You are on page 1of 13

The International Journal of Human Resource Management 10:3 June 1999 488 500

The coffee-machine system: how international selection really works

Hilary Harris and Chris Brewster


Abstract The literature on expatriate selection tends to present lists of criteria, with the implicit assumption that the process is formal and rational. The limited empirical work there is in this area suggests that these criteria have little impact on selection in practice. This paper argues that a deeper understanding of selection systems is needed and presents research showing the outcomes of different kinds of systems, suggesting that closed, informal systems predominate and outlining the practical and research implications of such systems. Keywords International selection; expatriate management; international manager competencies; IHRM; selection systems.

Introduction The rapid globalization of business has led to an ever-increasing need to `internationalize managers within organizations. Many large organizations now see international management experience as a prerequisite for promotion to senior management positions. International assignments ful l certain key purposes for both the organization and the individual. For the organization these include control and socialization in the prevailing culture and transfer of expertise and for the individual they are a major opportunity for personal development. These assignments do, however, present a considerable risk to the organization. In terms of cost, sending a manager on an international assignment can cost three times normal salary and on costs. Failure, de ned, not just in terms of premature return home, but as under-performance, can be, in its short- and long-term effect on business in the host country, very expensive. The extent to which selection practice matches this requirement is, however, questionable. Examination of the international manager-selection literature reveals a mismatch between theory and practice, with extensive lists of theoretical criteria relating to effective international managers, most of which resemble a cross-cultural `wish list in respect of the vast array of skills and abilities required. Selection processes are equally depicted as both formal and professional. How much the theory of international manager selection re ects actual practice has, however, been questioned (Brewster, 1991). This article reports the results of a study into international manager selection which queries the assumption that selection in this area is in practice a rational, objective process; outlines the implications of actual practice for effective international selecDr Hilary Harris, Cran eld School of Management, Cran eld, Bedford MK40 0AL, UK (tel: 1 44 (0) 1234 751122; fax: 1 44 (0) 1234 751806; e-mail: h.harris@cran eld.ac.uk). Professor Chris Brewster, Cran eld School of Management, Cran eld, Bedford MK40 0AL, UK (tel: 1 44 (0) 1234 751122; fax: 1 44 (0) 1234 751806; e-mail: c.j.brewster@cran eld.ac.uk). Copyright Routledge 1999 09585192

Harris and Brewster: The coffee-machine system

489

tion; and suggests a new research agenda for exploring the process of international selection. Selection of international managers We have focused discussion in this section on the selection of international managers. The selection of other types of expatriates technical specialists or short-term assignees for example will have many aspects of the selection systems in common. However, each type will have items which are distinct to ensure consistency we are examining just managerial appointments here; we draw conclusions about selection for other types of expatriates at the end. Selection systems In terms of selection systems, research evidence points to the use within companies of `pools of existing employees with high potential, from which future international managers are picked (BIC, 1991; Brewster, 1991). Initial assessment of `high potential would appear to be undertaken jointly by line managers and personnel specialists and often takes place at very early stages in an employee s career, sometimes involving selection of an elite corps straight from university or graduate business school. Sophistication of systems and de nition of high-potential employees vary widely between organizations. However, Business International Corporation s (1991) report Developing Effective Global Managers for the 1990s observes that assessments of high potential are usually based on: 1 the rm s needs, as de ned by business lines and strategic goals; 2 external market conditions; and 3 employee identi cation with the value system set by top management. The report also notes that, among most MNCs surveyed, `high potentials are proposed by their immediate supervisors and then screened by a committee composed of senior managers from the country of operation where the person works as well as of representatives of the parent-company s personnel department. Assessment centres are commonly used to assess initial potential, followed by annual reviews. Few companies explicitly tell employees they have been designated as `high potentials . First international postings for people identi ed in high-potential pools are often not implemented until the individuals have acquired a certain level of maturity and management experience within the domestic environment; typically the age range for rst postings would be late-20s to mid-30s. Selection criteria: the theory Research into selection criteria for international assignments shows a split between theory and practice. In surveys asking for general views on what makes effective international managers, the criteria mentioned as being critical differ from those reported as being used in practice. The literature on the criteria used for expatriate manager selection also has a tendency towards prescription and a heavy North American bias. There have been several reviews of this literature (Dowling et al., 1994; Mendenhall and Oddou, 1985). Phillips (1992) suggests that there is little or no difference between the personal qualities required for success in managing domestic or international business, but successful development of international business demands a higher level of skills and

