Organizational justice in Chinese banks: understanding the variable influence of guanxi on perceptions of fairness in performance appraisal

ABSTRACT Drawing on survey (n = 308) and interview (n = 22) data from three different types of banks in China, we found significant differences in perceptions of organizational justice in performance appraisal processes. The state-owned bank was perceived as significantly less fair in its appraisal procedures than both its city-commercial and foreign-owned counterparts. These differences could be explained, in part, by variations in the influence of guanxi on supervisor decision-making. This, in turn, was linked to differences between the banks in their organizational objectives, as well as to intra-organizational differences at the departmental level. These findings question the common assumption that national culture variables, such as guanxi, are extremely stable and have universal explanatory value in all organizational contexts.


Introduction
We investigate the variable influence of guanxi on justice perceptions and performance appraisals in China's banking sector. Specifically, we focus on sources of difference between a state-owned bank, a foreign-owned bank, and a city commercial bank. By examining performance appraisal in banks with different ownership structures we investigate whether different organizational cultures and objectives moderate the influence of guanxi on justice perceptions. Our aim is to discover if national cultural characteristics, such as guanxi, are more malleable than is often assumed in the existing literature, which tends to emphasize their inherently stable and enduring qualities. We argue that there is a pressing need to develop a more nuanced approach to the investigation of the relationships between guanxi, justice perceptions and HRM outcomes, such as performance appraisal. We do this through adopting an emic methodological approach, working from the inside out, combining both quantitative and qualitative data, to give voice to some of those who directly experience performance appraisals in Chinese banks.
In the following critical literature review, we introduce some of the contemporary debates on how national culture variables, including guanxi, are thought to shape HRM Theoretical debates in performance appraisal and organizational justice This study focuses specifically on how guanxi influences performance appraisals and justice perceptions, therefore it is essential to introduce the key theoretical approaches in these areas. It is well established, and unsurprising, that if an employee perceives that processes and evaluation procedures used in performance appraisals are unfair, then they are unlikely to be effective tools of motivation (DeNisi and Murphy 2017;DeNisi and Smith 2014;Heffernan and Dundon 2016). Employees show extreme dissatisfaction with the performance appraisal process if they consider the systems to be biased and political (Dello Russo, Miraglia, and Borgogni 2017;Yamazaki and Yoon 2016). Studies investigating these processes often draw on the distinction made between distributive and procedural justice in organizational justice theory (Greenberg 1987;Guest 2017;Silva and Caetano 2016). Measures of distributive justice are designed to reflect employee perceptions of the fairness of outcome, whereas measures of procedural justice capture employee perceptions of the fairness of procedures in appraisal processes (Greenberg 1986(Greenberg , 1990Lucas et al. 2016). When employees believe that performance appraisal systems are inaccurate or unfair, they are unlikely to take them seriously (Eib et al. 2017). Consequently, their effectiveness as a means of improving employee engagement and productivity are of limited value (Murphy and Cleveland 1995;Nair and Salleh 2015).
More specifically, Thurston (2001) identified two sets of factors associated with employee perceptions of distributive justice in performance appraisals. The first set relates to the nature of supervisors' decision norms in appraisals (Rowley andRamasamy 2016a, 2016b). Supervisors may distribute performance ratings among employees by conforming to some pre-existing social norms such as equity, equality, need and social status, rather than basing them strictly on the objective work performance of any one individual. While employees tend to consider performance ratings distributed in accordance with an equity norm as fair, those which are based on perceived social status tend to be regarded as unfair and unjust. The second set of factors identified by Thurston (2001) relate to perceptions of the supervisor's personal goals in appraisal, such as avoiding conflict or seeking personal favours (Iqbal, Akbar, and Budhwar 2015;Smith, Craig Wallace, and Jordan 2016). Understandably, employees regard performance appraisal results as unfair if they are perceived to be aimed at satisfying the supervisor's personal preferences and concerns, rather than attempting to objectively evaluate employee performance. However, if employees perceive that the supervisor is using the appraisal to try to teach them or motivate them, then outcomes are generally considered to be fair.

