung im Gesundheitswesen 2014). Job opportunities further vary dependent on the hierarchical position, with more job openings for physicians in lower positions and scarcer opportunities higher up the hierarchy (Deutsches Krankenhaus Institut e.V. and merirandum GmbH 2014).
However, the medical profession in Germany is also shaped by certain preconditions and requirements one must fulfil to practice medicine, including, among others, a license to practice medicine, a medical degree, and a certain degree of German language skills (Marburger Bund 2019). Furthermore, physicians must register with one of 17 State Chambers of Physicians in Germany (Bundesärztekammer 2020). These prerequisites result in more restricted labor market access, making it more difficult for immigrants in particular to enter the labor market.
The education of medical professionals in Germany includes, first, successfully completing medical studies at university and, second and optionally, specializing in a field of one's choice (ibid.). This specialist-in-training period lasts approximately five to six years and includes an exam (ibid.). The medical profession's hierarchical structure in German hospitals usually works in the following way: Physicians intending to become specialists are "specialists in training" and, after successfully completing their specialist training, become "specialists." If they do not start specialist training or continue working at an assistant level after becoming specialists, they are "assistant physicians." Leading positions include, in ascending order, "senior physicians," "leading senior physician," "chief physician," and "medical director." The main distinction in the outpatient sector is between practice owners ("resident physicians") and employees (again, "specialists," "assistant physicians," and "assistants in training") (Hellmann 2020).

Explaining the Subjective Employment Mobility of EU Physicians in Germany
I argue that three sets of factors are particularly important in explaining migrants' evaluations of their employment situation after migration, compared to before migration: their reasons for migrating; further migration-specific factors (e.g., origin region, duration of stay, language skills); and occupational factors in the destination country. As previous research on the relationship between geographic and labor-related mobility has mainly focused on objective employment mobility (e.g., Attias-Donfut and Dimova 2011; Chiswick, Lee and Miller 2005;Stanek and Ramos 2013), a framework to analyze migrants' subjective employment mobility is missing. Building on different research strands that aim to explain labor-related mobility, I offer below a framework to analyze migrants' subjective employment mobility. This framework combines migration-related factors and aspects of the occupation in which migrant professionals work and pays special attention to the importance of original migration reasons for understanding employment mobility. Combining these different sets of explaining factors allows a more comprehensive assessment of subjective employment mobility. I further argue that this analytical framework could be adapted for other professional groups, beyond the present case of EU physicians in Germany. The next sections further develop and explain the three sets of explaining factors.

Reasons to Migrate
Previous studies analyzing migrants' labor market integration indicate that migrants' original reasons for migrating influence their intentions and opportunities to find a job in the destination country, according to their qualifications and in line with their previous career development. Migrating for economic reasons suggests that income is crucial in the migration decision and might lead migrants to accept job positions below their qualifications (e.g., Parutis 2014). For example, Trevena (2013) shows that one reason high-skilled migrants from EU Eastern member-states might accept jobs below their skill level in the United Kingdom is earning money, along with the intention to improve one's situation upon return (see also Eade, Drinkwater and Garapich 2007;Parutis 2014;Ryan et al. 2009).
Those migrating for career-related reasons are triggered by the intention to find not only a job to earn money but also a good job with good prospects, especially if they are high skilled (Favell 2008). The importance of finding adequate employment is also shown by recent research on high-skilled emigration from Southern Europe before and during the European economic and fiscal crisis (Bartolini, Gropas and Triandafyllidou 2017). The significance of career reasons can also be stated for EU physicians who migrated to Germany (Becker and Teney 2020). We could assume that those migrants for whom career-related reasons were important in their migration decisions are more inclined to work on and achieve an improvement in their occupational situation, in terms of costs-benefits calculations and selfselection processes (Chiswick 1999). Furthermore, and in line with human capital approaches (see Sjaastad 1962), we assume that human capital investments influence employment outcomes, as these investments prepare migrants for jobs according to their original skills levels (see Emilsson and Mozetič2021 for high-skilled Latvians and Romanians in Sweden). For instance, Campbell (2016) demonstrates that workrelated, intra-EU migration to the United Kingdom, especially from Eastern EU member-states, leads to higher employment propensities for migrants than does family migration. Similarly, Wanner (2019) finds that professionally motivated migrants have higher levels of labor market incorporation in the destination country than do those who migrated for family or other reasons.
