Performance appraisal in Western and local banks in China: the influence of firm ownership on the perceived importance of guanxi

Abstract The Chinese cultural factor guanxi (personal relationships) has an important influence on the implementation and effectiveness of HRM practices in China. However, few studies have investigated how guanxi influences performance appraisal in organisations with different ownership structures. We investigated this issue by studying three banks operating in mainland China: a state-owned bank, a foreign-owned bank and a city commercial bank. Both quantitative (survey of 308 employees) and qualitative data (22 in-depth interviews) were used. Macro-level convergence and micro-level divergence were found among the banks. The reform of the performance appraisal system in the state-owned bank indicated a degree of convergence towards a form commonly found in Western banks, at least on the surface. However, there were differences at the implementation level in the three banks. The perceived influence of guanxi on performance appraisal was strongest in the state-owned bank and weakest in the foreign-owned bank. The reasons for this included generational differences in cultural values, the international experience of employees and managers, and the degree of alignment between senior managers and others over the purpose of appraisal. The study demonstrates the mutability of cultural characteristics in different organisational settings.


Introduction
Since the opening up of the banking sector to foreign competition in 2006, Chinese banks have experienced wide-ranging reforms. These include changes in ownership structure, increases in management accountability and the introduction of human resource management (HRM) practices based on linking pay to performance. The drive behind these reforms was partly influenced by the policy ABSTRACT The Chinese cultural factor guanxi (personal relationships) has an important influence on the implementation and effectiveness of HRM practices in China. However, few studies have investigated how guanxi influences performance appraisal in organisations with different ownership structures. We investigated this issue by studying three banks operating in mainland China: a state-owned bank, a foreign-owned bank and a city commercial bank. Both quantitative (survey of 308 employees) and qualitative data (22 in-depth interviews) were used. Macro-level convergence and micro-level divergence were found among the banks. The reform of the performance appraisal system in the state-owned bank indicated a degree of convergence towards a form commonly found in Western banks, at least on the surface. However, there were differences at the implementation level in the three banks. The perceived influence of guanxi on performance appraisal was strongest in the state-owned bank and weakest in the foreign-owned bank. The reasons for this included generational differences in cultural values, the international experience of employees and managers, and the degree of alignment between senior managers and others over the purpose of appraisal. The study demonstrates the mutability of cultural characteristics in different organisational settings.
of 'linking up with the international track' (yu guoji jiegui), which emphasises the advantages of convergence towards international practices and technical standards which, it is argued, may help China develop a more efficient, competitive and performance-orientated economic and business system (Guthrie, 1998(Guthrie, , 2009Wang, 2007). Performance appraisal is seen as especially important for the motivation and development of the highly skilled knowledge workers needed for a successful and competitive banking sector (Joshi, Cahill, & Sidhu, 2010;Kuvaas, 2006). As such, state policies in China have been built on the assumption that performance appraisals may also help to improve firm performance in the domestic banking sector (Cousin, 2011;Jing, Cui, & Li, 2015).
That said, however, there is a large body of research which argues that Chinese cultural values may influence the implementation of HRM practices, which were developed in the West, leading to divergence and differentiation in their use, influence and outcomes (Cooke, 2013;Cooke, Saini, & Wang, 2014). In particular, there has been much interest in whether or not guanxi might make elements of Western HRM practice difficult to implement (Nolan, 2011(Nolan, , 2015. The term guanxi is used to describe many different facets of Chinese personal relationships. It may refer to a specific dyad, to ways in which the dyad is developed and maintained, or to the norms and expectations governing social networks more generally. It is assumed by many to be a culturally distinct form of social networking because of its roots in Confucianism (Bond & Hwang, 1986;Warner, 2014a). However, beyond the influence of Confucian values, guanxi can also be understood as the product of historical, political and economic factors which encouraged a preference for particularist ties. Some argue that China's gradual shift towards a rational-legal bureaucracy, based on the principle of open market competition, will necessitate a decline in the importance of guanxi (Guthrie, 1998(Guthrie, , 2009. Others suggest the reverse, noting that the turmoil associated with economic transition may create greater uncertainty and could enhance the need to rely on guanxi connections (Yang, 1994(Yang, , 2002. These arguments will be elaborated below, but it is clear that they link to our current concern with whether East Asian HRM practices might be converging towards those found in the West or whether they are diverging away from international practices and developing their own particular form. This paper aims to contribute to this debate by focusing specifically on the extent to which the perceived impact of guanxi on performance appraisal varies in banks with different ownership structures. The reason for this is that there is emerging evidence that ownership form may moderate the influence of cultural factors through a range of mechanisms, including organisational culture, management values and recruitment practices (see Warner, 2014a for a review). Our research sites include a state-owned commercial bank, a city bank and a foreign-owned bank. We use a quantitative survey to assess employees' perceptions of the influence of guanxi on performance appraisal and in-depth semi-structured interviews to explore why differences may exist between the banks. In the following section, we will introduce existing debates on the influence of institutions, cultures and business systems on HRM practices. We will then describe our methodology and present our findings. We will show how the degree of convergence in appraisal practices in each bank is influenced by cultural and generational differences, the international experience of both employees and managers, as well as institutional politics and conflicting interests between different groups.

