Cross-Cultural Turnover Research
In our review of the cross-cultural research on employee turnover and retention we highlight two areas: the generalizability of turnover constructs and models to non-US countries; and the retention of international workers.
Understanding turnover in non-US countries
Much of the cross-cultural research on turnover focuses on whether turnover constructs that are deemed important in the turnover decisions of American employees are also relevant in the turnover decisions made by employees elsewhere (e.g., Morrell, Loan-Clarke, Arnold & Wilkinson, 2008). Cross-cultural studies have also tested whether turnover models developed and validated in the US can be corroborated in other countries. The cross-cultural literature on turnover posits that the environment perceived by employees determines whether, and to what extent, certain turnover constructs are important in the decision to leave an organization. Additionally, traditional turnover models may not apply in certain cultural contexts because of differences in employee values and perspectives. For instance, whether employees are working in an individualistic or collectivistic context may determine the extent to which community embeddedness is important to their decision to leave (Ramesh & Gelfand, 2010).
These cross-cultural studies typically test for generalizability in a single non-US sample (e.g., China, Mexico, India; Allen et al., 2009; Harman et al., 2009; Maertz et al., 2003; Morrell et al., 2008; Peltokorpi, 2013; Peltokorpi, Allen & Froese, 2015; Wocke & Heymann, 2012) or by comparing the results of a US sample to a non-US sample of employees (e.g., U.S. vs. India; Ramesh & Gelfand, 2010; Wang, Lawler, Walumbwa & Shi, 2004). Very rarely do these studies employ multinational samples, but there are some examples (e.g., Sturman et al., 2012).
The turnover models that have been tested in non-US cultural contexts include general models of turnover decisions which incorporate constructs such as perceived job alternatives, job search and the perceived cost of leaving (Maertz, Stevens & Campion, 2003), the performance-turnover curvilinear model (Sturman, Shao & Katz, 2012), the cognitive withdrawal model (Allen et al., 2009), the job embeddedness model (Harman, Blum, Stefani & Taho, 2009; Peltokorpi, 2013; Tanova & Holtom, 2008) and the unfolding model (Morrell et al., 2008).
Some studies have tested whether particular constructs are relevant in the turnover decisions of non-US employees. Kim, Lee and Lee (2013) tested the effects of supervisor and co-worker relationships on turnover intentions in China and South Korea. Yingyan (2010) examined demographics and facets of satisfaction as predictors of intentions to stay among Chinese employees. Wocke and Heymann (2012) examined the role of shocks and demographics in the decision to leave among South African workers. Peltokorpi and colleagues (Peltokorpi, 2013; Peltokorpi et al., 2015) tested the effects of job embeddedness on turnover in Japanese samples. Lastly, Ramesh and Gelfand (2010) investigated the generalizability of various job embeddedness constructs (e.g., organizational links, community links, and family embeddedness) by comparing American and Indian employees.
Generally, research indicates that many of the turnover models developed in the US predict turnover in other contexts (e.g., Aladwan, Bhanugopan & Fish, 2013; Allen et al., 2003; Maertz et al., 2003; Ramesh & Gelfand, 2010; Sturman et al., 2012; Tanova & Holtom, 2008). However, some constructs emerge as more or less important in these other countries, and the details of some models change due to cultural and environmental factors. In one study that employed a multinational sample, Sturman and colleagues (2012) concluded that cultural factors across 24 countries affected the curvilinearity of the performance-turnover relationship. Regarding the decision to leave, Maertz and colleagues (2003) found that the influence of the family was quite pervasive in the turnover decisions of Mexican employees. Morrell and colleagues’ (2008) findings challenged the generalizability of the unfolding model, as most leavers could not be classified using the model with a sample of nurses in the UK. The authors attributed this to variations in the labour market for nurses there.
Several cross-cultural studies have focused on staying by testing the job embeddedness model. Peltokorpi (2013) concluded that distinct cultural factors affected on- and off- the-job embeddedness among Japanese employees, such as in-group ties. Additionally, Peltokorpi and colleagues (2015) found that risk-aversion moderated the relationship between organizational embeddedness and turnover intentions in a Japanese sample. Ramesh and Gelfand’s (2010) results showed that person-job fit predicted higher retention in the US (an individualistic culture), but person-organization fit, organizational links and community links predicted higher retention in India (a collectivistic culture). Harman and colleagues (2009) found that job embeddedness was not a predictor of turnover intentions over and above job satisfaction for Albanian employees, despite this relationship being quite robust in samples in the US. They concluded that the volatile employment situation in Albania (e.g., many new businesses moving in, many migrant workers moving to the cities) may be a factor.
In summary, although many of the turnover relationships and models that have been developed and validated in the US hold in other settings, there are nuances in these relationships that can be attributed to the cultural context and environmental influences. Future research should continue to examine the role of these situational factors in turnover decisions. More suggestions on research in this area are discussed in the section on future research.