Family-related migration, by contrast, often negatively impacts labor market outcomes. So-called "tied" (Cooke 2013) or "family migrants" are more likely to be less integrated in the destination-country labor market (e.g., Boyle et al. 2001;Meares 2010;Zwysen 2018) and to face greater professional challenges than primary migrants (Wanner 2019). These larger challenges to labor market success become apparent, for example, in lower employment propensities and lower wages, especially for female family migrants (e.g., Campbell 2016;Chiswick 1999). The negative relationship between gender and employment mobility can be explained by the fact that family-related migration is most often female (e.g., Cooke 2013) and was triggered by reasons primarily unrelated to the labor market; thus, professional advancement may have been a secondary consideration for migration. A career penalty has also been demonstrated for high-skilled female accompanying migrants (Föbker 2019), although it is also important to note that family and work-related migration can be interlinked (González-Ferrer 2011).
Although most of the research described above observes labor market incorporation rather than subjective employment mobility, these insights serve as important contextual factors related to this article's focus. Based on my elaborations above, I hypothesize that the more strongly economic reasons influenced migrating physicians' decisions, the more likely they are to perceive an improvement regarding income and working conditions (H1a). The more strongly career-related reasons shaped EU physicians' migration decisions, the more likely they are to perceive an improvement of their employment situation after migration, compared to before, especially regarding the use of skills and career opportunities (H1b). Family-related migration might lead to a worsening of the employment situation regarding all dimensions (H1c).

Migration-Specific Factors to Consider
Furthermore, migration-specific factors can influence subjective employment mobility. Because previous research on the relationship between geographic and labor-related mobility has focused mainly on explaining objective employment mobility (e.g., Attias-Donfut and Dimova 2011; Chiswick, Lee and Miller 2005;Stanek and Ramos 2013), with less attention to subjective employment mobility, I build on these previous findings and show how they might also apply to subjective employment mobility.
The main reason for migrants' downward mobility at migration is a devaluation of human and social capital obtained in another country (e.g., migrants' origin country) that is explained, for example, by problems with recognition of qualifications, language knowledge, or unfamiliarity with specific working procedures in the destination country (Braun and Recchi 2008, 161;Chiswick 1977;Chiswick, Lee and Miller 2003;Fernández-Macías et al. 2015;Recchi 2009). While downgrading due to difficulties with recognition of qualifications is less applicable to physicians migrating within the EU, language knowledge and knowledge of specific working procedures are (Jansen, Hänel and Klingler 2018;Klein 2015;Öhlander, Wolanik Boström and Pettersson 2020;Wolanik Boström and Öhlander 2015). As physicians working in Germany must prove at least upper-intermediate German language skills (Marburger Bund 2019), there is less variation in language skills among migrant physicians than among migrants in general. However, physicians speaking the destination country's main language as a native language might face fewer obstacles. I, thus, hypothesize that EU physicians who are native German speakers perceive their situation as better, especially regarding professional aspects for which language is particularly relevant (career opportunities, use of skills) (H2a). As knowledge of specific working procedures increases over time, the duration of stay might also influence the comparison of the current employment situation with that before migration. I, thus, hypothesize that the longer the duration of stay, the more likely EU physicians perceive an improvement concerning those aspects of the employment situation that are most affected by increased knowledge of specific working procedures (use of skills, career opportunities) (H2b).