The influence of institutions, cultures and business systems on implementing performance appraisal practices in China
Serious reform of HRM in the Chinese banking sector began in the mid-2000s. The changes implemented included adopting a more instrumental approach to appraisal based on linking pay and promotion to the achievement of 'objective' goals, such as profit or efficiency-based targets (Bai & Bennington, 2005;Zhao & Du, 2012). In relation to the pre-reform period, performance appraisals in state-owned enterprises (SOEs) are now generally more market orientated and, at an aggregate level, this appears to correlate with improved employee and firm performance (Rowley, Benson, & Warner, 2004). Akhtar, Ding, and Ge (2008) argue that in market-based economic systems, firm performance is increasingly linked to the quality of human capital. As such they suggest that in the coming years, there will be further convergence in Chinese HRM practices towards those commonly found in the West which are perceived by many observers to have had a positive impact on skills development (Cousin, 2011). However, other authors note that Chinese SOEs tend to face significant constraints in the implementation of results-based performance appraisal processes not least because of long-standing institutional factors, such as conflicting policy pressures from local government, central government and trade unions (Ngo, Lau, & Foley, 2008). Moreover, the skills and experiences of managers in SOEs tend to be somewhat limited and initiating change can be difficult when many employees and supervisors continue to identify with pre-reform norms and values (Ngo et al., 2008;Warner, Goodall, & Ding, 1999).
A further source of divergence in HRM in East Asia stems from differences in national-level regulations and business systems. The 'national business systems' approach draws attention to the way in which interlocking social, economic and political institutions create a distinctive way of organising economic activity (Whitley, 1992;Witt & Redding, 2013). For example, in terms of how multinational corporations (MNCs) operate globally, this perspective emphasises the tendency for institutions from the home country of the MNC to guide the HRM practices used by the firm in its international operations (Bae, Chen, & Lawler, 1998). However, these, in turn, may be modified by the different regulatory and political environments of various host countries (Almond et al., 2005;Varma, Budhwar, & Singh, 2015). In relation to our concerns here, we might expect to find that foreign MNC banks will face a different set of challenges in implementing HRM practices than locally owned Chinese banks. As Edwards and Kuruvilla (2005) note, far from being homogeneous in their aspirations and outlook, many organisational and political groups within China have conflicting interests. Central government officials, local government officials, trade unions, company directors, middle management and employees do not readily co-operate (Cooke, 2002;Ding & Warner, 2001) and it is likely that different groups may resist HRM reforms and attempt to defend existing privileges. Consequently, while government officials may make 'rational calculations' about the organisational changes they wish to see implemented in SOEs, these may be contested both within and outside the organisation. Jin and McDermott (2013) note that reform in SOEs requires an understanding of the strategic actions of senior managers acting as change agents within the organisation as well as their ability to handle conflicting interests between internal and external stakeholders. Likewise, Rowley and Bae (2004) have noted that there is a need to identify not just the level of change in HRM (e.g. reforms to policies, practices and day-to-day schemes), but also the degree of acceptance of these changes by key groups (such as employees, cadres and management).
As well as the varying influences of institutions and business systems on different kinds of organisations in China, there is a large literature arguing that any HRM practice developed in an Anglo-American context cannot be applied without adjustment in Chinese societies because of its distinctive, particularistic culture (Cooke, 2013;Hofstede, 2001). Culture itself is a contested concept and definitions vary depending on the epistemological standpoint of the author (Hofstede, 2001;Jack & Westwood, 2009;McSweeney, 2002). Nevertheless, there is much empirical research on the influence of what might be called 'Confucian' culture on HRM practices in Chinese and other East Asian societies (Warner, 2010;Warner, 2014a;Yan, 2010). Indeed, despite evidence from the business research literature on the enduring influence of Confucian culture in management practice, it is possible to observe significant changes in Chinese urban society which indicate that 'traditional' Confucian values are not as widespread as they might appear to be at first glance. These include changes in marriage and kin relations (such as later marriage for women and increased divorce rates), the development of a youth culture, disaffection with elements of the political system and rapidly growing consumerism (Hansen & Svarverud, 2010;Lin & Ho, 2009;Pan & Xu, 2015;Yan, 2009).
All of these debates are important for advancing our understanding of the relative importance of guanxi in performance appraisal. Different banks are subject to different constraints and enablers in the institutional environment and may subsequently attract and retain staff with different cultural and occupational values. We would expect a Chinese cultural characteristic, such as guanxi, to be present in each of the banks in our study. However our research question seeks to investigate the extent to which guanxi's influence on performance appraisal varies in banks with different ownership structures. We would predict that our quantitative data will demonstrate that the influence of guanxi on appraisal will be stronger in State Bank than in either City Bank or Foreign Bank. Being embedded in a particular institutional heritage, State Bank will find it more difficult to address the influence of guanxi on performance appraisal, both because of the legacy of certain recruitment practices and of a clash of values between policy-makers, supervisors and employees. The aim of our qualitative research is to investigate in detail both employee and supervisor perceptions of the appraisal process in each bank. We will focus specifically on the themes of institutional environment, organisational culture and Chinese cultural values, especially guanxi, in the practice and experience of appraisal.

