Repatriation turnover revisited : A focus on South African multinational enterprises

Over the years, expatriation has been covered extensively in the literature with the challenges associated with foreign assignments and expatriate management receiving the lion’s share of the attention (Hill 2005:629; Stahl et al. 2009:91; Van Aswegen 2008:2; Vidal, Valle & Aragon 2008:1683). Recent studies, however, have directed the attention to the importance of expatriates and repatriates – ‘former international assignees who have returned to their domestic organisation after completing their international assignment’ (Berthoin Antal in Burmeister & Deller 2016:69) – as vehicles for knowledge transfer and organisational learning, the processes that constitute the foundation of organisational competiveness in a global economy (Lazarova & Cerdin 2007:405; Vidal, Valle & Aragon 2016:3). But with multinational enterprises (MNEs) experiencing problems in retaining employees after their return from international assignments (Goss & Hynes 2005:97; Lazarova & Cerdin 2007:404; Vidal et al. 2008:1683) many MNEs are not able to capitalise on this opportunity (O’Sullivan 2013:265; Shen & Hall 2009:793).


Introduction
Over the years, expatriation has been covered extensively in the literature with the challenges associated with foreign assignments and expatriate management receiving the lion's share of the attention (Hill 2005:629;Stahl et al. 2009:91;Van Aswegen 2008:2;Vidal, Valle & Aragon 2008:1683).Recent studies, however, have directed the attention to the importance of expatriates and repatriates -'former international assignees who have returned to their domestic organisation after completing their international assignment' (Berthoin Antal in Burmeister & Deller 2016:69) -as vehicles for knowledge transfer and organisational learning, the processes that constitute the foundation of organisational competiveness in a global economy (Lazarova & Cerdin 2007:405;Vidal, Valle & Aragon 2016:3).But with multinational enterprises (MNEs) experiencing problems in retaining employees after their return from international assignments (Goss & Hynes 2005:97;Lazarova & Cerdin 2007:404;Vidal et al. 2008Vidal et al. :1683) ) many MNEs are not able to capitalise on this opportunity (O'Sullivan 2013:265;Shen & Hall 2009:793).
To put this problem into context, according to Grant (1997:166), 15% of US repatriates leave their enterprise within 12-18 months after repatriation, increasing to 40% after 3 years.In a more recent study the Brooksfield Global Relocation Trends Report 2010 (in Kraimer, Shaffer, Harrison & Ren 2012:399) found that 38% of repatriated employees leave the employment of their MNE within the first year after repatriation.While the 2016 report found that 14% of MNEs saw an increase in their repatriation turnover, 12% saw a decrease and 75% saw no change (Brooksfield Global Relocation 2016).The cost of these failures can range from a loss of the initial development investment in employees when they take up employment with another MNE (Allen & Alvarez 1998:31;Caligiuri & Lazarova 2001 in Kraimer et al. 2012:399) to the inability to recruit high potential employees for future assignment as they see an international assignment leads to career derailment (Allen & Alvarez 1998:31;O'Sullivan 2013:265).

Background:
The topic of repatriation turnover as a major source of concern for repatriates and their multinational enterprise has been covered extensively in the literature over the years, with the literature showing that between 15% and 38% of repatriated expatriates leave the employment of their multinational enterprise within the first year after repatriation.However, no such study has focused on the repatriation of South African expatriates.

Aim:
The primary aim of this study was to determine if there is a correlation between the repatriation practices of South African multinational enterprises and their repatriation turnover rates.The secondary aim of the study was to determine why repatriated employees leave the employment of South African multinational enterprises. http://www.sajems.org Open Access South Africa is the second biggest emerging market source of foreign direct investment (FDI) into Africa (UNCTAD 2013) and ranked ninth overall with 28 FDI projects in Africa (Financial Times 2016:6).These figures are of particular concern given the fact that the availability of key skills is considered to be the greatest challenge facing CEOs in Africa.This while 88% of enterprises in PricewaterhouseCooper's (PWC's) 17 th Global CEO Survey indicated that the importance of global mobility has increased (PWC 2014).
Although the topic of repatriation turnover as a major source of concern for repatriates and their MNE has been covered extensively in the literature over the years (Lazarova & Cerdin 2007:404), no such studies have focused on the repatriation of South African expatriates (Vögel 2015).
However, in a study focusing on the preparation, support and training requirements of South African expatriates, it was found that discussing the expatriate's long-term career path prior to the international assignment and assigning a mentor to an expatriate were rated second and third respectively as the requirements with the greatest gap between what expatriates require from their MNE and what they received (Vögel, Van Vuuren & Millard 2008:36).Both of these have been identified as important elements in ensuring the successful repatriation of expatriates (Allen & Alvarez 1998:34-35; Ambrosius 2016:3-4).
Thus, given the increasing presence of South African MNEs in Africa, the shortage of key skills in the continent and the lack of knowledge of repatriation in South African MNEs, the primary aim of this study was to determine if there is a correlation between the repatriation practices of South African MNEs and their repatriation turnover rates.With a practice being defined as 'the customary, habitual, or expected procedure or way of doing something' (English Oxford Living Dictionaries 2017) and repatriation defined as the process of integrating expatriates back into the home enterprise (Allen & Alvarez 1998:30;Peng & Meyer 2011:497), repatriation practices can be defined as the customary, habitual or expected procedures followed by an MNE when integrating expatriates back into the home enterprise.The secondary aim of this study was to determine why repatriates leave the employment of South African MNEs once repatriated.

