Understanding discrimination in hiring apprentices: how training companies use ethnicity to avoid organisational trouble

Abstract Children of immigrants from non-EU countries face particular problems to access apprenticeship training in German-speaking countries. In this context this article asks how recruiters in small and medium sized companies (SME) make sense of national and ethnic origin when hiring new apprentices. The author proposes Boltanski and Thévenot’s theory of justification in order to conceptualise ethnic discrimination in hiring. Accordingly, the social body of a company consists of multiple interweaved (industrial, domestic, market) ‘worlds’ of social coordination and justification. In order to avoid organisational trouble and to guarantee the further existence of the company, these worlds claim different principle of personnel assessment, some of them penalising applicants of specific ethnic origin. Empirically, the article refers to apprentice recruitment in Switzerland and Germany. It illustrates that employers in SME expect trouble in the domestic and in the market world of the company when hiring school leavers they perceive as foreigners. Hence, discriminatory categories such as ethnicity are used as symbolic and organisational resources for trouble avoidance in hiring apprentices.


Introduction
Similar to Germany, Switzerland has become a major immigration country during the second half of the last century. Three out of ten lower secondary school students are of foreign nationality in Switzerland, and immigrants from the Western Balkans (the successor states of Yugoslavia) and Turkey constitute two of the largest immigrant populations from non-EU countries (Schnell and Fibbi 2016). Both groups include a large proportion of youth. They accounted for up to 40% of the foreign lower secondary school leavers 1 in the mid-2000s who were called upon to transition to upper secondary education, with vocational education and training (VET) being the dominant educational offer in Switzerland.
Again similar to Germany, apprenticeships in small and medium sized companies (SME: companies with 250 employees at most) constitute an important part of the Swiss VET system. 3/5 of youth cohorts leaving compulsory school benefit from a company-based apprenticeship, of which nearly nine out of ten are offered by SMEs (Buchholz et al. 2012). From the age of fifteen, school leavers have to search for a company who is willing to train them in one of over 200 training occupations. School-based VET also exists but four out of five VET offers are company-based (Seibert, Hupka-Brunner, and Imdorf 2009). By completing their apprenticeship, students attain a Federal VET Diploma, which enables their access to the highly segmented qualified labour market. 2 Holders of professional credentials obtain societal recognition and social security, and they face lower risks of youth unemployment and precarious work careers.
However, access to apprenticeship training is rather unequally distributed among school leavers of different nationality and ethnicity in Switzerland, comparable to Germany. Some immigrant students face problems to access apprenticeship -especially children of immigrants from non-EU countries, referred to as 'foreign youth' in public othering (Imdorf and Leemann 2012). 3 'Foreign' students face reduced chances to access apprenticeships compared to native applicants even if they leave compulsory school with comparable academic achievements (Beicht 2015;Buchmann et al. 2016). A recent correspondence study in Germany shows that training companies do less often respond to application letters from youth with Turkish family names than to letters from youth with German names (Schneider, Yemane, and Weinmann 2014). These findings point to a potential employer-sided resistance to hire 'foreign youth' as apprentices (Scherr, Janz, and Müller 2015).
One could object that, as a consequence of the demographical drop of school leavers in Switzerland in recent years, the problems of non-native school leavers to access apprenticeships have decreased during the last few years (Imdorf and Seiterle 2015;Seibert 2015). However, demographic scenarios of the Federal Office of Statistics (BFS 2015, 14) predict again an increase of school leavers as of 2016, which might anew rise the competition for apprenticeship places at the disadvantage of 'foreign' youth in the next years. Moreover, Germany currently faces the challenge to integrate a large number of young refugees in the VET system. Hence, an in-depth understanding of company-sided mechanisms that might hinder the access of minority youth to dual VET programmes seems crucial in order to improve their integration in post-compulsory education.
Previous VET research in German-speaking countries has primarily relied on the analysis of statistical residuals concerning the disadvantage of minority youth. However, statistical disadvantage does not necessarily need to be caused by employer discrimination, and until recently discrimination processes in recruitment for VET have hardly been researched. This is especially the case for Anglo-Saxon countries, where the few existing studies are very dated. They refer to the 80s (Lee and Wrench 1983;Waldinger and Bailey 1991) or even earlier (Marshall and Briggs 1967), and they do not offer convincing theoretical arguments to better understand why employers discriminate in contemporary VET. In Germany, research on ethnic discrimination in VET has only recently been developed, stimulated by the apparent ethnic penalty at the transition from school to VET in Germany in the 2000s, subsequent and with reference to a Swiss study on apprentice recruitment in SMEs by the author (Imdorf 2010). In order to offer a suitable, data-driven and policy-relevant theoretical framework of discrimination in apprentice recruitment, the author has published a series of analysis since the late 2000s (Imdorf 2009(Imdorf , 2010(Imdorf , 2012(Imdorf , 2013(Imdorf , 2015, together with several applications of the framework to the integration of minority youth into VET (Seibert, Hupka-Brunner, and Imdorf 2009;Buchholz et al. 2012;Imdorf and Leemann 2012;Imdorf and Seiterle 2015).
The main contributions of this article are therefore twofold. First, it fills the gap of missing Anglophone research on ethnic discrimination in apprentice recruitment, thereby offering an encompassing conceptual framework to understand current and emerging conditions of ethnic discrimination in apprentice recruitment. Second, the article discusses the main conceptual findings of the author's Swiss study in the light of both previous and most recent research on employer discrimination in German VET.