Sample and access to research sites
Three banks were selected for study: a state-owned bank (State Bank), a foreign bank (Foreign Bank) and a city commercial bank (City Bank). The research was conducted at the regional headquarters of each bank in a city in eastern China. Initial access to the banks was gained through gatekeepers who were already known to the authors. The professional status of these individuals varied between the banks, but each gatekeeper was able to assist in gaining permission from senior managers to administer the survey to employees. Selection criteria for inclusion in the survey were that employees had been in the bank for more than a year and had experienced the performance appraisal cycle at least once, from the start (the setting of the objectives for the coming year) to the end (receiving a performance rating from the supervisor). All potential respondents were then invited to participate on a voluntary basis and anonymity and confidentiality were guaranteed. Following the survey, in-depth semi-structured interviews with employees and supervisors were arranged on an individual basis. Selection criteria for the participants in the in-depth interviews were that they had been employed at their respective banks for at least three years.
State Bank has a long history of operating in the mainland and has already experienced numerous internal operational changes associated with the strategic reform of the banking sector. All of the managers and employees in the Regional Headquarters were from mainland China (as were the Board of Directors) and around 8% of the workforce had international experience defined as having either studied or worked overseas. The average age of all employees was 37 years old. Foreign Bank is headquartered in Europe 1 and has a much shorter history of operating in the mainland than the State Bank. As Foreign Bank is a global multinational company, the nationalities of its Board of Directors is dominated by professionals from the US and Europe. Within the Regional Headquarters under investigation, approximately 5% of staff were officially non-Chinese nationals (though around half of these were born in the mainland and later changed nationalities). Around 60% of all employees in Foreign Bank had international experience. The average age of all employees was 28 years old. City Bank was included because it stood as a hybrid organisational form between State Bank and Foreign Bank. It was established post 2000 and had adopted a financial and efficiency-based performance appraisal system at the outset. Consequently, unlike State Bank, it was not trying to reform an organisational culture based on the norms associated with long-standing state ownership. However, unlike Foreign Bank, all of its owners, managers and employees were mainland Chinese and were thus fully immersed in the culture of guanxi as it operates in urban areas. The qualitative interviews collected in City Bank were used to inform the analytical framework of the study more generally and we also use the statistical data collected from the employee survey in City Bank for comparison with State Bank and Foreign Bank. However, in our qualitative analysis below, we concentrate principally on the differences between State Bank and Foreign Bank as these illustrate most clearly the institutional, organisational and cultural factors which may influence convergence or divergence in HRM practices. The final sample for the employee questionnaire was 308 employees, with 108 valid questionnaires returned from State Bank, 100 from Foreign Bank and 100 from City Bank. A brief description of the sample and data sources collected are shown in Table 1.