Despite EU efforts to foster more equitable economic development across Europe (see the efforts made by the European Regional Development Fund (ERDF)), vast differences concerning Gross Domestic Product (GDP), (un-)employment rates, education, income (Statistisches Bundesamt 2016), and wealth remain between Western and Eastern European member-states (Mau 2004). During the European economic and fiscal crisis, gaps between Southern and Northern EU member-states increased, as Northern member-states recovered more quickly from economic downturn, which also influenced migration patterns (Barbulescu, Lafleur and Stanek 2015;Lafleur and Stanek 2017). These differences between EU regions might also affect migrants' perceived labor-related mobility. Migrants from countries with lower wages, higher unemployment rates, and, in many cases, worse health care systems (compared to the destination country) are likely to experience an improvement in their professional situation. I, thus, hypothesize that Southern and Eastern European physicians (especially those from the 2007 accession states-Bulgaria and Romania) perceive an improvement in their employment situation, compared to Western/Northern European physicians, particularly regarding income and working conditions (H2c). Furthermore, as many physicians from Eastern EU member-states report less, and less advanced, medical equipment in their home countries (Klein 2016), I hypothesize that Eastern EU physicians also experience an improvement regarding their use of skills (H2d). On the other hand, their perceptions of their employment mobility might further be influenced by the current employment situation (see below) in which career advancement might be easier for physicians from some regions than from others (Becker 2019). High-skilled professionals from Western and Southern countries report that they feel more positively treated, for example, by colleagues or superiors, due to their nationality than other migrants, while positive treatment is less often reported by Eastern European migrants (Becker 2019;Grigoleit-Richter 2017; see also Favell 2013). 4 This more positive perceived treatment seems to ease migrant physicians' day-to-day interactions, as they feel more accepted and welcome, but might also affect being offered a promotion. In a similar vein, results from the United Kingdom show that risks of over-education are substantially higher for EU Eastern than for Western migrants (Campbell 2016;Johnston, Khattab and Manley 2015;Zwysen and Demireva 2018). Thus, I hypothesize that Eastern Europeans are more likely to perceive worse career opportunities, as they are more likely to perceive disadvantages due to their nationality (H2e).

Assessment of Occupational Factors
Perceptions of change in one's employment situation might strongly depend on one's current position. For medical doctors, position entails the career stage, as well as where and in what specialty area they work. While opportunities are often more extensive for physicians at earlier career stages, positions higher up the hierarchy are scarcer (Deutsches Krankenhaus Institut e.V. and merirandum GmbH 2014, see also EU Physicians in Germany); thus, competition might be fiercer. Previous research analyzing whether opportunities to reach higher positions are evenly distributed has found that the likelihood of facing "glass ceilings" (i.e., limited career 4 Comparative research on how migrants from different countries, especially within the EU, face different degrees of (perceived) discrimination is scarce. Safuta (2018) introduced the concept of 'peripheral whiteness' to analyze how different distinction mechanisms are applied for white migrant domestic workers from Central and Eastern European countries in Belgium and Poland. Drewski, Gerhards, and Hans (2018) observed 'symbolic boundaries of nationality' between students at a European school in Brussels, indicating a status hierarchy ranging from Northwest-European to Eastern European students. opportunities) increases for higher positions with age, gender, and migration background (e.g., Cotter et al. 2001;Maume 1999;Morrison, White and Van Velsor 1987). We observe that for the medical profession in Germany, the share of female physicians has significantly increased over the last twenty years and surpasses the share of men at lower levels, although higher positions remain male dominated (Coy and Proba 2018;Deutscher Ärztinnenbund Deutscher Ärztinnenbund e.V. 2016). Furthermore, previous research shows challenges for women to be successful in the medical profession (Ziegler et al. 2017). A study of physicians' careers during specialist training indicates a slightly lower intention to take up a leading position among women, compared to men. However, this intention to take up leading positions decreased in recent years for all physicians, and more strongly for men than for women (ibid.).
Glass-ceiling effects concerning gender might be particularly apparent in Germany, where the modified breadwinner model in which women stay at home or are secondary earners is prevalent (Trappe, Pollmann-Schult and Schmitt 2015). The male breadwinner model results in relatively low shares of women in more senior and prestigious positions, as well as comparatively high shares of women in part-time employment (Minkus and Busch-Heizmann 2018). Migration scholars interested in gender dynamics in high-skilled professions in Germany note that highskilled female migrants often experience being perceived as less career-minded, especially if they are married and have children, which these high-skilled females explain at least partially by Germany being a conservative welfare state (Grigoleit 2012).
Employment opportunities also vary within a country, with higher labor shortages in rural areas (Hoesch 2012, see also EU Physicians in Germany). Thus, on the one hand, finding employment and reaching higher positions might be easier in rural areas, where competition might be lower. Séchet and Vasilcu (2015) show that Romanian physicians in France often fill positions in rural areas and in places with the highest labor shortages in the medical profession. Labor shortages in the medical profession exist for different medical specialty areas in Germany (Kopetsch 2010, see also EU Physicians in Germany). On the other hand, the availability of a large variety of advanced medical procedures is limited in rural areas, which might also limit the use of physicians' skills and further career prospects (for Sweden, see Sturesson et al. 2019). The relation between career development and working in rural areas is, however, theoretically under-developed and needs further exploration in future research.