Literature review and hypotheses
When the repatriation process is not executed properly, uncertainty is created, dissatisfaction is experienced with the repatriation process and the repatriated employee can perceive the international assignment as career limiting, ultimately resulting in resignation (Lazarova & Cerdin 2007:406;MacDonald & Arthur 2003:3).This experience is referred to by Herman and Tetrick (2009:71) as career derailment.Furthermore, negative repatriates can influence the rest of the workforce considering international assignments, as top performers will not be interested in international career opportunities (Herman & Tetrick 2009:72;O'Sullivan 2013:265;Vidal et al. 2008Vidal et al. :1685)).Meanwhile poor performance and departure after repatriation cause both a heavy expense burden and business risks for an enterprise (Baruch, Steele & Quantrill 2002:660).
Over the years a number of factors have been identified as contributing to repatriation turnover, namely the financial shock related to the forfeiting of allowances, the psychological shock related to the loss of autonomy and authority, a lack of career advancement and reverse culture shock resulting from unrealistic expectations that nothing in the home country has changed (Goss n.d.:5-8;Herman & Tetrick 2009:69,72;Lazarova & Cerdin 2007:406;MacDonald & Arthur 2003:6;Paik, Segaud & Malinowski 2002:635).Lazarova and Cerdin (2007:405;O'Sullivan 2013:265), however, put forward a second argument explaining repatriation turnover, stating that repatriation turnover occurs not so much as a consequence of a lack of organisational attention and support as due to the availability of better employment opportunities outside the repatriate's current employer and the need for individuals to manage their own careers.
Repatriation, however, remains a costly expense in terms of time, money and human capital (Kraimer et al. 2009:30;Stahl et al. 2009:103).But in the case of repatriation turnover not only will the opportunity for reverse knowledge transfer be forfeited but it will also be detrimental to the MNE's leadership and development as well as affecting the global career plans in terms of the loss of a cadre of international employees available for future placement (Kraimer et al. 2009:30;Stahl et al. 2009:104;Vidal et al. 2008Vidal et al. :1684)).
This being said, no matter how sophisticated the career development and repatriation management processes are, turnover should be expected in some cases (O'Sullivan 2013:266;Stahl et al. 2009:105).
According to Martin and Anthony (2006:620), the repatriation process consists of four phases, namely pre-expatriation, during expatriation, prior to repatriation and during repatriation.In order to examine the repatriation process in more detail, these four phases were used as a broad structure for this study.

Phase 1: Prior to expatriation
The focus during this phase is creating and building support for the programme and selecting the right candidates for international assignments coupled with career planning and orientation prior to departure.A major component of this phase is also the creation of a formal repatriation programme (Martin & Anthony 2006:621).The Worldwide ERC Global Benchmarking survey has, however, found that only 51% of enterprises had a formal repatriation programme.In particular, this was the case in 87% of Asian enterprises compared to 45% of US enterprises (Tinder n.d.:2).
During the initial phase, career development is planned and goals for the assignment are communicated to the expatriate, while a mentor should be assigned to monitor and work with the expatriates, including informing the expatriate of future assignments.Mentorship prior to departure can form a basis of future interaction and determine the frequency of communication (Allen & Alvarez 1998:34;O'Sullivan 2013:267).
Initially, the focus should be on adjustment, but thereafter the focus should move to career development.The development of a repatriation education programme (i.e.exploring the issues, concerns, challenges and rewards of living abroad) should also form part of this phase (De Valk 2004:42;Martin & Anthony 2006:622).
Prior to departure, career expectations should be defined in terms of how the international assignment fits into the employee's career path and possible domestic job opportunities on return.Conversations about the employee's post-assignment job placement should commence before the international assignment and continue throughout the expatriate period (Ambrosius 2018:3-4;Martin & Anthony 2006:623).An important briefing during this phase is to emphasise that expatriates must be realistic about the postassignment, back in the home country.These sessions should be mandatory for all employees and family members (Tinder n.d.:7).
Partner orientation is also part of this phase and has a substantial influence on the relationship of the expatriate with the enterprise.Ensuring that the partner is properly trained to relocate can increase the possibility of the expatriate completing the assignment and staying with the enterprise on return (Martin & Anthony 2006:623).In considering a foreign assignment policy, it should be kept in mind that such a policy should be a set of guidelines for relocating persons around the world while maintaining the corporate culture, and not rigid rules (Greenberg 2009:18).
According to Martin and Anthony (2006:622), in those instances where senior management is involved in formulating the policy and not merely approving the policy, there is a higher occurrence of best practice activities.In those cases, it is more likely that there are career discussions with expatriates, that expatriates are given special attention in succession planning, mentorship programmes exist and, perhaps more importantly, there is much greater planning for an expatriate's return.Jassawalla, Connelly and Slojkowski (2004:41) are of the opinion that the presence and quality of foreign assignment policy guidelines seem strongly linked to repatriation effectiveness.Hence the first Hypotheses: H 1a : There is no correlation between the implementation of a foreign assignment policy and the repatriation turnover rates of South African multinational enterprises in year 1 after repatriation.