Research design, procedures and hypothesis
We adopted a two-step explanatory mixed-methods design based on the philosophical standpoint of pragmatism (Creswell, 2013;Johnson & Gray, 2010). Two methods were used for data collection: (i) a quantitative survey of employee's perceptions of the influence of guanxi on performance appraisal and (ii) in-depth semi-structured qualitative interviews with both employees and supervisors. In the quantitative survey, a 3-item scale measuring guanxi's impact on appraisal outcome (measured using a 6-point Likert scale) was embedded in a longer workplace questionnaire which also contained measures of perceptions of organisational justice and employee-supervisor communications. The scale was adapted from Bozionelos & Wang (2007). The Cronbach's alpha for the guanxi and appraisal scale is .82. The items in the scale were: • Performance ratings are often distributed based on employees' guanxi.
• It is more important to have guanxi than to perform well, if employees want to get a good performance rating. • Those employees who have guanxi will get better performance ratings than their peers, even if their performance is not better than their peers.
As discussed earlier, for cultural and institutional reasons, SOEs often experience greater difficulties in implementing HRM practices than enterprises with other ownership forms (Ding & Akhtar, 2001;Ngo et al., 2008), thus it was hypothesised that:

Employee perceptions that guanxi positively impacts on appraisal outcome will be significantly higher in State Bank than in City Bank and Foreign Bank
The in-depth semi-structured interviews were conducted after the questionnaire had been administered and were aimed at investigating why the specific quantitative results might have emerged. The interview schedule included structured introductory questions aimed at eliciting demographic information and specific technical details about the design and implementation of the appraisal process. These were followed by open-ended questions aimed at encouraging participants to share their perceptions of the workplace environment and their personal experiences of the appraisal process itself. The data from the interviews were analysed using a data encoding thematic analysis process whereby themes were searched and identified through 'careful reading and re-reading of the data' (Rice & Ezzy, 1999, p. 258). In this study, we adopted a hybrid approach of thematic analysis, which incorporated both the deductive technique, generating themes from previous research and prior theories (Crabtree & Miller, 1999), and the inductive method, generating new themes from the raw data (Boyatzis, 1998). Qualitative themes were not used simply to augment quantitative results through triangulation, or to be numerically codified through a process of content analysis. Rather, their purpose was to assist in the interpretation and explanation of observed statistical patterns (Creswell, 2013). This is a key strength of using the two-step explanatory mixed-methods approach.

Quantitative results
A one-way ANOVA was used to compare the mean scores between the three banks on the scale which measured the perceived link between guanxi and performance appraisal. The descriptive statistics are shown in Table 2. There was a significant effect for the perceived influence of guanxi on performance appraisal at the p < 000 level for the three banks [F(2, 305) = 19.33, p = .000] (see Tables 3 and 4). The results of post hoc comparison LSD test shows that guanxi's impact on appraisal outcome in State Bank is significantly stronger than the counterpart in Foreign Bank, MD = .72, p = .000 < .01. Guanxi's impact on appraisal outcome in State Bank is also significantly stronger than the counterpart in City Bank, MD = .42, p = .000 < .01. Guanxi's impact on appraisal outcome in Foreign Bank is significantly weaker than the counterpart in bank City Bank, MD = −.30, p = .011 < .05 (see Table 5). The results show that employee perceptions that guanxi has an impact on appraisal outcomes is highest in State Bank, followed by City Bank, followed by Foreign Bank (see Figure 1). Thus, the hypothesis was supported.

Qualitative results: insights from the in-depth interviews
The perceived influence of guanxi on appraisal practices is most apparent in the differences between State Bank and Foreign Bank. Therefore, we will focus on a comparison between the qualitative data in these two banks in the following section, identifying key explanatory themes.