Ethnic discrimination in apprentice recruitment is best studied when it is most pronounced, that is when demand and competition for apprenticeship places are the highest (Baert et al. 2015). In Switzerland this was the case in the 2000s, when competition for apprenticeships has been increasing. Since 2007 a demographic decline of school leavers has resulted in a decreasing demand for apprenticeship positions and in an improvement of vocational opportunities for foreign youth (Granato and Ulrich 2014;Imdorf and Seiterle 2015;Bahl and Ebbinghaus 2015). Therefore, this article especially refers to the mid-2000s, when ethnic inequalities in accessing apprenticeships were pronounced in a context of apprenticeship shortage.
Given that ethnic inequalities in employment persist worldwide, this contribution may also offer new insights in ethnic employment discrimination beyond apprenticeship in youth labour markets. The recruitment of apprentices seems especially suited to study discrimination in hiring. Although the apprenticeship itself is highly regulated and based on a contract between the company, the state and the student's parents, there are no anti-discrimination regulations in Switzerland when it comes to the preceding selection activities. Additional characteristics of the hiring process cause pragmatic solutions in hiring: personnel specialists in SMEs can face numerous applications from adolescents without work experience but with a rather ambiguous future and they have only a short amount of time and little professional know-how available to deal with these challenges.
In the context of 'foreign youth' exclusion from apprenticeship in the last decade, this article asks how recruiters in SMEs make sense of ethnicity when hiring new apprentices. It aims at developing a theoretical model of personnel selection and (ethnic) discrimination that goes beyond the prevalent assumptions of individualised productivity (inherent in concepts such as statistical discrimination or implicit prejudice) when recruiters assess the quality of job applicants. It thereby counters widely used human and social capital arguments in the study of ethnic inequalities in employment. Rather, the theoretical explanation of discrimination accounts for the organisational motives underlying hiring practices in work organisations.
In the following the most prominent current explanations of discrimination in hiring are presented in Section 2 and integrated in a comprehensive theory of hiring in Section 3. Conceptually, the article discusses companies' motives to resist hiring 'foreign' youth by proposing a framework drawn from the French Economics of Convention -based on Boltanski and Thévenot's (2006) theory of justification -to capture the complexity of companies' personnel recruitment processes. By highlighting different principles of social coordination in the company this approach permits a comprehensive theorising of recruitment and selection from an organisational perspective. The qualitative research project of the author is briefly described in Section 4. Based on both the Swiss study and German research, Section 5 illustrates some major hiring barriers for 'foreign youth' in different organisational contexts of small training companies, and discusses them in the light of most recent research on apprentice recruitment in Germany. Section 6 concludes and suggests some policy responses to reduce employment risks for minority youth in Switzerland and Germany.

Previous modelling of discrimination in personnel selection
Human capital theory (Becker 1993) and social network theory (Holzer 1987;Granovetter 1995) are probably among the most popular and dominant concepts used by scholars to explain social inequalities in access to employment. However, field experiments such as audit studies or correspondence testing (Pager 2007) show that significant inequalities remain even controlling for human capital and social networks. While field experiments enable the measurement of discrimination in employment, they cannot explain why employers discriminate based on ethnicity, gender, or age to sort out job candidates in the selection process.
Beyond VET, research on discrimination in employment has a long tradition and a wide range of explanations have been emphasised (for a review see Pager and Shepherd [2008]). Most widespread are approaches which try to answer the question 'who might be least productive' among job candidates in the perception of the employer. They are based on the assumption, that applicants are equipped with an individual productivity or competence, which could be measured objectively under ideal conditions (Eymard-Duvernay 2008, 56). Models of statistical discrimination (Phelps 1972) emphasise the shortage of information about an applicant's individual productivity and the cost to assess it through adequate selection procedures. Employers resolve the problem by making generalisations about group-specific individual productivity based on beliefs supported by past experience or other forms of evidence. Models based on (implicit) prejudices (Quillian 2006), in contrast, emphasise a negative affective feeling (antipathy) of employers towards a target group and a poorly founded belief (stereotype) about productivity-related attributes of their members.
However, these concepts are limited because they reduce job applicants to employers' assumptions about their individual productivity. They neglect the complexity of social coordination and dependencies at the workplace, all of which contribute to the productivity of the company (Kirschenman and Neckerman 1991, 231). From this perspective, productivity is seen 'as the outcome of social relations at the workplace, rather than solely the outcome of workers' skills' (Shih 2002, 102). Thus, beyond human capital, horizontal and vertical social relations within the company come to the fore. In addition, companies are not only bound to their workforce, but also to their clientele whose needs and wishes they are willing to respect. Following Becker (1971, 75, 76) sex, race, religion and personality of sales personnel can be part of a consumer's evaluation of a retail store. Holzer (1996, 102), among others, has shown that employment of blacks is significantly higher at establishments that serve a black clientele than in those with mostly white customers. The clientele for a company's services or products may create race-specific demands for workers, especially in organisations in which employees interact with customers (Reskin, McBrier, and Kmec 1999, 343;Brief, Butz, and Deitch 2005, 130).