H 1b :
There is no correlation between the implementation of a foreign assignment policy and the repatriation turnover rates of South African multinational enterprises in year 2 after repatriation.
H 1c : There is no correlation between the implementation of a foreign assignment policy and the repatriation turnover rates of South African multinational enterprises in year 3 after repatriation.

Phase 2: During expatriation
The major components of this phase include: adjusting to the overseas environment, keeping up with organisational priorities and not being out of sight, out of mind, with communication being paramount (Greenberg 2009:18-19;Martin & Anthony 2006:623;O'Sullivan 2013:275-276).
Communication throughout the duration of the international assignment is considered to contribute to re-entry adjustment (Tinder n.d.:7).As a result, from a repatriation point of view, the focus of this phase should be on keeping the expatriate up to date with organisational changes.
During this phase, mentors can perform a critical role in updating expatriates on structural changes, revised policies and procedures and strategy updates (Allen & Alvarez 1998:34;Martin & Anthony 2006:624).In particular, the enterprise's objective in assigning a mentor is to establish a formal communications network involving the expatriate, the mentor and corporate human resources (Oddou 1991:306;Swaak 1997:30).
The so called 'out of sight, out of mind' syndrome can largely be attributed to the separation from the home country; not being involved can have disadvantages when important HR decisions are tabled and the expatriate is excluded from career-enhancing jobs.Mentorships can bridge this gap, but when mentors are reassigned, the loop of information is interrupted (Martin & Anthony 2006:624).The responsibility for keeping in contact is, however, not only the responsibility of the enterprise; the expatriate should also play a proactive role (Allen & Alvarez 1998:37-38;Tinder n.d.:7) by for example developing close relations with several other informal sponsors and developing rapport with several other managers who supervise positions considered to be attractive target repatriation positions (Allen & Alvarez 1998:37).As a result, the following Hypotheses were stated: There is no correlation between an expatriate manager perceiving assigning a mentor to a repatriate as being important and the repatriation turnover rates of South African multinational enterprises in year 1 after repatriation.

H 2b :
There is no correlation between an expatriate manager perceiving assigning a mentor to a repatriate as being important and the repatriation turnover rates of South African multinational enterprises in year 2 after repatriation.
http://www.sajems.orgOpen Access H 2c : There is no correlation between an expatriate manager perceiving assigning a mentor to a repatriate as being important and the repatriation turnover rates of South African multinational enterprises in year 3 after repatriation.

Phase 3: Prior to repatriation
During this phase the MNE should focus on creating realistic expectations, starting with repatriation counselling, conducting career planning and addressing issues that could lead to resignation (Kulkarni, Lengnick & Valk 2010:543;Martin & Anthony 2006:624).
De Valk (2004:42) stresses the importance of starting this process several months prior to the expatriate's return to his or her pre-assignment workplace.Keogh (2003:38) is of the opinion that although some MNEs start this process 2-6 months before the return of the expatriate, MNEs that effectively set goals start planning the return of expatriates even before the assignment starts.
As this phase focuses on creating realistic expectations, the intentions of the MNE should be clearly communicated to the expatriate.It is also essential to address family support as soon as possible during this phase.Enterprises should thus help to create these realistic expectations about the work and non-work lives before repatriation (Kulkarni et al. 2010:542;Martin & Anthony 2006:624-625).Programmes where expatriates and their families are invited to a 2-day seminar 6 months prior to repatriation can address issues like return adjustments, clarifying career objectives, expenses to be incurred on return, the elimination of allowances and lifestyle adjustment on return.(Martin & Anthony 2006:625;O'Sullivan 2002:598).Meanwhile Kulkarni et al. (2010:543) suggest sending the expatriate back to their home country prior to the official repatriation to see first-hand any changes before repatriation.
The role of the mentor during this phase continues to be important in keeping the expatriate informed, liaising with HR to finalise re-entry and conducting re-entry discussions (Conference Board Report in Martin & Anthony 2006:627;O'Sullivan 2002:598).As a result, it was hypothesised that: There is no correlation between when South African multinational enterprises start the repatriation process and their repatriation turnover rates in year 1 after repatriation.

H 3b :
There is no correlation between when South African multinational enterprises start the repatriation process and their repatriation turnover rates in year 2 after repatriation.

H 3c :
There is no correlation between when South African multinational enterprises start the repatriation process and their repatriation turnover rates in year 3 after repatriation.