Based on these considerations, I hypothesize that migrant physicians aiming for the highest positions in Germany (currently working as specialists) perceive their career opportunities as worse than before migration, as they would not face glass ceilings due to foreign nationality in their origin country (H3a). I also hypothesize that female physicians are more likely to perceive their career opportunities and use of skills as worse, especially if they aim for the highest positions (H3b). I further hypothesize that physicians in rural areas might perceive their career opportunities as better but their use of their skills as worse (H3c) and that physicians working in specialty areas with high labor shortages might perceive their opportunities as better (H3d).

Data and Methods
The analysis presented here is based on an original survey of physicians who held EU citizenship (except Germany) and migrated to Germany after 2003 (Teney et al. 2017). The sample is randomly drawn from the registers of 12 (out of 17) local State Chambers of Physicians which agreed to support our study. For practical reasons, the year of migration is, for the sampling procedure, defined as the year of registration at a local State Chamber of Physicians. The sample, thus, only includes physicians that fulfilled certain preconditions (e.g., medical degree, at least upper-intermediate German language skills) and intended to work as medical doctors in Germany. Those not working as physicians or those who left Germany are not part of the sample. This sample definition must be kept in mind when looking at the results, as physicians perceiving subjective upward mobility might be over-represented in the sample.
The survey was conducted between September 2015 and May 2016. The response rate was 23 percent, with 1,607 eligible cases (Teney et al. 2017). Due to the low response rate, I cannot exclude an uncontrolled bias as, for example, migrants satisfied with their situation might be more prone to participate in the survey. For the analysis, I excluded physicians who were already retired (defined as older than 65) or who migrated as children, as well as cases with missing values on central variables (e.g., year of migration, citizenship, year of birth), leaving 1,286 respondents. Missing values were imputed if possible (see below for more details). After a listwise deletion of cases with missing values for my dependent, independent, and control variables, the sample is 1,058. Due to the sampling procedure, the sample over-represents recent migrants. A comparison of respondent distribution across origin regions with the only existing reliable statistics on EU physicians (Bundesärztekammer 2016) indicates a deviation up to five percentage points between these statistics and my sample. 5 The over-representation of physicians from 2004 EU accession states and under-representation of physicians from Southern EU countries can be explained by the fact that physicians who migrated 5 Distribution in the sample vs. all registered physicians in Germany   after 2004, who are over-represented in the sample, mainly came from Eastern EU member-states. My analytical strategy is based on robust OLS regressions to test the effect of migration reasons, other migration-specific factors, and occupational variables on physicians' subjective employment mobility regarding different dimensions. Analyses are carried out using STATA.

Dependent Variables
Survey respondents were asked to compare their current personal and professional situations with their situations before they migrated to Germany on a five-point scale. The question reads as follows: "Now please compare your current personal and professional situation to your situation before you moved to Germany: How would you describe your current situation regarding the following areas of your life?" "personal income," "professional career opportunities," "working conditions," and "using your professional skills." The answer categories included "much worse," "worse," "about the same," "better," "much better," and "does not apply." Variables were coded from 1 ("much worse") to 5 ("much better"); "does not apply" was coded as missing. Missing values on one variable were replaced by the mean across the other variables but only if only one value was missing across all four variables.

Independent Variables
Migration reasons. Physicians surveyed were asked to rate different reasons why they had left their origin country and migrated to Germany on five-point scale ranging from "unimportant/irrelevant" to "extremely important." Based on previous research, in which the most important migration reasons were found to be economic (national economic situation, working conditions, income, and standard of living), career, and family reasons in both the origin country and Germany (Becker and Teney 2020), 6 I built mean indices for each set of variables. The indices allow for one missing value on the variables measuring family and career reasons and for up to two missing values on the variables measuring economic reasons. I confirmed each scale's goodness of fit with Cronbach's alpha (family reasons: 0.90, career reasons: 0.87, economic reasons 0.92).