The case of State Bank
The enduring influence of guanxi on performance appraisal Prior to the introduction of HRM reforms in 2006, the lack of a formalised appraisal process in State Bank had actively encouraged the use of instrumental guanxi practices. As one respondent said: If the employee had a good guanxi with the supervisor, for example he gave gifts to the supervisor on occasions, the supervisor would provide a good performance evaluation in exchange. Also, if the employee had a strong guanxi connection with some important person in the company, the supervisors would provide a good evaluation to him, in order to satisfy and develop guanxi with that important person for other purposes in the future.
It was this kind of behaviour which the Board of Directors at State Bank considered to be an impediment to firm performance and professional development and was a key reason for the initiation of HRM reforms. However, despite a high level of commitment to driving through reforms amongst the leadership team at Regional Headquarters, there were clear indicators that preferential treatment, based on employees' guanxi connections, still influenced the process. For example, respondents reported that the practice of 'guanxi recruitment' continued in State Bank. In the pre-reform period, State Bank had an explicit policy of recruiting the children of current employees and this was regarded as an important perk of the job. While the HR Director reported that this policy had been officially abolished in 2009, interviewees noted that a number of graduates were still being recruited into the bank if they had a guanxi connection with a senior manager. One employee described the recruitment of a colleague through this route:  Furthermore, interviewees noted that even if they themselves had been recruited through open competition, it was still necessary to develop guanxi relationships with line managers and senior managers within the bank. Indeed, employees mentioned that an individual's ability to develop guanxi is considered an important (if informal and undocumented) performance evaluation criterion, particularly amongst older supervisors. As such, employees actively sought to interact with managers outside of work by visiting their homes on special occasions and offering gifts. The organisational culture within State Bank was such that even new employees who were not well connected quickly adapted to these practices and sought to establish their own guanxi networks. As one young employee said: I did not have any guanxi when I first entered the bank. I found that the guanxi network is quite important here, so I have kept developing guanxi with two managers, the head of my workgroup and the head of the department. I understand this kind of investment will be beneficial to me in the future.

The consequences of performance appraisal: promotion, bonus and face
The reform of the performance appraisal system in State Bank was accompanied by a reform to the promotion system. An informant in the HR Department noted that prior to the reform, the normal procedure was that a manager nominated a preferred candidate for a position, without any formal assessment. The manager simply needed to provide a reason for the nomination which was often based on the strength of the personal connection between the employee and the manager. However, during the reform of the promotion system, the bank set up a 'scientific' promotion system based on quantitative talent assessment which included psychometric measures (such as personality and aptitude tests) as well as the numerical score achieved in the performance appraisal. As one respondent observed: ... if you have the ambition for future promotion, you should try your best to get higher ratings in appraisal. If you can get an 'excellent' or a 'merit' rating over several years, you have a better opportunity of being promoted to a managerial position.
Finally, appraisal outcomes are important for employees in State Bank because employees care about mianzi (or face). Due to a lack of confidentiality in the system, appraisal outcomes tend to become known by others in the workgroup and an employee can be humiliated by a poor appraisal outcome. As noted in the introduction, mianzi may interact with guanxi, especially when it comes to Confucian values of maintaining group harmonies (Chen, Chen, & Huang, 2013). Thus, as well as the material and instrumental concerns related to promotion and bonuses, employees in State Bank were strongly motivated to obtain better performance ratings because they wished to maintain mianzi. This, in turn, encouraged them to adapt to the guanxi practices of gift-giving and out-of-work socialising with supervisors as a means for advancing themselves within the guanxi network of the bank and for protecting mianzi.

Evidence of gradual attitudinal change within State Bank
Despite having the support of the top management team, the general perception of our informants was that the reform of the performance appraisal system had been met with considerable resistance. However, that said, there were some indicators of attitudinal change among at least some of the supervisors involved in appraisal. One appraiser in State Bank observed: The new appraisal system provides direction to us on how to make a proper evaluation for each employee ... now I can refuse to benefit those employees who have guanxi [within the bank] and I can use the new regulations as a good reason to explain my decision.
And an employee noted that he had observed change in other workgroups, though not his own: I know that the new appraisal system had a big impact on some supervisors. They strictly follow the new policy. I think the employees in their workgroup are lucky. However, with my supervisor, I don't think he has implemented the new policy. He still does what he did before in appraisals.
So for a few supervisors, the reform of the appraisal system in State Bank is a welcome measure. Their adoption of the new system may, over the longer term, create norm diffusion throughout the organisation as a whole, leading eventually to a decline in the influence of guanxi on appraisal. If so, this would echo Guthrie's (1998) idea that the development of a rational-legal system based on codified regulations and performance-based reward may gradually lead to a decrease in the influence of guanxi. For the time being, however, the information from our respondents indicates that while there may be surface-level change in HRM systems (indicating a degree of convergence at the macro level with international 'best practice'), guanxi recruitment, bias in the appraisal process and the tendency for employees' to actively develop guanxi with managers are lasting features of the organisation. The complicated guanxi network within State Bank is resistant to change and, according to our respondents, has an enduring impact on most supervisors' decision-making in performance appraisals.