H 4a :
There is no correlation between an expatriate manager perceiving the introduction of enterprise initiatives such as an orientation programme as being important prior to repatriation and the repatriation turnover rates of South African multinational enterprises in year 1 after repatriation.
H 4b : There is no correlation between an expatriate manager perceiving the introduction of enterprise initiatives such as an orientation programme as being important prior to repatriation and the repatriation turnover rates of South African multinational enterprises in year 2 after repatriation.
H 4c : There is no correlation between an expatriate manager perceiving the introduction of enterprise initiatives such as an orientation programme as being important prior to repatriation and the repatriation turnover rates of South African multinational enterprises in year 3 after repatriation.

Phase 4: During repatriation
The focus in the last phase of the process should be on reducing reverse culture shock, repatriation adjustment, managing expectations, welcome-home orientation and the repatriation role of the mentor.Using the newly acquired skills and valuing the international experience of the repatriate to the benefit of the enterprise is also critical during this phase (Janush 2003:52;Martin & Anthony 2006:626).
Reverse culture shock in particular is normally associated with re-entry adjustment, stress management and work performance (Martin & Anthony 2006:626).It is, however, in the enterprise's best interest to proactively address the repatriation phase of the assignment through training, mentoring and career planning (De Valk 2004:42).These proactive initiatives can include a re-entry seminar 1 month after repatriation (Tinder n.d.:9), counselling (preferably with an external consultant to encourage open and honest communication), assistance to partners seeking employment, customised sessions with children, and another follow-up 6 months later (Martin & Anthony 2006:627).
It should be kept in mind that re-entry adjustment is often more complex and more difficult than cultural adjustment when embarking on an international assignment with the duration of a foreign assignment affecting the re-entry adjustment.The longer the period of an international assignment, the more difficult the adjustment back into the home country will be (Martin & Anthony 2006:626;Vidal et al. 2008Vidal et al. :1685)).The time it takes for the repatriate to adjust back into the home country and new position can vary from 1 month to 3 years and should be managed by the enterprise as part of the re-entry adjustment (Janush 2003:52;Martin & Anthony 2006:627).
When looking at the management of expectations, the MNE should consider a number of factors such as expectations of expatriates to be placed in high-profile positions, applying their new skills and receiving the support and interest of colleagues as well as an acceptable living environment and manageable cost of living (Jassawalla et al. 2004:43;Martin & Anthony 2006:627).Repatriates themselves can address these expectations by scheduling meetings with their home country supervisor about their newly acquired skills and how these skills can be utilised on the job (De Valk 2004:42), as one of the great concerns of expatriates is that their knowledge should be put to use (Tinder n.d.:9).Based on the aforementioned discussion, it was hypothesised that: There is no correlation between an expatriate manager perceiving offering repatriation support to repatriates as being important during repatriation and the repatriation turnover rates of South African multinational enterprises in year 1 after repatriation.
H 5b : There is no correlation between an expatriate manager perceiving offering repatriation support to repatriates as being important during repatriation and the repatriation turnover rates of South African multinational enterprises in year 2 after repatriation.
H 5c : There is no correlation between an expatriate manager perceiving offering repatriation support to repatriates as being important during repatriation and the repatriation turnover rates of South African multinational enterprises in year 3 after repatriation.

Methodology Sample
With this quantitative study focusing on the repatriation practices of South African MNEs, the target population was all South African MNEs who employ expatriate staff.Consequently, the inclusion criteria used were that the MNEs had to be headquartered in South Africa and, secondly, that they had to employ expatriate staff.As a result of the specific inclusion criteria, coupled with the fact that the population size was difficult to ascertain, non-probability judgement sampling was used.
Next the unit of investigation was identified as the expatriate managers of these South African MNEs as they were deemed to be in the best position to provide information on the repatriation practices in their respective MNEs.In drawing the sample, two self-developed databases containing the contact details of South African MNEs were used.The first database contained the names and contact details of 170 South African MNEs with no indication of whether they employed expatriates or not.The second database contained the names and contact details of an additional 40 South African MNEs who employ expatriates.Next, the 170 MNEs on the first database were contacted telephonically to determine whether they made use of expatriates, and if so what the name and contact details of their expatriate manager were.The result of this process was that this database was reduced from 170 to 61 MNEs providing the researchers with a sample of 101 MNEs.The 101 expatriate managers at these MNEs were then sent a self-administered questionnaire electronically, with 41 expatriate managers completing the survey, translating into a 41% response rate.This response rate compares favourably with similar repatriation studies, with response rates ranging from 35% to 44% (Caligiuri & Tung 1999:769;Kraimer et al. 2009:33).
When looking at the sample characteristics, the wholesale and retail trade industry represented 27% of the 41 respondents, followed by the information and communication technology industry, with 15%, and the manufacturing and construction industries with 10% each.The enterprises in this study ranged in size from 465 employees to 70 000 employees, while the number of expatriates on assignment ranged from 3 to 6000.
Of the 41 respondents, 38 assign expatriates to African countries, followed by Europe, with 19 enterprises, Asia, with 17, South America with 11, North America with 9 and lastly Australasia with 8.

