The neglected state of organizational-level turnover studies in the Chinese context: a call for research

In this essay, the authors discuss the neglected state of organizational-level turnover research in the Chinese context. They provide a brief overview of the importance of turnover research in the organizational sciences, highlight the role of performance-related turnover rates research, and outline general theories and findings that appear in the Western and English-language literature. This evidence is compared with a dearth of studies using samples of Chinese organizations and in Chinese-language journals. They conclude by calling for additional theory and empirical studies on turnover rates.

major meta-analyses (Heavey et al., 2013;Park and Shaw, 2013). Theory tradition at the organizational level is also abundant; major, differential theories have been developed or brought to bear in economics, sociology, human resource management, and organizational behavior, to name a few Shaw, Gupta, et al., 2005).
In this editorial, we outline the importance of such organizational-level turnover rate research and note the striking dearth and major omission of this type of research in the Chinese context, and outline a call for contextual and industry studies in China to fill the gap.

The importance of turnover rate and organizational performance research
The organizational-level turnover literature is poly-theoretic; four main themes predominate . The first is a view of depletion via human capital or social capital losses. Per this perspective, higher turnover rates damage organizational performance by reducing the organization's ability to function effectively. The mechanisms between voluntary turnover and the ultimate financial performance of the organization are presumed to be workforce-related, viz., lower productivity and customer satisfaction as well as higher disruption, accidents and errors. A related view originated in Price's (1977) sociological perspective on turnover. Price's (1977) theorization takes a similar direction of thought, but instead presumes that turnover effects are not monotonic. Instead, he argues that the first instances of voluntary quits are the most damaging to performance (Shaw, Gupta, et al., 2005). In this view, a rise from zero turnover to moderate levels creates the most impactful loss of knowledge, skills, and abilities (Alexander et al., 1994;Shaw et al., 2013) and the most significant disruption-creating gaps in organizational communication networks (Shaw, Duffy, et al., 2005).
A third view originates in the organizational behavior literature and, perhaps, enjoys the most popularity, but the least empirical support, in terms of describing turnover's effects. Dalton and Todor (1979) and others lauded a functional view of turnover, arguing that moderate turnover rates bring new ideas and infuse the organization with stagnation-fighting energy from outside. In direct opposition to Price's (1977) arguments, the functional view presupposes that low-to-moderate turnover improves organizational functioning; Price (1977) suggests that these levels are the most damaging. In the inverted-U view, at higher turnover levels the negative depletion effects on performance will ultimately prevail. It is difficult to search the literature and find examples that support the inverted-U approach unequivocally; Glebbeek and Bax (2004) and Siebert and Zubanov (2009) offer some limited support (see, Shaw (2011), for a review). A fourth view is a classic contingency model that originated in the strategic HRM literature. Arthur (1994) argued and found that the relationship between turnover rates and performance was moderated by how much the organization had invested in its people (see also Guthrie (2001) and Shaw et al. (2013)). High investment organizations tend to perform better, but suffer greater productivity loss via turnover than organizations that invest little in their people.
In more recent years, researchers have attempted to synthesize and extend these founding theories by considering the role of replacements or hiring in assuaging the negative effects of turnover (e.g., Nyberg and Ployhart, 2013) and by factoring in the timing or clustering of turnover . These new advancements have spurred a spate of additional empirical testing and hastened accumulation the evidence (Call et al., 2015;Reilly et al., 2014). As noted by Shaw (2017), the morass of organizational-level theory may have created some initial confusion, but the upside to the situation was a vast amount of empirical testing of alternative formulations. Thus, in addition to the presumed practical value of examining these relationships, the empirical literature provided the fuel for two comprehensive meta-analytic studies (Heavey et al., 2013;, which confirm practically significant negative relationships between turnover rates and performance outcomes. These meta-analyses also show (1) little evidence to support other formulations other than practically important declines in performance as turnover rates rise, (2) few differences in turnover rate relationships across voluntary and total turnover operationalizations, and (3) some evidence that turnover rate effects differ across employment relationships and job types.