Procedural transparency reduces the influence of guanxi on performance appraisal
As shown in the survey results, employee perceptions that guanxi influenced performance appraisal outcomes were lower in Foreign Bank than in both State Bank and City Bank. Foreign Bank is a global multinational firm that implements a 'best practices' approach to HRM throughout its operations. It began trading on the mainland with a formal target-based appraisal system in place. From the information gained from both company documents and our informants, it appears that the appraisal system itself showed strong 'home country' effects and had not been adapted for use in Chinese societies, either formally or informally. The performance appraisal objectives matched those used elsewhere in the firm, namely a strong focus on performance targets which related to achieving financial goals and specialist skills development. Furthermore, appraisal procedures were made clear to all employees during induction and the process itself was subject to regular internal checks by senior managers outside the particular workgroup. Interviewees noted that the emphasis on financial performance and an awareness that procedures were monitored tended to restrict and reduce the influence of guanxi throughout the organisation. As one employee said: The decision-making process for performance appraisal in my bank is transparent ... I don't think guanxi will have a big impact on my supervisor's decision in the appraisal. The Department Head checks the results for every employee in every workgroup in the Department. The Head can clearly see each employee's performance in terms of every criterion in the report. This helps to reduce bias in the appraisal result.

Emphasis on performance and resistance to guanxi practices by supervisors
In each of the in-depth interviews, employees from Foreign Bank noted that guanxi does not, in their experience, have a particularly strong influence on supervisors' decision-making in performance appraisal. Employees attributed this to the performance-oriented culture of Foreign Bank which ensured that supervisors and employees generally focused on quantifiable results. Furthermore, participants noted that it is not common for employees to have informal interactions with their supervisors, such as having dinner together or engaging in gift-giving. This means that there is no real opportunity for employees to try and develop guanxi with supervisors. Interviews with the managers provided further support for this. They noted that the highly competitive and performance-oriented culture at Foreign Bank strongly influences their decision-making in appraisals, and that any attempt by the employee to develop guanxi is viewed as 'annoying' and counter-productive. An employee stated that: ... guanxi has very little effect on my supervisor's appraisal decision. The employees' 'real' performance matters a lot in his distribution of performance ratings. Guanxi is not so important here.

The consequences of performance appraisal: bonus and promotions
As with State Bank, the consequences of performance appraisal were improvements in promotion prospects and higher bonuses, however, the maintenance of face (mianzi) did not emerge as an important issue amongst employees in Foreign Bank. In terms of bonuses, our informants all noted that the outcome of the performance appraisal was directly linked to the employees' annual bonus, and that managers allocated the annual bonus based only on employees' ratings. Those with the highest ratings receive the highest bonuses. This contrasts with State Bank where managers consider other factors such as the guanxi connections of the employee and their seniority in the bank. Moreover, as one supervisor observed, because their own yearly bonuses are linked solely to the performance of the workgroup they manage, almost all supervisors are performance oriented. They do not tend to make decisions based on employees' guanxi connections in appraisal because, to make sure of an excellent team performance overall, they need to have high-performing and productive employees in their own group. Finally, the performance appraisal reports play a key role in the promotion system at Foreign Bank. Interviews with members of the HR team revealed that in order to be considered for career progression, employees' performance appraisal ratings need to be consistently high over several years. Clearly then, achieving good scores in a performance appraisal is important to an employees' career progression and income. Our informants were satisfied, however, that the appraisal system was generally fair and transparent, and that they would usually be judged on merit, not their guanxi connections.