490

The International Journal of Human Resource Management

qualities. This is because managers working abroad will be involved in a wider range of activities, roles and responsibilities than those required in the home market. Likewise, it has been suggested that the international manager has many characteristics of the effective manager operating in a less complex environment. The international manager, however, needs additional skills to reconcile the cultural problems created by the international environment. Possessing an awareness of the dif culties is not enough in this situation. This confusion may in some ways be related to the fact that the majority of the studies are not speci c in de ning the type of expatriate to whom such criteria relate (Tung, 1982). Equally, there is little consideration as to whether criteria will vary according to country, host-country role and number of expatriates employed (Bjorkman and Gersten, 1990; Tung, 1982; Torbiorn, 1982). Forster (1996), however, notes that the many different criteria presented in the literature tend to fall into three broad categories. These are technical competence at work; personality traits/attributes; and interpersonal social skills and personal and family situations. In his review of the literature relating to selection criteria, Brewster (1991) argues that successive authors have adapted previous categorizations of criteria, or developed new ones, so comparability between studies is limited. He notes that many of the American studies originate with the 1970 Business International list of fteen categories of skills for the international manager:
Experience, adaptability and exibility, technical knowledge of the business, competence, ability and past performance, managerial talent, language skills, potential, interest in overseas work, appreciation of new management and sensitivity, proper education, initiative and creativity, independence, good ability to communicate, maturity and emotional stability. (Business International, 1970)

A number of researchers have added to the list: for example, Hays (1974) emphasized the importance of `relational abilities and the `family situation ; Zeira and Banai (1985) introduced country of origin and `appearance (dress and looks) , the latter also discussed more recently by Stone (1991) as an unwritten but critical selection factor for women expatriates. Ashridge Management Research Centre s study of international managers (not necessarily expatriate managers) revealed clear agreement among respondents on attributes which appear desirable whatever the company s strategy. These include strategic awareness, adaptability in new situations, sensitivity to different cultures, ability to work in international teams, language skills and understanding international marketing (Barham and Devine, 1991). The list is signi cant in that four out of the top six characteristics identi ed represent more `soft skills, underlining the human relations aspect of international management and an ability to handle unfamiliar situations. Several authors have attempted to produce more discrete, wider categories. Rehfuss (1982), for example, identi es ve groups: `relational or interpersonal abilities; `cultural empathy (including motivation, language, maturity and an `x factor, operationally de ned as the ability to live abroad); technical skill; domestic performance; and spouse/family. Torbiorn (1982) identi es eight criteria: adaptability; language; motivation; level of education; social manners; family adaptability; medical status; status of the job. Mendenhall and Oddou (1985) identify three sets of individual cross-cultural skills as follows:

Harris and Brewster: The coffee-machine system

491

Self-ef cacy skills including reinforcement substitution, stress reduction and technical competence. Relational skills including relationship development, willingness to communicate, and language. Perceptual skills including understanding why host nationals behave and think in the way they do and making correct inferences as to the motives behind these behaviours. A key observation from this literature is the emphasis on interpersonal and crosscultural skills as determinants of success for international assignments. The stress on `soft skills re ects a more general departure from reliance on traditional `hard skills for successful management. Coulson-Thomas (1992) asked senior managers in ninety-one organizations to identify the qualities for effective international operation that are sought in members of a senior management team. Fourteen items were revealed. In order of preference, they were listed as: strategic awareness, customer focus, individual responsibility, communication skills, creativity, perspective, team player, objectivity, self-discipline, international awareness and perspective, breadth, transnational con dence and effectiveness, European awareness and perspective, language ability. It is interesting to note that, despite the imminent upgrading of the EU, these managers were still rating language ability as the least important characteristic of effective international operation. As a result of interviews with currently operating international managers, Barham and Wills (1992) identi ed a deeper, core competence, which is essentially holistic in nature, which underpins speci c behaviour competencies and skills. The authors labelled this a `being competence and split it into three interlinking parts: cognitive complexity, emotional energy and psychological maturity. Cognitive complexity refers to the ability to perceive several dimensions in a stimulus rather than only one (differentiation), as well as being able to identify multiple relationships among the differentiated characteristics (integration). Features of cognitive complexity include cultural empathy, active listening and a sense of humility. The second `being competence is emotional energy, which includes emotional self-awareness, emotional resilience and risk acceptance, together with the emotional support of the family. The nal `being competence relates to psychological maturity and represents a manager with a value system which helps them to formulate the dominant goals or themes which make their lives meaningful. Included in this competence is curiosity to learn, a `present orientation to time and personal morality. More recently, Birchall et al. (1995) de ned ten competencies related speci cally to the international manager s job. These were: global awareness, international strategy, international negotiation, international marketing, international nance, cultural empathy, addressing ethical dilemmas, building international teams, working with stakeholders and foreign language skills. Birchall et al. asked 102 respondents to rate the competencies on a scale of 15, with 5 indicating that the behaviour was of vital importance to successful performance overall. The top ve rated competencies as a result of this study were: international negotiation, global awareness, international strategy, international marketing and cultural empathy. Selection criteria: the practice Given the emphasis on interpersonal skills in management theory, it is somewhat surprising to nd evidence in the research into current practices of MNCs of the