Organizational turnover research in the Chinese context-a missing link
The meta-analyses noted above were conducted, like most meta-analytic studies in Western journals, using available English-language studies. For example, Park and Shaw (2013) analyzed data from more than 100 papers and 300 sample-level correlations. Strikingly, only 4 of those studies included data from Chinese companies. This is remarkable underrepresentation given the increasing volume of research being produced in greater China and by Chinese scholars worldwide. In addition, given the vast Chinese economy, significant global economic influence, and strident economic growth in the past several decades, the lack of contextual research from China on this critical topic in the field of management is curious and troublesome.
In our preparation for this editorial we sought to dig deeper, hoping to undercover a trove of research that was overlooked in the English-language meta-analyses that would somehow balance the view. To do so, we searched the literature to identify organizational-or unit-level empirical studies on turnover rates that appeared in Chinese language journals in the last 10 years. What we found confirmed our initial conclusion that organizational-level turnover research in China is neglected. We first searched CNKI database to identify articles on turnover using broad turnover-related key words; this search yielded 2294 articles. We then perused these articles to identify empirical studies that examined organizational-or unit-level turnover rates. In the end, only 16 papers met our inclusion criteria, two of which were published in Chinese social science citation index (CSSCI) journals (Chen et al., 2016;Mei and Zhao, 2016); the remaining 14 papers were published in non-CSSCI journals or magazines. Nine out of 16 articles were guided by formal theories, such as agency theory, tournament theory, equity theory, and social identity theory. In terms of analysis strategy, regression analysis was used in 13 of these articles, and the remaining three articles simply descriptive statistics or content analysis. Table 1 shows the summary information of these articles.
The clear theme among these 16 studies is a focus on top management teams (TMTs). This contrasts with the large Western literature where the focus tends to be on turnover rates among core employee groups or on turnover rates in the entire organization. Firm performance, pay structure, and pay levels were the most-studied antecedents of TMT turnover in the Chinese papers. As examples, Cai and Li (2016) examined whether firm performance and incentives would explain top management turnover. Taking an agency theory view, the authors found that firm performance and stock incentive were negatively related to top management turnover rates, but pay level was not significantly related. Drawing on tournament theory and fairness theory, Deng et al. (2016) predicted and found among 376 top management teams from listed A-share companies that TMT pay dispersion increased turnover rates and, further, that pay level attenuated this positive relationship. A telling finding is that only the paper by He and Liu (2011) investigated the performance-related consequences of turnover rates using key or core employees. Building on the human capital literature, the authors predicted that employee turnover rates would be negatively related to organizational productivity. The authors sampled employees from 14 plants of a company covering the years 2007 to 2010 and found that total and involuntary turnover rates related positively to defective rates, defined as errors caused by manpower, machine, material, method, measurement, and environment. In addition, voluntary turnover rates related positively to the defective rate caused by manpower.

A call for research
The preceding review makes the purpose of this essay-a call for additional organizational-level turnover and performance research-clear. Our search efforts were meticulous, if not exhaustive. These endeavors uncovered only a handful of Chinese studies, while hundreds of such studies conducted in Western samples appear in the Western literature. The reasons for the dearth are unknown, but we can offer some speculation. First, it is possible that researchers believe that turnover consequences are not as severe in this context. The meta-analytic evidence does show that the turnover-organizational performance correlation is somewhat weaker in Asian samples  but it is difficult to place confidence in this comparison given the small number of Asian-sample studies. Second, it is possible that it is more difficult to collect or obtain such data in China compared to North America, where most of the studies originate. In the North American context, key-informant research designs are plentiful (Gupta and Shaw, 2000). In these studies, turnover data are typically obtained from key informants (e.g., plant managers or HRM managers) in large samples of organizations within or across industries. These turnover reports are then placed into empirical models with archival data obtained from industry associations, government databases, or purchased from third-party suppliers (e.g., Huselid, 1995;Shaw et al., 2009). Although individual-level survey data sets seem to be relatively easy to collect within Chinese organizations, organizational-level research conducted with a key informant design appears to be lacking. Third, strong industrial associations often characterize or bring together organizations in Western industries. Researchers are often able to access organizations and data via industry associations and trade conferences. These connections lead researchers to rich data sources underpinning years' worth of within-industry studies, for example, in automobile manufacturing (MacDuffie et al., 1996;Jacobides et al., 2016) and trucking (Shaw et al., 1998;Shaw, Gupta, et al., 2005;Shaw and Gupta, 2007), to name only two. It is possible that these associations function differently in China, making it more difficult for researchers to obtain the data and access needed for large organizational samples.
Whatever the root causes, our contention is that such research is potentially fruitful and may offer insights that are not apparent in the existing broader literature. First, there is certainly a need to continue to test alternative theories of the turnover-performance relationships that already appear in the literature. Indeed, as Shaw, Gupta, et al., (2005) and others have pointed out, tests of differential or competing theories are needed in the literature, as are attempts to integrate these alternative theories (see Nyberg and Ployhart, 2013;Shaw et al., 2013). Second, it is certainly possible that turnover-performance dynamics operate differently in Chinese organizations, leaving open the possibility for new theoretical development and empirical comparisons against existing conceptual foundations. Third, there are many clear differences in employment relationships, workforce management, and the labor market in China. These contextual factors could be brought to bear as antecedents of turnover rates, or as moderating factors between workforce outflow and performance. Fourth, Chinese organizations face some temporal and seasonal challenges that are unique. For example, there are severe, yet somewhat predictable, mass turnover events after the Lunar Year Holiday and/or after bonus payments. In addition, many organizations near metropolitan centers deal with frequent influx and outflow of migrant workers, which creates administrative, hiring, exit, and productivity-related challenges that are not seen in Western countries where the residential population and labor market are more stable.
To conclude, we hope this call challenges and energizes Chinese researchers to consider the value in turnover rate studies at the unit and organization level. Such endeavors will surely balance the scales of empirical evidence. Additional studies in the fascinating Chinese context will also likely lead to facets of insights that, despite its 100-year history, have yet to be observed in the turnover literature.