Evidence for convergence
The clearest evidence of convergence towards Western performance appraisal process can be identified in the practices at the Regional Headquarters of Foreign Bank. One key reason for this was that there appeared to be a relatively clear alignment between the goals of performance appraisal, as advocated by the senior management team, and the attitudes and behaviours of the supervisors and employees actually involved in the process (Rowley & Bae, 2004). Other studies have shown that MNCs generally find it somewhat easier to implement Western-inspired HRM practices in China than SOEs (Cooke, 2013;Zhu, Warner, & Zhao, 2011). One reason for this is that the HR managers in MNCs are usually expatriates or local Chinese who have received overseas management education. These managers have been exposed to the 'best practices' approach and have international experience, often in the home country of the MNC, which they then bring into the organisation (Akhtar et al., 2008;Warner, 1993). In our case study, around 60% of the employees in the bank had international experience. We would argue that the educational and professional background of both the senior management team and the labour force in Foreign Bank created a form of normative isomorphism, or convergence, towards home country practices in the appraisal process (Edwards & Kuruvilla, 2005).
Moreover, employees at Foreign Bank are generally younger than those at State Bank (Table 1). As noted earlier, there is an indication that for some of China's younger generations (especially those based in urban areas with graduate education, professional aspirations and international experience), there may be some resistance to the use of 'traditional' cultural practices such as guanxi in the workplace (Lin & Ho, 2009;Yan, 2009). For this group, a reward system based broadly on merit may be more appealing than one which requires the cultivation of guanxi. They may feel that they are more likely to succeed in an MNC because there may be a closer alignment between their own desire for professional recognition and a company where the appraisal system generally does work to deliver a path for progression based on merit (Rowley & Bae, 2004). Of course, it may also be the case that those without guanxi tend to gravitate towards professional careers in MNCs because, by their very lack of connections, they are excluded from equivalent jobs in SOEs. Either way, our findings indicate a degree of convergence in the appraisal system used at the regional headquarters of Foreign Bank with those found in the country of origin, both in terms of basic system design and everyday practical application.
There is also some indication of convergence in State Bank, namely the strategic decision by the Board of Directors to implement a performance-based appraisal system in the Bank in the first place. This could be viewed as an attempt by officials to modify an element of the national business system by shifting assessment criteria in former SOEs towards improving firm performance as opposed to achieving social policy goals (such as providing secure, long-term employment) (Akhtar et al., 2008). The Board of Directors specifically instructed and incentivised the managers in the regional headquarters to initiate reforms to the performance appraisal system (though as our interviews have shown, this was not an easy task for managers). While some of our participants had observed compliance with the new system by a few supervisors, they were perceived to be a minority group. Generally speaking, there was a notable lag between the claims made about performance appraisal in official company documents and actual day-to-day practices.

Evidence for divergence
While the strategic decision to reform the performance appraisal system in State Bank may indicate a degree of convergence at the macro level, both our employee survey and our in-depth interviews indicate that guanxi connections still clearly influenced the process. As noted earlier, the organisational cultures of SOEs have long been shown to be resistant to change in relation to HRM practices (Ding & Warner, 2001;Warner, 2014aWarner, , 2014b. Explanations for this include institutional approaches which highlight the entrenched norms and values associated with the goals of management in the pre-reform period (Ngo et al., 2008;Warner et al., 1999). Other possible reasons focus on the nature of 'traditional' Chinese cultural practices such as guanxi which, it is argued, do not fit readily with the ostensibly individualistic and target-based orientations of Western HRM practices (Cooke, 2013;Poon, Wei, & Rowley, 2010). Our interviews demonstrate that the enduring influence of guanxi on performance appraisal drives divergence in HRM practices in State Bank. However, the 'type' of guanxi practice in State Bank appears to be highly instrumental and is cultivated principally to ensure career progression (Chen et al., 2013). It was not framed or justified by participants in terms of a 'special' Chinese characteristic based on demonstrating ganqing (sentimental feeling and special care within the dyad). Those who did not have guanxi connections prior to their arrival at the bank, and were personally ambivalent about engaging in practices of favour exchange and gift-giving, needed to adapt quickly in order to build a future for themselves.
Furthermore, in State Bank, there is a tension between the strategic objectives of the Board of Directors, who wanted to use performance appraisal to improve efficiency and competitiveness, and the use of guanxi practices by supervisors and employees as a means of evading the accomplishment of performance-related targets and for maintaining face within the workgroup. There are, therefore, conflicting interests between the major actors engaged in the appraisal process in State Bank. On the one hand, the main driver for convergence comes largely from an 'outside' stakeholder, namely the Board of Directors. While the managers at regional headquarters had been incentivised to drive through the reforms, they faced resistance from key 'internal' stakeholders (namely the supervisors and employees directly involved in the appraisal process). Divergence is likely to continue in State Bank as there is a lack of alignment between the goals of 'outsiders' (the Board of Directors) and the preferences of 'insiders' (supervisors and employees) in relation to the purpose of performance appraisal (Jin & McDermott, 2013;Rowley & Bae, 2004). There is a need to recognise that 'culture' itself can become an important element of internal organisational politics and cannot simply be described as an inevitable consequence of indigenous preferences (Edwards & Kuruvilla, 2005). That supervisors in State Bank resist the new appraisal process can be defined as 'political' , in that they are attempting to defend their interests and to co-opt employees into the process of evading the actual achievement of performance-orientated goals (Edwards & Kuruvilla, 2005;Ferner & Quintanilla, 1998). In brief, divergence occurs in State Bank not just because of the mismatch between Chinese culture and Western HRM practices, but because of organisational politics and conflicting interests between internal and external stakeholders in the Bank itself.