It is plausible to assume that training companies disadvantage 'foreign' apprenticeship candidates beyond their 'human capital' , which, after the completion of compulsory school, mainly consists of academic knowledge and skills beyond any substantial work experience. However, academic skills only matter secondarily for the successful completion of a company-based apprenticeship. Successful graduation from an apprenticeship requires only that apprentices comply with the minimal academic requirements of the vocational school, which they attend during one to two days per week. Hence, reports from lower secondary school, which precedes apprenticeship training, mainly serve recruiters to ascertain whether or not candidates are in a position to satisfy the minimum requirements of the part-time vocational school (Imdorf 2009). While large companies rely on low-cost methods for early selection and therefore use school certificates as a screening device, pre-selection procedures in small and medium-sized businesses concern the 'personal impression' made by the candidates (Stalder 2000;Schmid and Storni 2004) to assess the social fit of candidates right at the outset of the recruitment process. 4 This promotes the influence of discriminating categories, such as the stigma attached to being 'foreign' , when companies fill apprenticeship positions. Two older studies on discrimination in hiring apprentices (Lee and Wrench 1983;Schaub 1991) have indeed revealed that personnel managers question both the staff's willingness to accept 'foreign' youth as well as the latter's willingness to integrate and to subordinate to their supervisors. Managers furthermore expect higher dropout rates of 'foreign' youth from apprenticeship training.

An organisational model of personnel selection: satisficing multiple claims of social coordination
It has been emphasised that more general organisational approaches are needed to fully understand why employers discriminate (Feagin and Eckberg 1980;Pager and Shepherd 2008). 5 As far as the recruitment of apprentices is concerned, previous Anglo-Saxo and German research has lacked appropriate theoretical foundations to satisfactorily explain the exclusion of disadvantaged school-leavers from training companies.
Hence, the investigation of recruitment processes in training companies calls for a theory by which personnel selection and discrimination can be understood and which does justice to the complex social organisation of work organisations, especially in small training companies which form the backbone of the Swiss economy. The French Economics of Convention resolves both claims. This embodies a transdisciplinary approach, established in French economic sciences as well in French sociology, and is founded on the pragmatic theory of justification of Boltanski and Thévenot (2006). The core element is the observation and analysis of the forms by which people coordinate concrete action among themselves or with their environment, as well as the conventions contributing to the shaping of a common ground of perspectives, thus making it possible to coordinate and justify action (Dodier 1993). The term conventions denotes collectively established cultural forms, that are structured, evaluated and legitimised in relationships between actors. It refers to inherently social rules of coordination between actors that are considered to be just (Eymard-Duvernay 2004, 72, 92). The evaluation of the 'quality' of one actor is based on conventions in varying contexts of coordination with a unique principle of social order and justice termed by Boltanski and Thévenot (2006) as 'worlds' .

In search of acceptable and accepted apprentices
This article proposes to integrate the different mechanisms behind discriminatory hiring in one congruent theoretical framework of personnel selection. The starting points are organisational needs and constraints of training companies, the latter being forced to economically survive within their industrial market. 6 Sooner or later, the training of apprentices must result in a profit and contribute to the preservation of the company. At the same time, apprenticeships are part of the common good fabricated in training companies (Eymard-Duvernay 2004, 71): they offer professional education to school leavers and facilitate their labour market integration.
The main problem when companies appoint new apprentices consists of matching wellknown, experience-based internal categories of a company's requirements (including professional and social fit [Kohlrausch 2013]) with mostly unknown candidates who are mainly identifiable on the ground of distinctions offered by the company-external environment. Both the multidimensional organisational demands -which constitute the job profile but cannot be measured by focusing on a few simple skills (Eymard-Duvernay 2008, 65) -and the lack of information about candidates constitute major decision problems. As profit maximisation doesn't work operationally, the 'perfect candidate' cannot be found, and companies stop their efforts to search for adequate candidates once they have found one who sufficiently satisfies their needs (according to the concept of 'satisficing' , see Simon [1957], 204).
Such pragmatic matching of the apprenticeship position with 'the best candidate' needs to fulfil an essential precondition: the selected apprentice must be credible and legitimate in the eyes of others. Even though small businesses are rather inconspicuous and less interesting to the public than big enterprises (Pichler, Pleitner, and Schmidt 2000), they are still not hidden. Outside the company, customers and business partners are their public and critics. At the same time, companies have to explain themselves to their own employees to ensure their ongoing cooperation and motivation (Boltanski and Chiapello 2005). Thus the relevant public exists both within and outside a company's walls and potential criticism or protest of customers, as well from within the workforce, must be minimised.

Coordination at the workplace: legitimate criteria to assess job candidates
To handle the complexity and ambiguity inherent in personnel decisions, companies depend -according to Boltanski and Thévenot's (2006) -on orders of justification, which enable them to legitimise their choice. Personnel decisions can be justified if they refer to selection rules considered just and fair by both the employer and the relevant public; that is, in the case of SMEs, the workforce and customers. Selection criteria are agreed upon if they serve a common good which goes beyond the company's self-interest (Eymard-Duvernay 2004, 66). An important assumption of the theory of justification is that the provision of a public good through workplace action necessitates a specific form of social coordination of humans (but also among humans and things, such as tools, machines and equipment) to mutually adjust their behaviour. Work organisations thereby require from their members to act according to plural conventions of workplace coordination during a typical workday, such as: working efficiently at the workbench or at the assembly line; having a friendly chat with colleagues during the break; dealing successfully with customers at the front office, and so on. Accordingly, recruiters have diverse but limited principles of coordination (conventions) and respective orders of justification at their disposal to assign quality to applicants in the selection process and to justify their decision. Job candidates possess quality if they are considered capable to coordinate in a specific situation at hand (Boltanski and Thevenot 1999, 363). Whereas Boltanski and Thévenot (2006) and Boltanski and Chiapello (2005) have proposed seven 'worlds' of coordination in enterprises all together, this article discusses those three conventions which have proven especially helpful for understanding the exclusion of 'foreign' youth in SMEs in the context of recruitment of new apprentices: the industrial, market and domestic conventions. 7 For an outline of these conventions the subsequent paragraph refers to Diaz-Bone (2009, 242, 243).