492

The International Journal of Human Resource Management

continuing adoption of more traditional criteria for selection of expatriates. Brewster s 1988 survey of international personnel executives from European multinational corporations identi ed the following top criteria used to select expatriates: `technical expertise; language; family support; potential; knowing company systems; experience; marital status; medical status; independence and motivation . Likewise, in the Ashridge survey (198990), the top factors were identi ed as: `Technical skills/expertise for the job; potential; knowledge of company systems; understanding the market and customers; language; necessary part of career path; family support; knowledge/ understanding of culture good (Barham and Devine, 1991). (It should be noted that Brewster s survey asked for unprompted criteria, while the Ashridge survey indicated a set of criteria, with space for additional criteria and ranking.) These ndings appear to correlate with Brewster s (1991) observations concerning the two main ndings from research into selection practices among MNCs. The rst nding is that expatriates are primarily selected on the basis of their technical competence alone (Baliga and Baker, 1985; Harvey, 1985; Mendenhall et al., 1987; Miller, 1972; Tung, 1981; Zeira and Banai, 1984). The second nding is that there is an underlying assumption of the universal nature of managerial skills (Baker and Ivancevich, 1971). The reliance on technical competence may well relate to technical expatriates rather than managerial level expatriates. However, Miller s (1972) early explanation of companies preference for technical competence still appears to have relevance today. In his view, companies perception of international selection as a high-risk operation leads to a tendency to place too much emphasis in recruitment on technical and managerial quali cations, to ensure that the job can be done competently. This view appears to be supported to some extent by subsequent research highlighting the lack of agreement with respect to de ning successful performance for expatriate managers and evidence of the resulting general confusion relating to criteria relevant to ensuring success. Antal and lzraeli (1993) argue that, in the face of uncertainties about the role of expatriates, organizational need for certainty in this high-risk area leads managers to select others who are most similar to themselves and, consequently, presumably more likely to be seen as trustworthy and predictable. Research into actual selection procedures again tends to support Miller s early explanation. Brewster (1991) notes widespread reliance on personal recommendation for expatriate postings from either a specialist personnel staff member or line managers. This results in more or less predetermined selection interviews which consist more of negotiating the terms of the offer than determining the suitability of the candidate. Despite differences in research ndings, it would appear that the majority of organizations do not interview the spouse and/or family (Brewster, 1988; Tung, 1982; Bjorkman and Gertsen, 1990). Formal testing for potential expatriates is also limited, with restricted use of personality and psychological tests and a general suspicion of the validity of cultural awareness or adaptability tests (Brewster, 1988; Forster, 1996; Scullion, 1994; Tung, 1981, 1982; Bjorkman and Gertsen, 1990). Typology of international manager-selection systems We developed a typology of international manager-selection systems from the general literature on formalization of selection procedures and from the literature relating to expatriate management-selection practice, which identi es four possible variations of selection systems in the eld of international management.