Instrument
The self-developed questionnaire contained both structured and unstructured questions.The first three questions ensured that the participant complied with the inclusion criteria and agreed to participate in the study voluntarily, followed by eight demographic questions.The third section of the instrument focused on Martin and Anthony's (2006) fourphase approach to repatriation, namely prior to expatriation, during expatriation, prior to repatriation and during repatriation.In particular, seven statements tested the importance of actions prior to expatriation, five statements tested the importance of assistance during expatriation, six statements tested the importance of assistance before repatriation and eight statements tested the importance of initiatives during repatriation.These statements were tested on a five-point Likert scale where the respondents had to indicate their level of agreement with the statements by choosing from the following options: 'Not at all', 'Slightly', 'Moderately', 'Very' and 'Extremely'.Finally, the last 10 questions focused on general repatriation questions, in particular the MNEs' repatriation turnover rates and the reasons for repatriates leaving the MNEs.The repatriation turnover rates were tested using an open-ended question and the reasons why repatriates leave the MNEs were tested using a multiple-choice, multiple-response question.The instrument was then pre-tested by five employees from a South African MNE with extensive expatriate and repatriation management knowledge.

Analysis
Due to the low number of responses in this study (n = 41), Cronbach's alpha was not used as the minimum sample size for coefficient alpha is commonly suggested at 200, 300 or 500 (Yürdugül 2008:403).An exploratory factor analysis prompts a similar debate regarding sample size.However, MacCallum et al. (1999:84-89) argue that the minimum sample size for an exploratory factor analysis is dependent on the level of communalities of item variables.Thus, due to the small sample size, an exploratory factor analysis was conducted on each of the four constructs focusing on the four phases of the repatriation process in order to determine if each represents a single construct.Eigenvalues greater than 1 determined the number of factors, while the loading for each variable was scrutinised to decide which variables to classify under each factor.Based on the eigenvalues, two factors emerged for actions prior to expatriation (construct 1), two factors for assistance during expatriation (construct 2), one factor for assistance prior to repatriation (construct 3) and two factors for initiatives during repatriation (construct 4).Each variable was classified under the factor in which it loaded the highest, except for construct 3, where only one factor emerged.No variables had to be eliminated due to double loading or low loadings (loadings smaller than 0.300).The results for each construct are subsequently discussed.

Actions prior to expatriation
Factor 1 was named 'preparation', as it represents the enterprise's actions to prepare the expatriate and family members for the foreign assignment.Factor 2 was named 'employee selection' as it involves the selection of the right candidates to fill the expatriate positions (see Table 1).The two factors account for 67.22% of the total variance in the data space (see Table 2).

Actions during expatriation
Table 3 shows the factor matrix with sorted rotated factor loadings, where Factor 1 was named 'expatriate support', as it summarises the enterprise's actions to assist with the adjustment process upon arrival as well as during the assignment.Factor 2 was named 'staying connected to the home country', as it involves all communication initiatives and home visits to prevent the 'out of sight, out of mind' syndrome.The two factors account for 69.14% of the total variance in the data space (see Table 4).

Actions prior to repatriation
Factor 1 consists of six variables and accounts for 61.21% of the total variance in the data space (see Table 5).
This factor was named 'assistance prior to repatriation' to align with the third phase of the employee repatriation process, namely 'actions prior to repatriation' (see Table 6).

Actions during repatriation
This construct was made up of two factors which cumulatively account for 75.35% of the total variance in the data space (see Table 7).Factor 1 was named 'repatriation support', as it combines the actions necessary to reintegrate the repatriate and his or her family back into the domestic environment.Factor 2 was named 'utilising and valuing international experience', highlighting the importance of assigning the repatriates to challenging positions to benefit from their newly acquired skills and to guide future assignees, if required (see Table 8).
Finally, Hypothesis 1 was tested using the Mann-Whitney U test, while Hypotheses 2-5 were tested using Spearman's correlation coefficient.

Findings Demographic profile
The majority of South African MNEs in this study assign their expatriates for periods ranging from 25 to 35 months (37%), followed by 13 to 24 months, 36 to 47 months and longer than 48 months, with 17% each.When looking at the age of South African expatriates, the majority are between the ages of 35 and 44, followed by 45 to 54.
Responses on the gender of expatriates showed that male expatriates (89%) still outnumber female expatriates (11%), a finding that supports an earlier study by Vögel et al. (2008:35).

Foreign assignment policy and repatriation turnover rates
Of the 41 MNEs that participated in this study, 31 (76%) had an implemented or operational foreign assignment policy.This was followed by eight (20%) whose foreign assignment policy was under development and two (5%) who felt that a foreign assignment policy was not required.
Based on the findings of the Mann-Whitney test (see Table 9) no statistically significant difference could be found in reported repatriation turnover rates in year 1, 2 and 3 after repatriation between enterprises that had implemented a foreign assignment policy and those that did not.
As a result none of the following three Hypotheses could be rejected: There is no correlation between the implementation of a foreign assignment policy and the repatriation turnover rates of South African multinational enterprises in year 1 after repatriation.