Performance appraisals in Chinese organizations
Studies which have focused specifically on the effectiveness of appraisals in Chinese organizations tend to investigate the influence of well-established national cultural characteristics (Kang and Shen 2016;Vaiman and Brewster 2015). For example, respect for age seniority makes employees less likely to participate in open dialogue with supervisors about performance (Varma, Budhwar, and Singh 2015;Zhang et al. 2016); the need to maintain face mitigates against peer appraisals or critiques of supervisors (Poon, Wei, and Rowley 2010;Wang et al. 2017); and a concern with group harmony may promote a preference for recognizing 'softer' performance criteria, such as punctuality, loyalty and team effort, regardless of performance (Cooke 2013;Wei and Taormina 2011). Others have looked at the detrimental effects of perceived subjectivity and bias in the appraisal process. For example, a number of studies have identified that when employees perceive that their peers are favoured over others, either because of their connections or the appraiser's desire to reward seniority, this is demotivating and reduces the effectiveness of the appraisal process (Cheng 2014;Gu and Nolan 2017). Importantly, Chinese cultural factors, such as an employees' guanxi (personal connections) and seniority, seem to have a greater negative impact on performance appraisal practices in state-owned enterprises than in other types of organizations (Bai and Bennington 2005;Gu and Nolan 2017).
A number of studies have examined the use of performance appraisal in China's banking sector, the industry focus of this study. The reform of the state-owned banks began in the mid-2000s and involved a major restructuring of the sector, including the development of quasi-commercial city banks and opening up to foreign competition (Cousin 2011;Nolan 2010). In state-owned banks, changes included adopting a more instrumental approach to performance appraisals based on linking pay and promotion to the achievement of objective goals, such as profit or efficiency-based targets (Liu et al. 2017;Wang et al. 2017;Zhao and Du 2012). However, state-owned enterprises tend to face significant constraints when implementing results-based performance appraisals for a number of reasons (Akhtar, Ding, and Ge 2008). These include the enduring influence of institutional factors, such as conflicting policy pressures from local government, central government and trade unions, the limited skills and experiences of managers and the tendency of the state sector to be somewhat protected from competitive pressures (Guo, Huy, and Xiao 2017;Ngo, Lau, and Foley 2008;Nolan 2018c;Whelan and Muthuri 2017). Finally, cultural characteristics which favour allocating rewards on the basis of guanxi connections and seniority may continue to go relatively unchallenged in the state-owned sector (Cooke 2013;Nolan 2011;Wong 2017). On the other hand, multinational companies and private enterprises face greater commercial pressures which may mitigate (though not eliminate) these cultural preferences (Ma et al. 2016;Ding, Goodall, and Warner 2000;Gu and Nolan 2017).

Research questions and hypotheses development
On the basis of the above research literature, we can see that the use of guanxi practices in performance appraisal could be problematic in organizational settings where there is high potential for conflicts of interest (Chen, Chen, and Xin 2004). These conflicts can arise when a supervisor has to appraise employees who are low-performing, but well-connected in terms of guanxi. In such cases, the supervisor may seek to ingratiate themselves with the employee, especially if the employee's guanxi connections are with the supervisor's own managers. This process, may, in turn, have detrimental effects on employees who are high-performing, but not well-connected, as they will not be fairly evaluated in relation to their well-connected, but lower-performing peers. We argue that this situation is more likely to occur in the state-owned sector. As such, this is the site where we would expect employees' perceptions of organizational justice to be most disrupted. In other words, this is the organizational context in which supervisors may be more likely to merge formal (managerial) and informal (guanxi) relationships. Thus, our hypotheses are: Hypothesis 1: Employees' perceptions of distributive justice in relation to performance appraisal will be significantly lower in the state-owned bank than in either the foreign-owned bank or the city-commercial bank.
Hypothesis 2: Employees' perceptions of procedural justice in relation to performance appraisal will be significantly lower in the state-owned bank than in either the foreign-owned bank or the city-commercial bank.
We investigate reasons for any differences in our qualitative interviews, prioritizing the voice of those who actually experience these practices in their daily working lives. In our qualitative data, therefore, we aim to answer the following research question: Why do employee perceptions of distributive and procedural justice in performance appraisal processes vary in banks with different ownership structures?