Three 'worlds' to understand the hiring of apprentices in SME
The convention of the industrial world pursues the principle of medium and long-term planning of action for which resources are to be used productively. The flows of action are accordingly gauged in line with an efficient organisation of work. Accordingly, the actors' quality is assessed by the degree of productivity in their contribution to the manufacturing of products and services. Persons (and things) are worthy when they are efficient, productive and operational. Relationships can therefore 'be said to be harmonious when organised, measurable, functional, standardised' (Boltanski and Thevenot 1999, 373). In this industrial world, the quality of a job candidate can be assessed and measured on a scale of professional capabilities based on technical performance. New apprentices who are not yet professionally qualified are chosen on grounds of more general 'hard skills' and virtues like punctuality, orderliness or dependability. Hence, the selection of apprentices is very much governed by the recruiters expectation about future production sequences being supported rather than disrupted by the candidate's abilities. In this world, efficient performance is the main principle of equitable recruitment as well as of the company's justification of its personnel decision.
By contrast, the convention of the market world accentuates the relational principle of price and competition within a shorter time window. In this world, apprentices become 'costly' if they drop out from training with consequential costs to the company, or if they obstruct the sale of a product or service. The nature of the relationship between actors is strategic and co-ordination is aligned opportunistically to the market situation. If recruiters assess the school reports of candidates to anticipate future trouble at the vocational school, and thus the risk of dropping out from apprenticeship, they assess their worth on the ground of the market convention. Furthermore, apprentices adopt an additional value once they get in direct contact with the company's clientele. Through her or his appearance, language and manners, an apprentice may appeal more or less to customers and hence influence a company's customer retention. Customers may prefer not to deal with employees (e.g. sales people) from minority groups, who in turn become 'reified' and obtain a proper market value (equal to a commodity). Referring to Becker's (1971) concept of taste-based discrimination, social features such as ascribed social identity (gender and ethnicity, see Jenkins [1996], 141, 142) may well be of relevance in selecting personnel.
Finally, the needs and requirements of a company's domestic world codetermine who will be hired. The corresponding domestic convention of social coordination underlies interpersonal relations and expectations between co-workers, which in small enterprises are quite often in line with the traditional, patriarchally structured family. The differentiation of professional and personal relations is suspended in this world and a person is intimately engaged with a familiar surrounding at the workplace (Boltanski and Thevenot 1999, 362;Eymard-Duvernay 2008, 61). The domestic form of coordination generalises a trust and authority relationship, which pertains to family kinship, and which is directed to faithfulness and loyalty in the company. Contrary to the first two aforementioned conventions, the time perspective in the domestic convention tends to hinge on tradition and hence on the past. Accordingly, the social bond between people is based upon a generalisation of kinship and respect for tradition and lineage (Boltanski and Thevenot 1999, 370). Furthermore, this world is characterised by characteristics such as respect, dependency and by traditional gender roles and relations. 'It's a neighbouring space, polarised by the opposition between what is far or close, what is here or there, who are associates or strangers' (Thevenot 2001, 414). Company members may refer to the 'spirit of the house' and candidates are getting assessed on how they 'fit in' and how they may integrate into the same 'house' . Hence, somebody receives quality if trust is placed in a job candidate on the grounds of his or her 'social proximity' , circle of friends, relatives and personal relationships. The worth of a future apprentice is thus measured in terms of the person's trustworthiness, loyalty and 'character' -that is by his or her social acceptability (Jenkins 1984) -and by their evidenced capacity to accept domestic control structures, show respect for senior member and to recognise their authority, as well as the gender-specific division of labour.
The three worlds defined here generate multiple social orders at the workplace. Their (industrial, domestic, market) conventions mediate the adjustment of mutual expectations and requirements between employers, employees, customers and apprentices. They can be viewed as major principles of social relations in the context of smaller training companies guaranteeing the coordination between the training company and the (future) apprentice. Therefore they provide the basis for the evaluation and selection of new apprentices and the respective evaluation criteria of candidates can be considered as the main moral hiring principles. 8 As new apprentices have to comply with requirements of multiple (industrial, market and domestic) worlds, they basically represent a compromise which allows for local and temporal compatibility between different modes of company coordination and interaction (Thevenot 2001, 410, 416). The empirical study that follows has assessed if and how the different conventions of work coordination and employee relations affect how recruiters in SMEs use ethnic categories (such as the stigma of being 'foreign') in their organisational task to match jobs to applicants in the apprenticeship market. It is assumed that employers use, amongst others, ethnic characteristics as indicators in assessing apprenticeship candidates' qualities with respect to the central conventions underlying the organisation of work and training. Hence, the empirical question is how ethnic categories are interpreted in the apprenticeship allocation process and how this affects immigrant youth's placement.