Other migration-specific variables. Origin region was measured by physicians' citizenship. If more than one citizenship was indicated, I used the citizenship listed first (if an EU citizenship, except Germany). I attributed respondents' citizenship to the following categories: Western/Northern EU, Southern EU, Eastern EU 2004(accession 2004), and Eastern EU 2007/13 (accession 2007) (see Footnote 5). Duration of stay was measured in years since arrival in Germany. Language skills were measured as German as a native tongue (or at C2 level, according to the common European reference frame which indicates "proficient users" on "mastery" level, see Council of Europe 2020) yes versus no, as most physicians had a high language skill level due to the administrative preconditions to register as physicians in Germany.
Professional variables and employment situation in Germany. The current employment situation contained four categories: "leading position," "specialist," "in training," and "assistant." For "leading position," I coded the following positions: acting senior physicians, senior physician, leading senior physician, chief physician, and resident physician. Gender was coded male versus female. The specialty area was asked for with an open question, and physicians could indicate whether they worked in a specialty area or obtained a degree in this specialty area. Specialty areas with more than 5 percent of the observations were coded as separate categories, specialty areas with a share lower than 5 percent were coded as "other." Missing values were subsumed under the reference category (internal medicine, which is the largest category and for which we would not assume any (dis)advantages). An additional variable indicated whether the information was missing. Regional opportunities were measured with a binary variable (city vs. countryside) (see Table 1 for an overview of all independent and control variables).

Descriptive Results
The majority of physicians surveyed indicated that their overall employment situation in Germany was better than their employment situation before they migrated (M = 3.61). The four dimensions-personal income, career opportunities, working conditions, and use of professional skills-strongly related to each other, and all helped explain the same concept of subjective employment mobility (see Online Appendix, Table 1A for a principal component analysis). However, there was visible variation concerning the four dimensions (see Figure 1). Almost three fourths of physicians surveyed perceived their current personal income as better than before they migrated, and about 59 percent perceived an improvement concerning their working conditions. Only about half of physicians surveyed, however, thought that their career opportunities and use of professional skills were better in Germany than before migration, while about one third perceived their career opportunities and use of skills as about the same as before migration. Between 16 percent and almost 20 percent assumed a worsening of their occupational situation regarding the four dimensions. As this distribution shows, a large share of physicians surveyed perceived an improvement of their situation. Nevertheless, many also thought that their situation was about the same, while about a fifth reported a worsening of their employment situation through migration. This perception of a worsening of their employment situation is particularly interesting, as a large majority migrated for career-related reasons (see Table 1). Table 1 gives an overview of all independent and control variables used in the analysis. The data confirm the importance of career-related reasons in the migration decision (M = 4.20). About a fifth of the sample held Western/Northern European citizenship or citizenship from a Southern EU member-state, about a fourth came from Bulgaria, Romania, or Croatia, and the largest group was from countries that joined  the European Union in 2004 (about 35 percent). The mean duration of stay was 7.58 years, with a large share having spent less than six years in Germany. More than a fifth of the sample held a leading position at the time of the survey, while the largest group was still in specialist training (about 41 percent). The majority worked in cities and in the specialty areas of internal medicine and surgery. Slightly more physicians were female (55.48 percent). The majority were 35 years or younger (53.12 percent), almost a third were between 36 and 45 years old (29.02 percent), and 17.86 percent were older than 46 (not displayed in Table 1).

Multivariate Analyses
As stated above, I offer an analytical framework to analyze migrant professionals' mobility. The framework combines migration-related factors and aspects of migrant professionals' occupation by paying special attention to the importance of original migration reasons to understand migrants' employment mobility. Based on my elaborations, I derived different assumptions on these different sets of explanatory variables (see Data and Methods). To test my assumptions about the effects of migration reasons, other migration-related factors, and occupational aspects on subjective employment mobility, I ran several OLS regressions on each dimension of the employment situation: income, working conditions, use of skills, and career opportunities. The main analyses display the effects of migration reasons, other migration-specific variables, and occupational variables on all four dimensions. Additional analyses observe interaction effects concerning gender and occupational position. Figure 2 presents the effects of migration reasons, other migration-specific variables, and occupational factors on surveyed physicians' perception of the change of employment situation for each dimension separately (see Online Appendix, Table 2A for the exact values of the regression coefficients and their standard errors). 7 The results show the expected effects of economic reasons for migration on the perceived improvement of personal income and working conditions. Those who migrated for economic reasons were significantly more likely to perceive an improvement concerning income and working conditions (H1a). The analyses also show that economic reasons were significantly positively linked to career opportunities. Furthermore, career reasons positively correlated with the perceived improvement 7 As the residuals were not completely normally distributed (Skewness: -1.10 for income, -0.65 for working conditions, -0.39 for use of skills, -0.46 for career opportunities), I ran robust OLS regressions. Tests of multicollinearity were assessed with the post-estimation command estat vif in Stata and indicate no problems (mean vif: 1.73, max. vif: 3.51). I also ran linear probability models on each dependent variable as robustness checks, where the dependent variable was coded as 0 ("worse or the same") or 1 ("better"). These models show substantially the same results (see Online Appendix, Table 3A for more details).