Directions for further research and limitations of the study
This study raises a number of important theoretical and empirical issues which require further examination. Our interviews show that in State Bank, employees find that they have to develop guanxi in order to progress their careers, even if they have been recruited in an open search. Furthermore, participants in State Bank were more sensitive to the need to maintain face (mianzi) given the deeply embedded nature of guanxi networks within the bank. In contrast, guanxi and mianzi were not perceived as major limitations to the implementation of performance appraisal in Foreign Bank. Here we saw that the effect of culture may be moderated by ownership structures, managers with international experience, organisational culture and generational differences. Variance in these particular factors may be important reasons why divergence in HRM practices continues, not just in Chinese SOEs, but within East Asia more generally. We would therefore encourage researchers to study these themes in more detail in other contexts. It would be interesting, for example, to study the influence of guanxi practice on the appraisal process in the banking sectors of other Chinese communities, such as Taiwan, Hong Kong and Singapore. Cross-national comparison on the role of other forms of indigenous East Asian networks on performance appraisal (such as Japan and Korea), examined in a range of other knowledge-intensive sectors (such as technology-based industries and other professional services), would also help to extend the findings beyond the Chinese banking case (Horak, 2014(Horak, , 2015Horak & Klein, 2015).
Methodologically, it would also be useful if the findings recorded here could be replicated using larger samples in a longitudinal design. For example, in State Bank, our statements about historical change are based on the subjective perceptions of our participants collected at one point in time. If possible, conducting a study based on time series data in an SOE which is in the process of actually implementing a new appraisal process would add further support to our findings. Such a design would also give more substantive weight to our claims about which factors may influence convergence or divergence in HRM practices in China and East Asia more generally. In relation to our qualitative sample, we interviewed 22 individuals and therefore this part of our report cannot be used to make statistical generalisations. Nevertheless, despite these limitations, we believe that our research has developed important explanatory themes which can now be investigated in other studies. Finally, a number of authors have noted that guanxi and guanxi practice are rather difficult and sensitive research topics because they may have negative connotations in a business context, often being synonymous with corruption and bribery (Bian, 1994;Guthrie, 1998;Nolan, 2010). As such, participants may tend to deny any knowledge of guanxi or lie about their own actions in relation to the practice. However, like other guanxi researchers (e.g. Chen et al., 2013;Guthrie, 2009;Yang, 2002), we found that participants were actually prepared to express a wide range of views on both guanxi and guanxi practice. We would agree with Yang (2002), who argues that, given the sensitive nature of guanxi research, gathering authentic data requires that the researcher spend time in the field building trust with participants, which is what we attempted in this study. While not without its problems, we would suggest that providing there is an element of researcher reflexivity in both the interview process itself, and in the subsequent interpretation of the data, then the kind of data we have gathered here has a high degree of ecological validity. That is, our qualitative data were collected using methods which are well suited to capturing people's opinions and values as they are expressed and experienced in their everyday lives in naturalistic settings (Creswell & Miller, 2000).

Conclusion
We found evidence of convergence at the macro level in the design of performance appraisal systems, in that each bank attempted to use assessment criteria based on financial and efficiency-focused targets. However, an investigation of micro-level variables, such as cultural and generational differences, international experience, home country effects and institutional politics, provided evidence of significant divergence between the banks. Our study shows that when there is a clear alignment between the strategic goals of the organisation and the values of employees and supervisors, then convergence may occur. This was the case in Foreign Bank where guanxi played a relatively limited role in the appraisal process. In State Bank, however, there was a conflict between the Board of Directors, who were driving the changes in the performance appraisal system, and the supervisors and employees, who continued to place value on maintaining and developing guanxi relationships over and above the achievement of performance-related goals. Our study shows that the influence of guanxi is not homogeneous, but rather varies depending on the ownership structure of the bank. Consequently, we would suggest that divergence in HRM, both in China and other Confucian-influenced societies, is likely to occur not so much because of the assumed immutable nature of certain elements of indigenous 'culture' , such as guanxi, but rather because of the institutional heritage of any given sector or organisation. Note 1. We understand that the nationality of the bank is important for evaluating the effect of 'home country' on organisational culture. However, in order to fully protect the anonymity of Foreign Bank we cannot provide specific information about its nationality other than to say that it is headquartered in Europe.

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