Empirical research on ethnic discrimination in hiring apprentices
In the last two decades empirical analysis of ethnic discrimination in hiring apprentices has been absent until very recently in the German-speaking research. This article therefore draws on the one hand on older research from the 1980s on apprentice recruitment in Germany (Gaugler et al. 1978;Gillmeister, Kurthen, and Fijalkowski 1989;Schaub 1991;Janssen and Polat 2005) to illustrate the proposed model of discrimination in hiring apprentices. On the other hand, newer data from a Swiss study was used for a secondary analysis by the author (Imdorf 2010) and gave rise to the conventionalist model of hiring presented in this article. Finally, two recent studies (Kohlrausch 2013;and especially Scherr, Janz, and Müller 2015) that have assessed some of its claims are used to verify the model. Whereas the older German research was mostly of empirical nature and lacking a consistent theoretical framework to understand discrimination in hiring, the author's own study was originally informed by some decisive theoretical assumptions to direct the research process. Shifting the perspective from the applicant school leavers to the recruiting company aimed at understanding the recruitment of apprentices from a company's point of view rather than to problematise the endowment of youth with talent and skills. Thus a focus was placed on the main drivers of decision-making when SMEs appoint their apprenticeship places to school leavers.
A qualitative survey of SMEs based on semi-structured face-to-face expert interviews with recruiters allowed for a rich picture of both well-known but also unexpected organisational factors which matter when companies appoint new apprentices. The study's data collection in 2004/2005 focused on hiring processes in select industries: automobile painter (trained in paint shops), mechanic (in garages), carpenter (in joiner's workshops), dental and medical assistants (in dental and medical surgeries) and clerk (in SMEs' offices, multiple branches). The selection of specific industries was based on their numeric prevalence in the overall apprenticeship offers available to school leavers and on their composition of SME. Furthermore the chosen industries differ in their accessibility by nationality and gender. The sample consisted of 81 SMEs in German-speaking Switzerland, mostly micro companies (25 companies with one to nine employees, including apprentices) or small companies (31 companies with 10-49 employees). Three out of four interviewees were men, and nine out of ten interviewees were Swiss nationals (no comparative analysis of native and immigrant recruiters was conducted). Half of the interviewees were either owning or managing the company. The other half had managerial positions or they were responsible for the apprentice training.
To examine how applicant characteristics such as their other-ascribed ethnicity or nationality become relevant in the context of apprentice selection, recruiters were not directly asked about the significance of these categories. Rather, the interview schedule focused on organisational characteristics (such as needs and constraints associated with offering apprenticeships), the reconstruction of the recruitment and selection procedures, and the most important selection criteria. However, the respondents mostly started to talk about applications of 'foreigners' on their own. In the remaining cases, the interviewer triggered talk on the issue by asking for the share of native and minority youth in the applicant pool. As organisation-related arguments used by companies to justify their selection and to make sense of ethnic categories were of special interest, the interviewer tried to guide the interviewees into a mode of reasoning and argumentation once nationality-or ethnicity-related discourses were taken up. The audiotaped and transliterated interview texts were interpreted using argumentation analysis (Toulmin 1958).

Empirical findings on barriers for 'foreign' youth to access apprenticeship
As far as technical and manual aspects of industrial coordination in the company were concerned, the employers interviewed in the author's study did not anticipate major problems around candidates for apprenticeships they perceive as 'foreigners' . On the contrary, ambition, willingness to perform, studiousness and tenacity have been attributed by employers to 'foreign' apprentices since the 1980s (Gravalas and Braun 1982;Gillmeister, Kurthen, and Fijalkowski 1989;Schaub 1991;Janssen and Polat 2005). The latter have been described by some employers as even more disciplined, reliable and hardworking than their native young colleagues. Fears of employers that 'foreign' youth possessed insufficient language skills in order to successfully complete a training course, have already been considered arbitrary in the 1970s (Gaugler et al. 1978). Hence, from the employers' view, 'foreign' youth are not perceived as a risk to the functioning of the industrial world of training companies. This points to important requirements of both the market and the domestic world in the hiring process, two worlds where the quality of 'foreign youth' proved to be particularly contested.

Ethnic issues in the market world
Previous studies have pointed to the fact that companies attribute a higher risk to 'foreign' apprentices of leaving training or denying employment after training. During the 1970s and 1980s, German recruiters anticipated that the former would return to their home countries, and they expected an early marriage or pregnancy in young women (Gaugler et al. 1978;Gravalas and Braun 1982;Schaub 1991). In contrast and more recently, especially in this Swiss study, employers have primarily attributed school and language deficits that could provoke a failure in the vocational school to 'foreign' youth (an argument already present in the older research: Gravalas and Braun 1982, 125), even so their manual skills are not up to discussion (likewise highlighted in Schaub [1991], 31). While some employers anticipate drop out from training due to academic underachievement at the VET schools, others fear production losses during the training, because the students would need to be exempted from the work place for remedial courses. Both expectations relate to risk aversion in the market world of the company. Kohlrausch (2013) could confirm the relevance of the foreseeable completion of the VET training for the assessment of apprenticeship candidates.
Furthermore, the quality of 'foreign' apprentices in the market world has been put into question with regard to the advancement of customer ties, when apprentices get in direct contact with the company's clientele. Gaugler et al. (1978) have stressed the unsuitability of 'foreigners' for such posts that require 'intensive communication links with the regular German clients' . Gravalas and Braun (1982) as well as Schaub (1991) showed that customers could feel disturbed by the 'mentality' of 'foreigners' .