Methodology
Sample and access to research sites As outlined in our introduction, three banks were selected for investigation: a state-owned bank (State Bank), a foreign-owned 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 in 2013. Initial access was gained through gatekeepers who were already known to the authors. Selection criteria for inclusion in the survey were that employees had been in the bank for more than one year and had experienced the performance appraisal cycle at least once. All potential respondents were then invited to participate on a voluntary basis and anonymity and confidentiality guaranteed. Following the survey, semi-structured interviews with employees and supervisors were arranged on an individual basis. Selection criteria for the semi-structured interviews were that people had been employed at their banks for at least three years (see Table 1 for sample details).

Research design and procedures
We adopted a two-step explanatory mixed-methods design based on the philosophical standpoint of pragmatism (Creswell 2013;Johnson and Gray 2010). Two methods were used for data collection: i) a quantitative survey and ii) semi-structured qualitative interviews. Quantitative analysis consisted of the multiple comparison of mean difference of relevant scales by using one-way ANOVA and Levene's test. The data analysis strategy used with the qualitative data was thematic analysis. This incorporated both the deductive technique, generating themes from the data which echoed with previous research and prior theories (Crabtree and Miller 1999), and the inductive technique, generating new themes from the raw data (Boyatzis 1998;Creswell 2013). We followed the well-established phases of thematic analysis: familiarization with data, generating initial codes, searching for themes among codes, reviewing themes and defining and naming themes (Creswell 2013).

Measures
Quantitative measures of distributive and procedural justice were drawn from the scale developed by Colquitt (2001). Some wordings of the items were modified as appropriate in order to fit the cultural context of the study.

Distributive justice
Distributive justice was measured by four factors: (1) 'My performance rating reflects the effort I have put into my work', (2) 'My performance rating is appropriate for the work I have completed', (3) 'My performance rating reflects what I have contributed, (4) 'My performance rating is justified, given my performance'. Cronbach's alpha for the measure was 0.87.

Procedural justice
The measure for procedural justice had seven items: (1) 'I have been able to express my views and feelings during the procedures of performance appraisal', (2) 'I have had influence over my performance rating arrived at by those procedures of performance appraisal', (3) 'Those procedures of performance appraisal have been applied consistently', (4) 'Those procedures of performance appraisal have been free of bias', (5) 'Those procedures have been based on accurate information', (6) 'I have been able to appeal my performance rating arrived at by those procedures of performance appraisal', (7) 'Those procedures of performance appraisal have upheld ethical and moral standards'. Cronbach's alpha for the measure was 0.71.