H 1b :
There is no correlation between the implementation of a foreign assignment policy and the repatriation turnover rates of South African multinational enterprises in year 2 after repatriation.

H 1c :
There is no correlation between the implementation of a foreign assignment policy and the repatriation turnover rates of South African multinational enterprises in year 3 after repatriation.

Mentors and repatriation turnover rates
A weak positive statistically significant correlation was found between an expatriate manager perceiving assigning a mentor to a repatriate as being important and the repatriation turnover rate in year 2 after repatriation (r = 0.342, p < 0.05) (Spearman's correlation coefficient), while no statistically significant correlation could be found for years 1 and 3 after repatriation (see Table 10).

Given these results, hypothesis H 2b , stating
There is no correlation between an expatriate manager perceiving assigning a mentor to a repatriate as being important and the repatriation  turnover rates of South African multinational enterprises in year two after repatriation, could not be accepted.
While Hypotheses H 2a and H 2c could not be rejected.
H 2a : There is no correlation between an expatriate manager perceiving assigning a mentor to a repatriate as being important and the repatriation turnover rates of South African multinational enterprises in year 1 after repatriation.

H 2c :
There is no correlation between an expatriate manager perceiving assigning a mentor to a repatriate as being important and the repatriation turnover rates of South African multinational enterprises in year 3 after repatriation.

Start of the repatriation process and repatriation turnover rates
When looking at the start of a repatriation process, it was found that 44% of the respondents commence with the repatriation process 6 months before repatriation.This is followed by 29% who start 2-5 months before repatriation, while 12% start 1 month before repatriation.Only 5% of the respondents start with the repatriation process more than 6 months before repatriation, while 5% provide no repatriation support to their returning expatriates at all and 5% start with the process less than 1 month before repatriation.
From Table 11 it can be seen that there is a moderate negative statistically significant correlation between when the enterprise starts with the repatriation process and the repatriate turnover rates in year 1 after repatriation (r = -0.361,p < 0.05) (Spearman's correlation coefficient).No statistically significant correlation could be found between the start of the repatriation process and repatriation turnover rates in years 2 and 3 after repatriation.
Consequently, hypothesis H 3a could not be accepted: There is no correlation between when South African multinational enterprises start with the repatriation process and their repatriation turnover rates in year 1 after repatriation.
While Hypotheses H 3b and H 3c , could not be rejected: There is no correlation between when South African multinational enterprises start with the repatriation process and their repatriation turnover rates in year 2 after repatriation.

H 3c :
There is no correlation between when South African multinational enterprises start with the repatriation process and their repatriation turnover rates in year 3 after repatriation.

Repatriation initiatives and repatriation turnover
A weak positive statistically significant correlation was found between an expatriate manager perceiving the introduction of enterprise initiatives such as an orientation programme as being important prior to repatriation and the repatriation turnover rates in year 1 after repatriation (r = 0.330, p < 0.05) (Spearman's correlation coefficient).A moderate positive statistically significant correlation was found between an expatriate manager perceiving the introduction of enterprise initiatives such as an orientation programme as being important prior to repatriation and the repatriation turnover rates in year 2 after repatriation (r = 0.434, p < 0.01) (Spearman's correlation coefficient) (see Table 12).No statistically significant correlation could be found, however, for year 3.
Based on these finding the following two Hypotheses could not be accepted: There is no correlation between an expatriate manager perceiving the introduction of enterprise initiatives such as an orientation programme as being important prior to repatriation and the repatriation turnover rates of South African multinational enterprises in year 1 after repatriation.
H 4b : There is no correlation between an expatriate manager perceiving the introduction of enterprise initiatives such as an orientation programme as being important prior to repatriation and the repatriation turnover rates of South African multinational enterprises in year 2 after repatriation.
Meanwhile hypothesis H 4c could not be rejected: There is no correlation between an expatriate manager perceiving the introduction of enterprise initiatives such as an orientation programme as being important prior to repatriation and the repatriation turnover rates of South African multinational enterprises in year 3 after repatriation.

Repatriation support and repatriation turnover rates
A moderately positive statistically significant correlation was found between an expatriate manager perceiving offering repatriation support to repatriates as being important during repatriation and the repatriation turnover rates of expatriates in year 2 after repatriation (r = 0.446, p < 0.01) (Spearman's correlation coefficient) (see Table 13).No statistically significant correlation could be found for years 1 and 3 after repatriation.
As a result hypothesis H 5b could not be accepted: There is no correlation between an expatriate manager perceiving offering repatriation support to repatriates as being important during repatriation and the repatriation turnover rates of South African multinational enterprises in year 2 after repatriation.
Hypotheses H 5a and H 5c , could not be rejected: There is no correlation between an expatriate manager perceiving offering repatriation support to repatriates as being important during repatriation and the repatriation turnover rates of South African multinational enterprises in year 1 after repatriation.
H 5c : There is no correlation between an expatriate manager perceiving offering repatriation support to repatriates as being important during repatriation and the repatriation turnover rates of South African multinational enterprises in year 3 after repatriation.