The more recent Swiss study has confirmed, although rather rarely, such reservations towards 'foreign' candidates. A dentist in private surgery, for example, had the gut feeling, 'that good, long standing Swiss patients could jump off because of [foreign] names' . The owner of a small travel agency revealed that she could never imagine having a girl in her office wearing a (Muslim) headscarf, 'even though she would be cute and neat and smart' . She knew from gossip that some of her customers 'really have something against such foreigners' . Scherr, Janz, and Müller (2015) verify the importance of customer expectations towards ethnic minorities when training companies recruit new apprentices.
Finally, several companies highlighted that hiring specific apprentices could destabilise the market world by restructuring the company's clientele in an undesirable way. A dentist worried that the employment of an Albanian apprentice could attract Albanian patients 'who have no appreciation of normal dental hygiene' . A garage owner devaluated potential clients from Albanian and Kosovan origin: 'Albanians and Kosovo, simply Yugoslavs in general, we don't have them among our Audi-customers. For, they are unbearable guys' . He further pointed to his second-hand car market, where we have them and where we grapple with them. Once we had one [as a worker] back at the car wash; and afterwards we had ten, twelve of them around the house. So, we said, we don't need that (…). Thus, we said generally not to take any foreigners anymore, well, first and foremost, no Kosovars. And ninety percent of the applications where the like.
The quotes point to an unexpected variation of the market logic: 'foreign' apprentices could destabilise a company's market through attraction of undesirable new customers. To prevent this, some companies avoid hiring youth of a corresponding ethnic background. This seems to be an efficient strategy for companies to constrain the market demand in their own interest, according to their own needs.

The stigma of being 'foreign': interpretations in the domestic world
Beyond ascriptions towards 'foreign' youth in the market world, the training companies' domestic world's requirements for social coordination seem to be crucial to understand discrimination in hiring which affects school leavers perceived as 'foreigners' . Both previous research as well as the Swiss study show that the quality of maintaining harmonious personal relations at the work place can be denied to 'foreigners' . As far as the empirical data from the Swiss study is concerned, claims of the domestic world have proved to be by far the most significant organisational arguments to justify the exclusion of 'foreign' youth from training.
The earlier studies on the allocation of apprenticeship places have already shown different nuances how training companies problematise 'foreign' youth due to various requirements of the domestic world. On the one hand employers talked of fears that staff could react with prejudices and resentments towards 'foreign' apprentices (Gillmeister, Kurthen, and Fijalkowski 1989), that 'foreigners' stay among themselves and face difficulties integrating in a German team (Schaub 1991;Janssen and Polat 2005), or that rivalries could erupt between different groups of 'foreigners' (Gillmeister, Kurthen, and Fijalkowski 1989). Both Scherr, Janz, and Müller (2015) and the Swiss study have confirmed such expectations. Some employers perceive 'foreign' apprentices to be misfits in an existing team. Exemplarily, a garage owner pointed to the risk that his staff could refuse to accommodate a 'foreign' apprentice: I wouldn't feel ashamed to have somebody here with another skin colour. However I have to accept the opinion of others (…). I'm not against the mix of peoples in principle. But my people back in our garage, for instance my mechanics … These are Swiss. I don't have foreigners.
Others anticipated social conflicts or workplace bullying between their employees once they recruit an apprentice from a different -perceived as inappropriate -national background. In the words of a dentist: Our aim is to be a team. I have to check what our surgery needs at the moment. I'm not xenophobic at all, but … This [a recent case of workplace bullying] has elicited very unpleasant tensions, and I said, in any case, I don't want to insert anybody from this group [Kosovo-Albanians] in the existing team any more. For I have seen new conflicts coming up. Thus, I said my main focus would really be a Swiss apprentice again.
In the Swiss study the preference for 'locals' turned out to be a very common argument as they seemed best to 'fit' in the pre-existing workforce. With 'locals' the interviewed recruiters referred to 'Swiss or the like' such as Italians, Spaniards or Germans. But the term always implied the exclusion of 'foreign youth' of non-European Union origin, especially nationals of the successor states of Yugoslavia or from Turkey. Additionally, some ethnic backgrounds -particularly Muslim, as further confirmed by Scherr, Janz, and Müller (2015) -were equated with attributes of 'alien mentality' or 'improper faith' , and therefore incompatible with the company's own culture. Such assumptions and attributions mainly refer to the organisational need for harmonious horizontal social relations. 9 While these previous arguments look towards ensuring a harmonious workforce, which facilitates greater manageability, other arguments point more directly to the employers' need for manageable apprentices. Schaub (1991) already reported that gender stereotypes of recruiters combined with ethnic prejudices affected Turkish young men's chances of being hired. The latter were expected to be rude to female supervisors, or to show too little respect for their trainers. Gender-culture related assumptions about ethnic milieus were also used in the interviews of the Swiss study to justify the exclusion of 'foreigners' . In line with Schaub's findings, and recently confirmed by Scherr, Janz, and Müller (2015), recruiters ascribed culturally determined macho-like behaviour (proudness, violence) to young male immigrants, especially those of Muslim faith (ibid.), which was seen as problematic in cases where such apprentices work under the supervision of a female. The Swiss study disclosed additional arguments why youth perceived as 'foreign' may trouble harmonious vertical social relations in the company. For instance, and in line with the findings of Scherr, Janz, and Müller (2015), recruiters avoided hiring too many members of the same minority group to prevent 'ethnic gathering' . Apprentices may elude social control once they start communicating in their native language. With regard to 'foreign youth' , recruiters finally expected a lack of family support in cases where their child faces difficult periods during the apprenticeship. This latter concern matters because companies prefer to rely on parents to ensure that the apprentice sticks to job and vocational school obligations even in the advent of problems.