Quantitative results
Employees' perceptions of distributive justice in relation to performance appraisal A one-way ANOVA was used to compare the mean scores on the scale used to measure employee perceptions of distributive justice in relation to performance appraisal. The means of this variable for the three banks are: 3.24 for State Bank, 3.68 for Foreign Bank Average age (years) b 37 28 29 a Defined as having studied overseas or had international work experience. Information derived from key informants and annual reports. b Information derived from key informants and annual reports. and 3.60 for City Bank. The descriptive statistics are shown in Table 2. The p-value for Levene's test (<0.05), indicates that a significant difference exists between the variance in the population (see Tables 3 and 4). A significant difference was found between State Bank and the other two banks. There was, however, no statistically significant difference between Foreign Bank and City Bank (see Table 5). To summarize: employees in State Bank expressed significantly lower distributive justice perceptions than those in Foreign Bank and City Bank, thus, Hypothesis 1 is supported.
Employees' perceptions of procedural justice in relation to performance appraisal A one-way ANOVA was used to compare the mean scores on the scale used to measure employee perceptions of procedural justice in relation to performance appraisal. The means for this variable for the three banks are: 3.32 for State Bank, 3.62 for Foreign Bank and 3.48 for City Bank. The descriptive statistics are shown in Table 6. The p-value for Levene's test (<0.05), indicates that a significant difference exists between the variance in the population (see Tables 7 and 8). A significant difference was found between State Bank and the other two banks, with employees in State Bank expressing lower procedural justice perceptions  than those in Foreign Bank and City Bank. A significant difference was also observed between Foreign Bank and City Bank with employees in City Bank expressing lower procedural justice perceptions than those in Foreign Bank (see Table 9). To summarize, employees in State Bank expressed significantly lower procedural justice perceptions than those in both Foreign Bank and City Bank, thus, Hypothesis 2 is supported.

Qualitative findings
Qualitative data is key to interpreting and explaining the differences identified in the quantitative data described above. It is here that we begin to understand the interaction between organizational context and cultural characteristics. Following Tsui (2006), we worked with the qualitative data from the 'inside out', giving voice to those who actually experience performance appraisal processes in their daily lives. Specifically, we identified important perceived differences in each organization in relation to the influence of guanxi connections, mianzi and seniority on both the ratings awarded and the outcomes of appraisals. These are best described in relation to differences in perceptions of distributive and procedural justice. Important themes are summarized in Table 10 and are discussed in  detail below. Themes are illustrated with salient quotes from participants (below each quote is a participant number which matches the participant details as described in Appendix 1).