Harris and Brewster: The coffee-machine system

493

The rst two variations relate to the nature of selection procedures. The expatriatemanagement literature identi es the use of both `open and `closed selection procedures in organizations. An `open system is one in which all vacancies are advertised and anyone with appropriate quali cations and experience may apply and candidates are interviewed with greater or lesser degrees of formalized testing. Selection decisions are taken by consensus among selectors. In contrast, a `closed system is one in which selectors at corporate headquarters choose, or nominate to line managers, `suitable candidates. In this situation, there may be only one manager involved in the selection process at head of ce. The candidate is informed only when agreement about acceptability has been reached between head-of ce personnel and the line manager. The selection interview in this process consists of a negotiation about the terms and conditions of the assignment. The second two variations of the selection process relate to the existence of formal and informal systems operating at organizational level. As has been discussed before, substantial evidence exists of the mediating effects on the formal organizational systems of informal mechanisms, leading to unintended outcomes with respect to stated organizational policy. In this way, four distinct categories of selection processes can be derived. A closed/informal system re ects a scenario in which individual preferences of selectors, which may be more or less unclear, will be allowed to determine who is seen to be acceptable due to the lack of in uence of formal systems, the lack of open debate about criteria and the lack of accountability engendered by the fact that employees are unaware that the process is happening. Under this typology, it is argued, the individual preferences of selectors may be inconsistent and incoherent in relation to identifying and assessing characteristics of effective international managers.

FORMAL OPEN c c c c c Clearly de ned criteria Clearly de ned measures Training for selectors Open advertising of vacancy (internal/external) Panel discussions c c c c c c c c c

INFORMAL Less de ned criteria Less de ned measures Limited training for selectors No panel discussions Open advertising of vacancy Recommendations Selectors individual preferences determine criteria and measures No panel discussions Nominations only (networking/reputation)

CLOSED c c c c c

Clearly de ned criteria Clearly de ned measures Training for selectors Panel discussions Nominations only (networking/reputation)

Figure 1 Typology of international manager-selection systems

494

The International Journal of Human Resource Management

In an open/informal system, although employees have access to vacancies, decisions as to who should be selected are usually arranged between relevant managers on the basis of personal recommendation and reputation and, although candidates may be put forward for interview, the selection decision is made before any formal interview takes place. In this scenario, the tendency for managers to select `clones of existing managers is increased. This type of system will see formal selection criteria agreed; however, the extent to which these and the match of candidates are debated and discussed may be limited. This will lead to decreases in consistency and in coherent thinking concerning the key characteristics of effective international managers and less attention to formal criteria. A closed/formal system involves selectors assessing candidates against formal criteria and discussing candidates match with them. However, the lack of personal contact with the candidate and the fact that the eld of potential applicants is determined by the selectors, with the attendant risk of omission of suitable candidates, may allow individual preferences of selectors to be re ected in the nominating process. The in uence of networking and reputation is a key feature of this type of system. Within an open/formal system, selectors assess candidates against formalized criteria and determine the best ` t through continual comparison of their own assessments against other selectors assessments, thus constraining the use of individual preferences and ensuring a questioning of assumptions. This will be re ected in more consistency in evaluations and greater clarity in thinking in relation to the critical components of effective international management. This type of system is also likely to produce a close match between individual selectors schemas of the ideal job-holder and formal selection criteria. The objectivity of this type of system will be enhanced by the use of psychometric and other tests. Exploring practice We explored the empirical behaviour in organizations selecting international managers through two separate procedures. First, we carried out interviews with the senior international HRM specialist in nine major `blue-chip international rms. These were all UK-based household name organizations, with extensive experience in international operations, amounting in all cases to more than half their total operations, and all with relatively large numbers of expatriates. They were in the air transport, petrochemicals, telecommunications, heavy engineering, pharmaceuticals, brewing and distribution and international aid sectors. All were operating in many countries including both First and Third World states. These were detailed interviews lasting over an hour and a half in all cases, and much longer in some, where we used a few open questions to encourage the interviewee to discuss expatriate policies, actual practice in recruitment and selection, and some of the advantages and problems of their current approach. We found that, without exception, the interviewees were very ready to talk about these issues and very open in describing the `warts and all reality of their organizational process. Second, we selected three of the organizations which seemed to us to stand at different points in our analytical framework for more detailed analysis. Two of the three organizations were situated at opposite ends of the closed/open continuum, with one representing a closed system and the other an open system. The other was positioned in the middle of the closed/open continuum. For the purposes of this paper, the three organizations are identi ed with the names Amstar (closed system), Cirus (open system) and Brymay (hybrid system).