Figure 2. OLS Regression on Perceived Improvement as Dependent Variable (by Dimension).
Note: The results present the values of the regression coefficients with their confidence intervals. Significance levels are indicated with *** p < 0.001, **p < 0.01, *p < 0.5. Reference categories: region of origin: EU Northern/Western; native German: no native German speaker; occupational position: leading position; gender: male; specialty area: internal medicine; region in German: city; student: no student. Not significant effects for the specialty area are omitted in the output.
concerning use of skills and career opportunities. The values of 0.180 (use of skills) and 0.135 (career opportunities) indicate that an increase in the importance of careerrelated reasons for migration by one unit increased the probability of a perceived improvement of the employment situation by 18 and 13.5 percentage points, respectively. These positive correlations between career-related migration reasons and perceived improvement concerning use of skills and career opportunities support the assumption that those for whom professional development was crucial for their migration decision were more likely to, indeed, experience such an improvement concerning these aspects of their occupational situation (i.e., regarding use of skills and career opportunities) (H1b). These relations between career-related reasons and perceived improvement of their situation hold true independent of whether physicians had a job offer before migration (see Online Appendix, Figure 1A for further analysis). The effects for family reasons were, however, low and not significant in all four models. A negative effect for those physicians who were motivated by family-related reasons to migrate can, thus, not be confirmed by my analyses (H1c). That there was no effect of family reasons on subjective occupational mobility could be partially explained, on the one hand, by the fact that surveyed physicians' migration decisions were often not determined by only one reason. On the other hand, family-related reasons entailed not only the traditional meaning of (mainly) female migrants migrating for their spouses' career but also, for example, that migrants chose a country where both spouses can pursue their careers, especially when they are from different origin countries (Becker and Teney 2020). Furthermore, a family penalty might be less applicable for physicians, especially from within the European Union, due to high labor shortages in Germany (Becker 2020, 25ff.).
Turning to the other migration-related factors in the model, the effect for German as a native language was significant and positive for use of skills and career opportunities, confirming H2a. Native-German speaking physicians, compared to nonnative speakers, were more likely to experience an improvement in their employment situation compared to before migration, with a probability of 27.0 percentage points (use of skills) and 37.2 percentage points (career opportunities). Length of stay was very weakly correlated with the dependent variable and not significant across all four models. This non-significant effect is surprising and contradicts the assumption (H2b) that increased work experience in the destination country results in an improvement of one's employment situation. However, it must be considered that a high share of the surveyed physicians had migrated rather recently. Physicians from Southern andEastern (2004 and2007/13) EU member-states were more likely to perceive an improvement in their employment situation, compared to Northern/Western EU physician migrants. These effects were stronger for Eastern than for Southern EU physicians and (except for income) stronger for Eastern European physicians from Bulgaria, Romania, and Croatia than for those from the 2004 accession states. However, these effects were only significant for all origin regions concerning working conditions and for Eastern EU physicians (accession 2004) concerning income, partially confirming H2c. Furthermore, physicians from Eastern Europe (2007/13) were more likely to experience an improvement concerning use of skills, partially confirming H2d. The positive correlation between origin region and working conditions, income, and use of skills can, to some extent, be explained by the difference between EU regions concerning welfare levels (see also Migration-specific Fators to Consider). The negative effects for Eastern Europeans concerning career opportunities cannot be confirmed (H2e).