Reasons for repatriation turnover
Finally, Table 14 provides a summary of the most frequently cited reasons for repatriates leaving the employment of South African MNEs with a reduction in compensation once repatriated being the most cited.Two reasons came in joint second, namely the fact that their newly acquired skills are not utilised or valued by the MNE and the lack of challenging assignments on return home.

Discussion
With repatriation failure rates of 9% in year 1, 11% in year 2 and 13% in year 3, South African repatriation failure rates were found to be lower than those of other studies, where failure rates of between 23% and 25% were reported in year 1 after repatriation, increasing to 40% in year 3 (MacDonald & Arthur 2003:3;Tinder n.d.:2).
So what can we learn from these MNEs?As indicated by Martin and Anthony (2006:620), the repatriation of employees from an international assignment should start even before the expatriate departs for the assignment and should continue throughout the assignment up until the expatriate is successfully repatriated to his or her home country.However, according to Jassawalla et al. (2004:39), 'most foreign assignments seem hastily designed, signalling the poorly planned, ad hoc nature of foreign assignments from the expatriate manager's perspective', not only contributing towards uncertainty, but also anxiety among expatriates.The presence and quality of a foreign assignment policy, however, not only lessens this anxiety but also seems to contribute to the confidence with which expatriates leave for and return from foreign assignments, and as a result is strongly linked to repatriation effectiveness (Jassawalla et al. 2004:41).
Having said this, no statistically significant correlation could be found between having an implemented foreign assignment policy and the repatriation turnover rates of South African MNEs.This supports the view of Jassawalla et al. (2004:41) that not just the presence but also the quality of such a foreign assignment policy is important for repatriation success.
When looking at specific practices, Holtbrügge and Ambrosius (2015:279), O'Sullivan (2013:267) and Allen and Alvarez (1998:34) are of the opinion that appointing a mentor to an expatriate not only assists with the successful repatriation of an expatriate, but also supports an expatriate during the international assignment.In fact, mentors have a role to play during each phase of the international assignment (Anderson 2001:35;Oddou 1991:306;Scelba 1996:42;Swaak 1997:30).This was also found to be the case for the South African MNEs in this study with a weak positive statistically significant

Reason for leaving Frequency a Rank
Lost ground in their careers (e.g.colleagues have been promoted in their absence) 9 5 The fact that newly acquired skills are not utilised or valued by the enterprise  correlation being found between expatriate managers perceiving the appointment of a mentor to be important and reduced repatriate turnover rates in year 2 after repatriation.This shows that in MNEs where it is considered important to appoint a mentor to an expatriate, the repatriation turnover rates have been lower in year 2 after repatriation.Klaff (2002:42) however, is of the opinion that the mentor's association with the repatriate should continue for at least 6 months after the return home.Having said this, Vögel et al. (2008:36) in a study focusing on the preparation, support and training of South African expatriates found that South African MNEs are not providing the mentoring support required by their expatriates.This phenomenon is also supported by Pattie, White and Tansky (2010:367) who found that only 7% of enterprises provide a mentor for repatriates to guide them through the re-entry process.
Preparing for repatriation, Keogh (2003:38) is of the opinion that MNEs should start with a formal repatriation programme at least 6 months prior to the return of an expatriate to his or her home country.Such a programme -which should involve both the expatriate and his or her family -can address issues like clarifying career objectives, expenses to be incurred on return, the elimination of allowances and lifestyle adjustment on return, among other things (Martin & Anthony 2006:625;O'Sullivan 2002:598).This was supported by the findings of this study when a moderate negative statistically significant correlation was found between when the MNE starts with an orientation programme and the repatriate turnover rates of South African MNEs in year 1 after repatriation.In this study, 44% of MNEs heeded the call by Keogh (2003:38), starting with a repatriation programme 6 months in advance of the repatriation and 5% started even sooner.
A weak positive statistically significant correlation was also found between expatriate managers perceiving the introduction of repatriation initiatives such as an orientation programme to be important and repatriation turnover rates in year 1 after repatriation.Meanwhile a moderate positive statistically significant correlation was found between expatriate managers who perceive the introduction of initiatives such as an orientation programme to be important and repatriation turnover rates in year 2 after repatriation.Consequently, not only was it found that in South African MNEs where initiatives such as orientation programmes were considered to be important that repatriation turnover rates were lower, but also in those MNEs where such initiatives were started sooner repatriation turnover rates were lower in year 1 after repatriation.
These programmes benefit not only the MNE but also the returning expatriate (De Valk 2004:42;Keogh 2003:38).For the repatriate, proper advance planning and communication can help reduce anxiety and resentment and can ultimately have an impact on productivity and turnover rates (Vermond 2001:33), while for an enterprise, it can result in improved retention and return on human capital (Jassawalla et al. 2004:40).Financial changes affecting the repatriate's lifestyle was the primary reason cited by South African MNEs for repatriates leaving the enterprise, and a reduction in job status and autonomy in terms of responsibility and prominence in the community was the fourth most cited reason.Thus, addressing these and other issues prior to returning home are of the utmost importance.Having said this, in a study by Vidal et al. (2008Vidal et al. ( :1691)), it was found that for Spanish repatriates, the change in compensation was not considered to be a factor in their perception of the effectiveness of the repatriation process and consequently in their desire to leave the MNEs.
Lastly, the study found a moderately positive statistically significant correlation between South African MNEs who perceive offering repatriation support to repatriates during the final phase of the repatriation process to be important and repatriation turnover rates in year 2 after repatriation.This indicates that in those South African MNEs where repatriation support was considered to be important, repatriation turnover rates were lower in year 2 after repatriation.According to Allen and Alvarez (1998:37), these support initiatives can include creating a repatriate directory to assist the enterprise to utilise former expatriates for special assignments.By calling on a repatriate's international experience an MNE sends a clear signal to the repatriate that they are valued and appreciated.This is of particular importance given that not valuing and utilising their newly acquired skills and the lack of challenging assignments on return home were the joint second most reported reason for repatriates leaving the employment of South African MNEs.