Harris and Brewster: The coffee-machine system

495

The methodology employed for the second stage of the research consisted of a casestudy approach. Information about international manager-selection processes was collected from two sources: the rst through semi-structured interviews with HR personnel and key selectors within the individual organizations; the second via an examination of organizational literature in the form of policies and administration forms, etc. A critical part of the research was seen to be the identi cation of personally held beliefs about the characteristics of effective international managers. To try to ensure an unbiased summary of the characteristics of effective international managers (both on the researcher s part and that of the individual selectors), repertory grid technique was chosen as an integral part of the case-study design. This consisted of asking individual selectors to compare and contrast up to nine international managers whom they knew personally, split into categories of highly effective, moderately effective and not effective. As far as possible, women international managers were included in the sample.1 The results from the repertory grid interviews yielded a set of `constructs , or statements concerning effective/non-effective international manager behaviour which were further analysed using the grid analysis package (GAP) developed by Slater (1972). An explanation of how to interpret the results from this package is available from Smith (1986). In this article we draw on the research data and analysis identi ed in the rst two procedures. Findings Our detailed exploration of the nature of the international manager-selection process within the three organizations allowed them to be plotted onto the typology of international management selection systems. Amstar was placed in the closed/informal quadrant, Cirus was placed in the open/formal quadrant, while Brymay fell across the quadrants on the typology, with different systems being used for different appointments. The description of the system was seen to fall almost equally between the open and closed quadrant, but in terms of degree of formality it was argued that it fell more into the informal quadrant. This positioning was seen to indicate a very hybrid system in which there were real tensions between espoused formal policy and current organizational practice. The degree to which differences in selection processes resulted in the posited outcomes with respect to the use of selectors individual preferences in selection decision making was explored via the repertory grid analyses. These aimed to address the extent to which the type of selection process resulted in the posited outcomes with respect to clarity and consistency of thinking in relation to effective international managers and the degree to which the constructs derived from the repertory grid interviews with selectors matched formal company criteria. Degree of clarity, consistency and link with formal criteria Three key analyses were used. These were: the extent to which the grids depicted a clarity of thinking in relation to the characteristics of effective international managers; the degree of consistency both within and across individual selectors grids within each organization; and the degree to which the individual grids re ected formal selection criteria within the organization. Clarity of thinking The constructs emanating from the repertory grid analyses indicated that there was a large degree of difference in terms of sophistication of

496

The International Journal of Human Resource Management

thinking about characteristics of international managers among the selectors at Amstar: the only construct which is used consistently being mobile. The focus of the grids with respect to determinants of effective performance appears to be based more on determinants of effective domestic performance. In contrast, selectors in Cirus were seen to be clearer in their thinking about the characteristics of effective international managers. In discussion with staff at head of ce, the constructs which were seen to be speci cally related to the needs of working in an overseas environment were: linguist; effective communicator with regional groupings; good relations between eld and head of ce; good emissaries; understands responsibilities towards local environment and agencies; values different approaches to work in cross-cultural respect; culturally sensitive; strong interpersonal skills; secure with feedback inter-organization and more exible/adaptable. The analysis of clarity of thinking in relation to characteristics of effective international managers in Brymay revealed more of an emphasis on speci c international skills than was the case in Amstar, but a more limited scope of constructs than with Cirus. The constructs could be grouped around the themes of: con dence based on experience overseas; manages interface between centre and eld; more internationalist and cultural empathy. Most of the selectors, however, still rated effective business-management capabilities as the most important distinguishing factors between good and bad international managers. Consistency Consistency within grids was evaluated by a qualitative examination of the number of times a particular construct was used more than once in relation to the three sorts of effectiveness of international manager. Degree of consistency of thinking across individual selectors grids was assessed by a qualitative analysis of the constructs identi ed as most distinctive to the selector, to see to what extent the constructs arising were replicated across grids at an organizational level. Analyses within Amstar indicated a lack of consistency both within and across grids. There were wide variations in the constructs used to describe effective international managers by the same selector and comparisons across grids revealed no areas of commonality in terms of the constructs used to describe effective international managers. In contrast, the results from Cirus showed a high degree of consistency within grids, with the majority of selectors using a common set of criteria to describe the effective managers. The degree of consistency across grids was also seen to be quite high, with four common themes emerging: consultative management style; team management skills; linguistic/communication skills and professional management skills. In the case of Brymay, the degree of consistency of thinking within individual selectors grids re ected a mixed picture. Two of the grids showed little or no consistency in criteria related to effective international managers, while the other four showed slightly more consistency. Consistency of thinking across grids was limited. Only two common themes were detected in more than one grid. These were high energy and business sense. Match with formal criteria In terms of linkage with formal criteria, the results were in line with the general propositions from the typology. Amstar s closed/informal international management-selection system meant that it was not possible to identify formal criteria for the majority of selection cases. However, a key component of selection is the potential rating of the individual which is assessed as a result partly of the annual performance and development reviews. This is supported by the focus on determinants of effective domestic management behaviours revealed by the constructs elicited from selectors within Amstar. An assessment of the degree of coherence between individual selectors constructs and formal assessment criteria contained within