Turning to the occupational effects, those who worked as specialists were more likely to perceive their situations as worse than before migration. Especially concerning career opportunities, being a specialist, compared to being a leading physician, decreased the likelihood of subjective employment upward mobility. These results are in line with the above-mentioned assumption (H3a) that migrant physicians might, at a certain career stage, perceive worse opportunities than they would in their origin country. However, perceived downward employment mobility is shown not only for career opportunities but also for all measured aspects of their occupational situation. Furthermore, those who were still in training were more likely to perceive their income as worse than before migration, compared to those in leading positions. The gender effect was not significant in the four models. The absence of a significant effect for gender on subjective employment mobility also holds when examining the interaction between gender and occupational position regarding career opportunities. I conducted an additional robust OLS regression model which expanded the models above by adding an interaction term between gender and occupational position (Figure 3). The interaction term between gender and current occupational position was not significant. However, the main effect for specialists remained stable. These results show that male specialists, in comparison to leading physicians, perceived their career opportunities as worse than before.
The assumption that female physicians in higher positions perceive subjective downward mobility (H3b) cannot be confirmed by my analysis. The effect size and direction of all other variables remained more or less stable in the extended model.
Whether a physician worked in a city or the countryside had a significant negative effect on their use of skills but not concerning career opportunities (see Figure 2), partially confirming H3c. This result supports the perception of a disadvantage of those working in rural areas. Those physicians working in the medical field "surgery" perceived their situation as significantly worse regarding all four dimensions, and those in working in the medical field "psychiatry" as better regarding income and working conditions, than before (compared to physicians working in "internal medicine"). These correlations only partially support my assumption that physicians working in specialty areas with high labor shortages might perceive their opportunities as better (H3d). The effect for surgery can, moreover, be explained by the fact that this specialty area is known to be highly competitive. However, the number of cases per specialty area limits the analysis. Note: The results present the values of the regression coefficients with their confidence intervals. Significance levels are indicated with *** p < 0.001, **p < 0.01, *p < 0.5. Reference categories: region of origin: EU Northern/Western; native German: no native German speaker; occupational position: leading position; gender: male; specialty area: internal medicine; region in German: city; student: no student. Not significant effects for the specialty area are omitted in the output.

Conclusion
This article set out to better understand the subjective employment mobility of EU physicians in Germany. Subjective employment mobility is understood here as migrants' perception of the employment situation before migration compared to after migration. As physicians' employment situation is shaped by different conditions, I analyzed different dimensions of the occupational situation. The main research interests were, first, to observe whether the group under scrutiny perceived an improvement in their employment situation regarding income, working conditions, use of skills, and career opportunities through migration and, second, to determine which factors explained why some physicians perceived their employment situation regarding these dimensions as better or worse. In the absence of a theory to analyze the subjective dimension of the relation between geographic and labor-related mobility, I offered an analytical framework for assessing migrants' subjective employment mobility-one which can be applied beyond the case of intra-EU migrant physicians. This framework combines migration-related and occupational aspects and pays special attention to the importance of initial migration reasons.
Combining these different factors allows us to approach the complexity of circumstances which influence employment mobility. On the one hand, migrants have different intentions to migrate, and these different intentions might influence their motivations and restrictions to find a job according to their skill level (see Reasons to Migrate). On the other hand, migration-related factors such as mastering the destination country's language influence career opportunities. Additionally, and often on a more subtle level, migrants' origin country/region seems to affect how they perceive the benefit of their migration choice (see Migration-specific Factors to Consider). The former two aspects are, however, interlinked and dependent on the profession in which one works. Labor shortages might explain why labor market access is easier. Nevertheless, reaching higher positions is harder, especially for professionals who obtained their degree abroad, making it crucial to consider the professional structure, as well as the point in time in migrants' career at which they transition to another country. Analyzing subjective employment mobility is crucial to better understand the effects of labor-related migration for migrants themselves and allows us to grasp dimensions of the occupational situation which are hard to measure objectively (e.g., use of skills, career prospects). Furthermore, subjective employment mobility might influence migrants' satisfaction, as well as their intentions to stay, move further, or return to their origin country.