Recommendations
This study contributes towards our understanding of repatriation in South African MNEs by showing that repatriation practices, such as the appointment of a mentor to assist the expatriate during each phase of the expatriation process, providing an orientation programme to assist expatriates with the repatriation process, starting with such a programme well in advance of their repatriation and providing repatriation support to expatriates during the repatriation process, can contribute to reduced repatriation turnover rates.
The study has, however, also identified that South African MNEs are still faced with repatriation turnover -albeit lower than reported cases in other countries.In particular, it has been determined that four of the eight reasons for repatriation turnover relate to job-related issues once repatriated.This highlights the importance of linking the successful completion of an international assignment to a specific position once repatriated.Preferably, this position should be communicated to an expatriate prior to the commencement of the assignment.
Although a positive correlation was found between providing initiatives such as orientation programmes to expatriates prior to their repatriation and lower repatriation turnover rates, addressing the negative effect of financial changes once repatriated needs to receive special attention during such a programme with this being the primary cause of repatriation turnover.Preparing an expatriate for reduced status back in the home office and country should also receive considerable attention.
Finally, although a correlation was found between providing repatriation support and lower repatriation turnover rates, feelings of isolation and rejection still persisted among repatriates.Consequently, repatriation support initiatives should focus on reducing these feeling of isolation and rejection.

Suggestions for future research
This study set out to determine if there was a correlation between the repatriation practices of South African MNEs and their repatriation turnover rates.In doing so a distinction was made between repatriation turnover rates in years 1, 2 and 3 after repatriation.In the process, correlations were identified in some years while not in others.The study did not, however, set out to determine why there might be discrepancies between the 3 years; as a result, future research should focus on determining: • Why there is a positive statistically significant correlation between an expatriate manager perceiving assigning a mentor to a repatriate as being important and the repatriation turnover rates in year 2, but not in years 1 and 3. • Why there is a negative statistically significant correlation between when the enterprise starts the repatriation process and the repatriation turnover rates in year 1, but not in years 2 and 3. • Why there is a positive statistically significant correlation between an expatriate manager perceiving the introduction of enterprise initiatives such as an orientation programme as being important prior to repatriation and the repatriation turnover rates in years 1 and 2, but not in year 3. • Why there is a positive statistically significant correlation between an expatriate manager perceiving offering repatriation support to repatriates as being important during repatriation and the repatriation turnover rates in year 2 but not in years 1 and 3.

Limitations of the study
A limitation of this study is the small sample size, as this limited the use of more advanced statistical analysis.The small number of respondents could possibly be attributed to the fact that only expatriate managers working for MNEs headquartered in South Africa, and not the repatriates themselves, were surveyed.
in job status and autonomy in terms of responsibility and prominent role in the community 18 4Feeling of isolation and rejection, as repatriates feel they no longer had the option to provide multiple answers.

TABLE 6 :
Factor matrix with sorted rotated factor loadings.

TABLE 5 :
Eigenvalue and variance explained.

TABLE 4 :
Eigenvalues and variance explained.

TABLE 3 :
Factor matrix with sorted rotated factor loadings.
a , All loadings less than 0.300 were indicated as 0.000

TABLE 2 :
Eigenvalues and variance explained.

TABLE 1 :
Factor matrix with sorted rotated factor loadings.
a , All loadings less than 0.300 were indicated as 0.000.

TABLE 9 :
Implemented foreign assignment policy and repatriation turnover rates.

TABLE 8 :
Factor matrix with sorted rotated factor loadings.
a , All loadings less than 0.300 were indicated as 0.000.

TABLE 11 :
Start of the repatriation process and repatriation turnover rates.

TABLE 10 :
Assigning a mentor and repatriation turnover rates.

TABLE 14 :
Reasons why expatriates leave after repatriation.

TABLE 13 :
Repatriation support and repatriation turnover rates.

TABLE 12 :
Repatriation initiatives and repatriation turnover.