Trouble avoidance and exclusion of 'foreign' youth in multiple worlds
The empirical findings from the Swiss study as well as from previous German studies illustrate well the important logic of organisational problem avoidance in multiple worlds when companies select apprentices. A Swiss dentist put it as follows: 'Well, one gets pragmatic. You just avoid any trouble' . Apprentices pose both a potential risk and a chance for the continuance of a company's diverse worlds and they need to contribute to the reproduction of the latter. Thus, to appoint a new apprentice means to assess and to avert the risk that he or she could perturb any of the respective worlds that matter. To allow for the company's further existence, each world tends to reproduce itself according to its importance for a specific company by claiming its own hiring principle of personnel evaluation and justice. As a consequence, any promising job candidate has to prove his/her industrial, market and domestic acceptability and suitability (Jenkins 1984) in the recruitment process. Hence, the recruited new apprentice represents a compromise of the main conventions of coordination which hold a specific company together.
Modelling the recruitment of apprentices from the perspective of such multiple organisational worlds that are closely intertwined in training companies allows us to interpret the problems 'foreign' youth faces in finding an apprenticeship place. A first important finding is that the candidates' nationality, or the stigma of being a foreigner adhering to specific national categories, proved hardly relevant for the functioning of the industrial world. Hence, the assumption that immigrant youth is less productive than their native colleagues, and that employers refuse to hire the former in consequence of statistical discrimination in the industrial world, can be rejected. Rather, the findings show a variety of presumptions and fears beyond hard skills towards those national groups employers perceive as foreigners. By hiring 'foreign' youth, employers consider it inevitable that the apprentices will cause organisational trouble in various worlds of the company. Scherr, Janz, and Müller (2015) have verified that ethnicity can serve as an indicator for trouble expectation in the company. This has been particularly evident with regard to the domestic world, where 'foreign' youth is expected to disturb horizontal and vertical social relations at the workplace. The stigma of being foreign has further proved significant in the market world, where 'foreign' youth is expected to be potentially costly due to risks of early drop out or adverse effects on customer ties.

Conclusion
The article draws on Boltanski and Thévenot's (2006) theory of justification to conceptualise apprentice selection and the resistance of training companies to hire 'foreign' youth from an interdisciplinary perspective. By highlighting different principles of social coordination in the company (i.e. social relations based on efficiency, domestic cohesion and cost avoidance) this approach permits a comprehensive theorising of hiring discrimination in work and training organisations. Hiring primarily follows a logic of 'suitability for the company' , which attempts at an early stage to shield the company from the risk of incurring difficulties with recruited apprentices (and employees).
Accordingly, any promising apprenticeship candidate has to prove his/her industrial, domestic and market quality and suitability in the recruitment process. Given the core values of the different 'worlds' of a small or medium sized training company, those applicants at risk of being refused are those who (1) prove to be inefficient, barely productive and badly organised in the industrial world; (2) costly in the market world as they could drop out from training or contest existing or desirable customer ties; or (3) cannot be trusted to smoothly integrate in the domestic world. Together these 'worlds' account for multiple social orders in small training companies, which allow for justifying the selection of new apprentices (and the exclusion of others) towards the own workforce and towards customers.
A conventionalist approach to hiring enables us to illuminate from a company's perspective the phenomena that are colloquially referred to as discrimination in job assignments. The recruitment of apprentices and of employees more generally generates 'discrimination' if the company selection process is not primarily driven by the industrial world's requirements for performance. The justification of industrial meritocracy is no longer given if the coordination principles and the expectations of the domestic and the market world are asserted, i.e. the motives of enhancing social company integration, cost avoidance and customer ties.
Referring to research on apprentice recruitment in Germany and Switzerland, the article reveals why 'foreign' youth face particular problems to access apprenticeship places in smaller companies: their problems are partially caused by training companies that expect to get into trouble by hiring them. When recruiters infer inferior qualities to 'foreign' candidates they refer to the requirement of social coordination in two specific worlds of their company. In the domestic world, training managers fear that the in-company integration of 'foreign' apprentices may fail. 'Foreign' trainees are considered to be rejected by the workforce, as well as to be managed with difficulties by their supervisors. Furthermore, applicants who are perceived as foreigners are expected to drop out from the apprenticeship or to irritate desirable customer ties. Both scenarios pose a risk to the company's well-being in the market world.
In contrast, the industrial convention proves to be of less systematic significance when small training companies make sense of applicant's national origin. Apparently the ethnic background and nationality of applicants (and the resulting potential stigma of being 'foreign') is not linked to organisational trouble-making due to inefficiency in the industrial world. Candidates' will to work, punctuality and other perceived work virtues seem to be weighed irrespective of it and 'foreign' youth are therefore accepted in the industrial world.