The descriptive results presented here show that a large percentage of physicians surveyed perceived their employment situation as better than before. However, we also observe large differences regarding the single dimensions of their employment situation (i.e., use of professional skills, career opportunities, working conditions, personal income), ranging from 49 percent perceiving an improvement regarding use of skills to 74 percent perceiving an improvement regarding personal income. The multivariate analyses showed that migration reasons, other migration-specific factors, and occupational variables helped explain physicians' subjective employment mobility. First, regarding migration reasons, results indicate that the centrality of career-related reasons in the migration decision went hand in hand with a perceived improvement concerning the surveyed physicians' use of skills and career opportunities, while the importance of economic reasons seemed to be strongly linked to a perceived improvement of personal income and working conditions. Family-related reasons, however, did not seem to influence physicians' subjective employment mobility. Overall, physicians surveyed did perceive an improvement concerning the dimensions of their employment situation which they originally intended to improve. The extent to which this perceived improvement was influenced by high labor demands for physicians must be further explored by comparing the results to other high-skilled professions. On a more general level, the results emphasize that the original reasons to migrate must be considered when analyzing migrants' (subjective) employment mobility.
Second, other migration-specific factors influenced physicians' subjective mobility to a varying extent. Duration of stay in the destination country seemed to have no influence on subjective employment mobility. Thus, increased knowledge of the workings of Germany's medical profession seemed not to influence physicians' assessments of their employment situation compared to before migration. This result is surprising, considering previous findings of a U-curved relation of labor-related mobility (e.g., Lubbers and Gijberts 2016). Factors in the origin country, such as income levels and working conditions, might influence migrants' evaluation of changes in their employment situation more strongly. On the other hand, this seeming absence of a U-curved relation could suggest that high-skilled migrants within the European Union face less (perceived) downgrading. Despite these possible explanations, it must be kept in mind that the majority of the sample arrived in Germany rather recently; thus, the effects of duration of stay might not yet have come into play. German native language skills positively correlated with subjective employment mobility in areas of work most reliant on language skills. However, as most native-German-speaking migrant physicians were from Austria, we cannot exclude that the effect of being a native German speaker could be a country-specific, rather than a broader language effect, as previous research has indicated that Austrian physicians often take assistant-in-training positions in Germany to circumvent the long general training period before specializing in Austria (Becker and Teney 2020). Differences concerning the origin region can be observed but were only found to be significant for some dimensions and regions. The effects regarding origin region were most likely linked to welfare differences and shed light on how physicians might use their migration to a country with better income prospects, working conditions, and use of skills as a career move.
Third, this article shows correlations between physicians' occupational position and their perceived career opportunities compared to before migration, particularly for male specialists. The finding that specialists perceived their career opportunities as worse compared to before migration, in comparison to the reference group (i.e., physicians in leading positions), is in line with the assumption concerning glass ceilings which migrants might not face in their origin country, showing that at least to some extent, even high-skilled migration within the European Union is not frictionless.
Despite these contributions, this article does have certain limitations. Given the sampling strategy, my data only allow us to draw conclusions for a specific group of European physicians in Germany: those working in Germany's medical profession. Those not working as physicians and those who left Germany are not part of the sample. This sample definition might lead to under-representation of those perceiving their situation as worse, although we do not know why physicians return, stay, or migrate further. Second, the cross-sectional design cannot trace the development of physicians' employment situation over time or track how their subjective employment mobility might have changed since arrival. Thus, further analyses should trace the development of their subjective employment situation and the effects of subjective employment mobility on return intentions.
On a more general level, this article's results inform understandings of labor-related high-skilled migration. High-skilled professionals are often assumed to face fewer obstacles in migrating for professional advancement (Favell, Feldblum and Smith 2007;Ryan and Mulholland 2014). This supposedly smooth mobility holds particularly true for migration within the European Union, which is based on the idea of an integrated European labor market in which EU citizens can migrate without, at least in theory, facing disadvantages and discrimination due to nationality. On the one hand, this article's results give an example of how high-skilled intra-EU migrants are, indeed, able to use their EU mobility rights to improve their employment situation. On the other hand, the analyses also indicate the limits to this idea of an integrated European labor market and obstacles even for high-skilled migrants. As it shows, not only do migrants from different EU regions unequally perceive the benefits of migration, but even high-skilled migrants face glass ceiling effects when aiming for leading positions. While my analyses are based on only one professional group, these results may also hold true for other migrant professionals, although the specific occupational context must be considered to understand labor-related migration.