Harris and Brewster: The coffee-machine system

497

the annual performance and development review revealed very little correlation with the formal system in relation to competencies, but a greater degree of agreement in relation to appraisal qualities. In three out of the four grids, more than half of the formal competencies listed in the annual review of performance were not seen to be linked to the constructs relating to effective management behaviour, from either an international or a domestic perspective. More linkages could be detected between the grid constructs and the appraisal qualities, with all but one of the grids containing constructs which could be subsumed into the formal list of appraisal qualities. Three main qualities were seen to be important in all grids, these were: achievement motivation, capacity to motivate and helicopter vision. Cirus demonstrated a high degree of coherence with formal selection criteria. The analysis of the link of constructs to the formal criteria included in the person speci cation for eld director/programme manager positions showed that by far the majority of constructs tted within the of cial list of selection criteria. The linkage was seen to be extremely strong in relation to consultative, team management skills, with less priority ascribed to the very speci c international aptitudes such as awareness of the cultural, social and political environment. An analysis of the extent to which the constructs re ected formal criteria within Brymay was dif cult to assess due to the variance in usage of any formal job speci cations. The emphasis on business performance could be seen to relate to the way in which performance was measured in Brymay, where performance against key result areas (KRAs) was the main method of determining good or bad performance. The grid from the head of ce personnel manager showed more agreement of constructs with formal criteria set out on the of cial interview assessment form used in country manager selection, although this is hardly surprising given that the manager was involved in the creation of the form. Discussion One of our key ndings is that much of the literature fails to encompass the reality of expatriate selection. In many organizations the selection of the relevant expatriate falls under what we came to call the `coffee-machine system . It was this expression, coined by one cynical expatriate we spoke to, that we adopted to summarize what we found to be the most frequent approach to expatriate selection. The coffee-machine system is the most common form of expatriate selection. What happens is that a senior line manager is standing by the coffee machine when he (usually a man) is joined by a colleague:
`How s it going? `Oh, you know, overworked and underpaid. `Tell me about it. As well as all the usual stuff, Jimmy in Mombai has just fallen ill and is being own home. I ve got no idea who we can get over there to pick up the pieces at such short notice. It s driving me crazy. `Have you met that Simon on the fth oor? He s in the same line of work. Very bright and looks like going a long way. He was telling me that he and his wife had a great holiday in Goa a couple of years ago. He seems to like India. Could be worth a chat. `Hey, thanks. I ll check him out. `No problem. They don t seem able to improve this coffee though, do they?

What happens next is that the organization s processes are brought in to play to legitimize the decision that has, in effect, already been taken. Personnel les will be