Future research should further explore the potential of the proposed framework to prevent the exclusion of minorities in training and employment. For instance, it is widely known that word-of-mouth recommendations play a crucial role in recruitment and selection, especially in small companies. A local recruitment approach through social networks enforces the domestic logic of selection (Eymard-Duvernay and Marchal 1997), as informal networks allow for more trustworthy information about candidates than do alternative channels to recruit young workers (Miller and Rosenbaum 1997). Recruitment channels differ in their implementation and weighting of specific orders of justification and they go along with specific values and constraints to judge job candidates (Eymard-Duvernay 2008). While trustbased recruitment through social networks gives way to domestic values, more 'objectified' anonymous recruitment procedures may accentuate industrial requirements. Recent research in Germany on apprentice selection and discrimination that draws on the proposed model of hiring discrimination indeed confirms some of its main assumptions (e.g. Kohlrausch 2013; Scherr, Janz, and Müller 2015).
The relatively homogeneous sample of small and private companies that were analysed in the Swiss study limits its explanatory power with regard to other types of work and training organisations. One can assume that work organisations vary in the way they handle their different coordination principles, resulting in different hiring practices and outcomes, which are more or less in favour of disadvantaged minorities. Comparative analysis on the level of companies as suggested by Thevenot (2001, 418) could be used in future research to test the assumption, that different organisational forms and origins of training companies (e.g. small vs. large training companies; single companies vs. training circles; rural vs. urban companies) go along with more or less (non-) discriminatory opportunity structures and recruitment practices (Petersen and Saporta 2004). For instance Hunkler (2014) provides evidence that there is only little room for hiring discrimination in a training circle with a highly professionalised recruitment department, and Seibert (2015) found that larger training companies in urban areas show more openness towards foreign apprentices than smaller companies in rural regions. Hence, the organisational approach proposed in this article may offer testable assumptions about how structural features of training organisations account for discrimination in hiring.
Finally, the proposed theoretical framework allows for drawing some preliminary policy implications. The theory of justification suggests that public disapproval of discriminatory staffing policy by either the buying public or the workforce can be effective to exert pressure on the company. Another policy might be to increase companies' awareness of the minority youth's productive potential by communicating positive experience of training companies with the respective apprentices, and by promoting company placements for disadvantaged youth. Last but not least, the reorganisation of dual VET in the form of training circles, which comes along with the outsourcing of the recruitment task to a professionalised entity, seems promising to reduce discrimination in hiring (Imdorf and Leemann 2012;Imdorf and Seiterle 2015). Notes 1. In this article the notion of school leavers refers to students who completed compulsory education (and not to students who dropped out from school). 2. Many Swiss training occupations can be combined with a Federal Vocational Baccalaureate (FVB), which enables access to Universities of Applied Sciences. However, an apprenticeship is a precondition to enrol in a FVB programme. 3. Public discourse in contemporary Switzerland concerning 'foreigners' is best described as a process of othering, denoting immigrants or their descendants as alien. The stigmatisation of being 'foreign' thereby depends on historical and regional contexts, and it occurs regardless of whether somebody is foreign-born or not. At present, youth referred to as 'foreigners' in Switzerland are mostly from families that emigrated from the former Yugoslavia and those with a Turkish background (Imdorf 2015). In the following, this article uses the mundane, nonscientific term 'foreign' in quotation marks to stress the social constructedness and ascribed nature of this category. 4. Furthermore, smaller companies may pre-select based on written applications. In case of apprenticeship candidates those contain ethnic information in the first and family name of the applicant, in the names of his/her parents, brothers and sisters, the place of birth, and last but not least, in the application photo, which is normally included and a common practice in Switzerland and Germany. 5. The subsequent description of the theoretical model of discrimination in hiring refers to a previous contribution of the author published in German (Imdorf 2015). 6. The following theoretical assumptions originate from diverse organisational theories such as the work of March (1994) on organisational decision-making, Luhmann's (2000) theory of organisation, as well as neo-institutionalist conceptions (Meyer and Rowan 1977;Scott 2001). 7. The remaining four conventions that do not significantly enhance our understanding of ethnic discrimination in placement with small and medium-sized companies are: the convention of fame, as companies do not search for renowned youth; the civic convention, as the recruitment of apprentices is not publicly regulated and SMEs do not need to legitimate themselves through civic engagement; the network convention of social coordination (asking for flexibility and mobility), which rather matters for the coordination of apprentices in large companies and in training circles (Imdorf and Leemann 2012), but not in SMEs; and the inspired convention (coordination based on inspiration), as minority youth do not rank behind native youth with regard to their vocation for the training occupations they apply for. 8. The different worlds, which fundamentally intertwine and mutually shape work organisations, are historical and social constructions (Boltanski and Thevenot 1999, 369). Rather than primarily essential for the economical survival of the company based on profit maximisation, the different orders of coordination and justification are first of all materialised cultural belief systems. 9. Scherr, Janz, and Müller (2015) show that explicit negative ascription mainly concerns subgroups of migrants. They point out that migration background can also be positively connoted in some cases. The Swiss study has indeed revealed that foreign language skills can be an asset for minority youth in some sectors, especially for Italian-and Spanish-speaking applicants in service-oriented businesses with clients of respective linguistic origins. However, the language skills of 'foreign youth' -Albanian, Turkish or South Slavic -were never appreciated even though many companies also had respective clientele. Research focusing on stereotypes about ethnic minority youth in specific industries, occupations, or companies of different size is currently missing.

Disclosure statement
No potential conflict of interest was reported by the author.