498

The International Journal of Human Resource Management

scrutinized. Simon will probably have an informal discussion with the manager concerned and, if still interested, will be interviewed, but the interview will be more like a negotiation about the terms and conditions under which the job will be done, rather than what an external observer would recognize as a selection interview. There will not usually be any other candidates. Either just prior to the interview or once an agreement on Simon s transfer has been reached, the international human resources department will become involved, dealing with the nancial aspects, the physical transfer arrangements, the family issues, the ights and so on. Among the organizations we studied, this pattern is by far the most common form of selection. This is the reality of the `closed/informal cell in our model. We found a smaller number of organizations where this kind of selection occurred for some appointments whereas for others there was a more hybrid system, with an ostensibly `open/formal system being heavily in uenced by informal practices. We have so far identi ed only one organization where an `open/formal approach was consistently followed for all international management appointments. The implications of these forms of recruitment are very much as predicted in the model. First, the careful analysis of candidates against some list of ideal-type criteria sits rather uncomfortably with the reality of the coffee-machine system. These lists may have some theoretic value in prescribing what should be included, but they bear very little relationship to the criteria in the selectors minds during this process. Hence, in part, the fact that repeated research shows that, despite the prevalence of these lists, the major criterion applied in the practice of expatriate selection continues to be technical ability or current job performance. Second, the likelihood of the different selectors having quite different criteria in their minds is high. Among the three organizations where repertory grid techniques were applied to the selectors to identify the personal constructs which lay behind their selection decisions there was a clear gradation depending upon their selection system. In Amstar, where the coffee-machine system held sway, selectors evidenced not only very different assumptions about what made for a successful expatriate, but even individuals tended to have mixed and sometimes contradictory constructs. In Brymay, there was a greater coherence, but still marked differences in views. In Cirus, where all expatriate selection was undertaken in a formal and open manner, there was clear coherence and much greater consistency in thinking about the characteristics of effective international managers in relation to the speci c needs of that organization. Third, the manner in which the coffee-machine system acts to restrict the pool of potential candidates is obvious. Almost by de nition, the pool is limited to those subordinates well-known to the selector and the other managers with whom they come into contact. It is likely to involve only the technical specialists in their eld of work, even where that is irrelevant to the post under consideration. This is seen to be particularly problematic for women, given the fact that between 85 and 95 per cent of international managers are currently men. Within a selection context where the nature of the vacancies re ect a male-typed bias, there appears to be even more need for selection systems to ensure that potential `prejudice on the part of selectors is constrained by a process which forces them continually to question their assumptions about women s suitability and, critically, their acceptability in international management positions. The nature of such a system is also more likely to engender debate about the extent to which criteria for selection follow equal opportunity principles. Within such a system, individual preferences of selectors should be more consistent and coherent as a result of the constant discussion and debate in which the likelihood of equal opportunity issues being raised is increased. However, our research into selection

Harris and Brewster: The coffee-machine system

499

systems for international assignments points to a preponderance of systems where primarily subjective knowledge of an individual determines who is seen to ` t in best with the existing organizational norms (Brewster, 1991; Scullion, 1994). Fourth, the prescriptive advice to selectors which insists on involving the family in the selection decision is rather pre-empted by this approach. It is, of course, possible to check retrospectively, but still prior to the transfer, that the family is not going to create problems during the assignment. But this is a long way from what is generally proposed as good practice. Fifth, the ability of the organization to take a strategic view of expatriation is severely limited. Almost inevitably, the role of the international HR department in these cases becomes one, familiar to many researchers in the area, of dealing with the nancial, physical and social issues that fall out from the selection, rather than having an input into whether an expatriate assignment is required or advantageous, or what kind of assignment it could be for what kind of expatriate. The line managers involved have little overall responsibility for a strategic view of international assignments. As a result, the organization continues to take reactive decisions about expatriation. Finally, the model as proposed here needs further testing and development. On the evidence presented here, it helps us to understand what is happening in the reality of international organizational selection. However, it does not, as presented, take into account the size or sector of organizations, their relative stages of internationalization or the proportion of their business conducted internationally. The model has been tested only against a particular group of major, experienced, British MNEs. Wider testing may enable these and similar additional criteria to be included in the model, or to be applied as relevant antecedents of explanation of the model. Note
1 For an explanation of the method and applications of the repertory grid, see Bannister and Fransella (1986) and Stewart et al. (1991).

References
Ali, A.J. (1991) `Expatriate and Saudi Managers Perceptions of Japanese and U.S Competitiveness: A Survey , Business Horizons, 34(6): 358. Antal, A. and Izraeli, D. (1993) `A Global Comparison of Women in Management: Women Managers in Their Homelands and as Expatriates . In Fagerson, E. (ed.) Women in Management: Trends, Issues and Challenges in Managerial Diversity. Beverly Hills, CA: Sage. Baker, J. and Ivanevich, J. (1971) `The Assignment of American Executives Abroad: Systematic, Haphazard or Chaotic? , California Management Review, 13(3). Baliga, C. and Baker, J. (1985) `Multinational Corporate Policies for Expatriate Managers: Selection, Training and Evaluation , SAM Advanced Management Journal, 50(4): 318. Bannister, D and Fransella F. (1986) Inquiring Man:The Psychology of Personal Constructs. London: Croom Helm. Barham, K. and Devine, M. (1991) The Quest for the International Manager: A Survey of Global Human Resource Strategies . London: Ashridge Management Guide/Economist Intelligence Unit. Birchall, D., Hee, T. and Gay, K. (1996) `Competences for International Managers , Singapore Institute of Management, Jan.: 113. Bjorkman, I. and Gertsen, M. (1990) `Corporate Expatriation: An Analysis of Firms and Countryspeci c Differences in Scandinavia , unpublished paper quoted in Brewster (1991). Brewster, C. (1988) The Management of Expatriates , Human Resource Research Centre Monograph 2. Cran eld: Cran eld Institute of Technology.

500

The International Journal of Human Resource Management

Brewster, C. (1991) The Management of Expatriates . London: Kogan Page. Business International Corporation (BIC) (1991) Developing Effective Global Managers for the 1990s. New York. Coulson-Thomas, C. (1992) Creating the Global Manager. Maidenhead: McGraw Hill. Dowling, P.J., Schuler, R.S. and Welch, D. (1994) International Dimensions of Human Resource Management, 2nd ed. Boston: PWS-Kent. Forster, N. (1996) The Management of Expatriates: A Survey of UK-Based International Organisations. In press. Harvey, M. (1985) `The Expatriate Family: An Overlooked Variable in International Assignments , Columbia Journal of World Business, Spring: 8492. Hays, R. (1974) `Expatriate Selection: Insuring Success and Avoiding Failure , Journal of International Business Studies, 5: 2537. Mendenhall, M., Dunbar, E. and Oddou, C. (1987) `Expatriate Selection, Training and Career Pathing: A Review and Critique , Human Resource Management, 26(3): 33145. Mendenhall, M. and Oddou, G. (1995) `The Dimensions of Expatriate Acculturation: A Review , Academy of Management Review, 10: 3947. Miller, E. (1972) `The Selection Decision for an International Assignment: A Study of the Decision-Makers Behaviour , Journal of International Business Studies, 3: 4965. Phillips, N. (1992) Managing International Teams. London: Pitman. Rehfuss, J. (1982) `Management Development and the Selection of Overseas Executives , Personnel Administrator , 27(7): 3543. Scullion, H. (1994) `Staf ng Policies and Strategic Control in British Multinationals , International Studies of Management and Organization , 24(3): 86104. Slater, P. (1972) Notes on Ingrid and Grid Analysis Package. Manchester: University of Manchester Regional Computing Centre. Smith, J.M. (1986) `An Introduction to Repertory Grids Part Two: Interpretation of Results , Graduate Management Research, Autumn. Stewart, V., Stewart, A. and Fonda, N. (1981) Business Applications of Repertory Grid. Maidenhead: McGraw-Hill. Stone, R. (1991) `Expatriate Selection and Failure , Human Resource Planning, 41(1): 918. Torbiorn, I. (1982) Living Abroad: Personal Adjustment and Personnel Policy in the Overseas Setting. New York: Wiley. Tung, R. (1979) `US Multinationals: A Study of their Selection and Training for Overseas Assignments , Academy of Management Proceedings , 39: 298301. Tung, R. (1981) `Selection and Training of Personnel for Overseas Assignments , Columbia Journal, 23: 12943. Tung, R. (1982) `Selection and Training Procedures of US, European and Japanese Multinationals , Human Resource Management, 23: 12943. Wills, S. and Barham, K. (1994) `Being an International Manager , European Management Journal, 12(1): 4958. Zeira, Y. and Banai, M. (1984) `Present and Desired Methods of Selecting Expatriate Managers for International Assignments , Personnel Review, 13(3): 2935. Zeira, Y. and Banai, M. (1985) `Selection of Expatriate Managers in MNCs: The Host Environment Point of View , International Studies of Management and Organisation , 15(1): 3351.