Distributive justice
Perceived fairness in relation to actual performance ratings Informants in State Bank perceived that some supervisors made appraisal decisions based primarily on an employees' guanxi connections and their seniority within the bank, rather than their actual work performance. While informants acknowledge guanxi as a characteristic of Chinese culture, they were not satisfied with the situation at State Bank and perceived its influence on performance appraisals as unfair. For example, one young interviewee in State Bank claimed: In my department, actual performance is not the only determinant of the appraisal outcomes. Those employees who have good guanxi with their supervisor or have strong guanxi connections outside of the bank can obtain better performance ratings than those who don't. Also the supervisor tends to provide better performance ratings to those with higher seniority, in order to show respect to them, regardless of their real performance. As a young person who does not have guanxi, I think this is unfair and it really demotivates me. Interviewee #7 Another employee at State Bank made similar observations about the unfair influence of guanxi on performance ratings, describing how he felt his own contributions were often undervalued and unacknowledged: The levels of guanxi quality between supervisor and each employee varied significantly in my workgroup. I did not have a close or special guanxi relationship with my supervisor. This, I think, is the reason why over the past five years I have not been able to obtain a proper performance rating. I have worked hard to perform better than my peers but I feel my actual performance is not acknowledged. Because of this, I became very sensitive to supervisors' decisions. I always noticed whether or not the supervisors made fair appraisal decisions and looked for supporting evidence for their assessments. I noticed that those colleagues with guanxi would receive better grades from the supervisor, regardless of objective performance. Interviewee #5 Table 10. Summary of explanatory themes identified in qualitative data.
Interviewees vary in each bank according to the following themes: Distributive justice • perceived fairness of performance appraisal in relation to actual performance ratings (State Bank employees questioned their supervisors' neutrality and lack of bias. They felt that those employees who had strong guanxi connections were given more favourable results, regardless of their actual performance) • perceived fairness of performance appraisal in relation to appraisal related outcomes (State Bank employees did not see any link between the appraisal itself and HR outcomes such as bonuses and promotion) Procedural justice • employee involvement in goal setting procedures (goal setting processes are not well-understood in State Bank, nor do State Bank employees feel empowered to contribute to the goal-setting process) • voice and a fair hearing (appeals processes are well-established in City Bank and Foreign Bank, but not in State Bank) • formal justification of performance rating to employee (employees in State Bank are not given an explanation of the appraisal decision, whereas formal meetings are held in both Foreign Bank and City Bank) Overall, our informants indicated that, in State Bank, supervisors seek to advance their own careers by building good relationships with those employees who, themselves, had influential guanxi connections. Supervisors did this by giving those with strong guanxi connections more favourable evaluations than those without, regardless of their actual performance. In line with research based on the equity norm (Thurston 2001), it is not surprising that those without these connections regard the practice as demotivating and, importantly, a form of favouritism and political manoeuvring on the part of the supervisor within the organization.
Cultivating guanxi is often described as an enduring and pervasive characteristic of workplace organization in China and one reason why Western HRM practices, such as performance appraisal, may have limited effectiveness in motivating and rewarding highly skilled professionals (Warner 2014). However, we would argue that guanxi influence may vary according to the structure and objectives of any given organization. The influence of guanxi in State Bank endures because of an institutional heritage which sees state-owned enterprises in areas of strategic importance, such as finance and banking, shielded from competitive market processes (Nolan 2018b). This, in turn, means that the 'objective' performance of employees does not matter as much as it might in those organizations that are exposed to more intense market pressures. This is more clearly illustrated when we compare the appraisal practices in State Bank with those in Foreign Bank and City Bank. In Foreign Bank, for example, where there is no state support at all, tangible results in the business departments really mattered. Employees' guanxi was perceived to have a very limited impact on supervisors' decisionmaking in appraisals and our informants perceived that their ratings reflected their performance with reasonable accuracy. As one manager in Foreign Bank pointed out: Actually, senior managers evaluate my own personal performance based on the performance of the workgroup as a whole. The actual and real performance of my staff has a great impact on the outcome of my own personal performance appraisal. So, in order to retain and motivate talented employees in my group, I will make sure to award those who perform well with better performance ratings and more bonuses. Interviewee #12 While informants indicated that the pursuit of fair and equitable results was the norm for supervisors' decisions in appraisals in Foreign Bank, this does not mean that the use of guanxi was completely absent in the organization as a whole. One participant from the HRM department noted that in those areas of the bank which were focused more on supportive, administrative functions (rather than in business departments where sales and profit targets were of key importance) employee guanxi connections did seem to occasionally influence appraisals: In business departments guanxi has little influence on appraisals. But in those administrative departments, such as the HR department where I work, guanxi does seem to impact supervisors' decisions. This I think is because the majority of the appraisal criteria used in HR department are subjective and not so easily quantified as those in the business departments. This means that those employees who have good guanxi with supervisors tend to interact more frequently with them and so can influence their views about the quality of the work. This is more easily done when the objective is unquantifiable. So in these cases the supervisor tends to consider that these employees have better performance than others and give them better performance ratings. Interviewee #10 Similarly, in City Bank our informants felt there were occasions when guanxi had some effect on supervisor decision making, but that, in the main, appraisals reflected actual performance. To the extent that City Bank was a hybrid organization, partially run by local government, there was perhaps more of an inclination for supervisors to attempt to build connections through their employees than in Foreign Bank. Nevertheless, as participants noted, performance really mattered in City Bank and this tended to limit the influence which employee guanxi could have on supervisors. As one participant noted: Basically, my supervisor makes decisions regarding appraisals primarily based on employees' performance. However, as I know from experience, in a small number of workgroups, guanxi still influences supervisors' decisions in evaluation. Generally speaking, I believe that we should all concentrate first on improving our work performance. However, we should also try to maintain and develop good guanxi with our supervisor. This is like a double insurance. It will help our performance appraisal as well as our promotion and career. Interviewee #20 Overall, in City Bank, our informants perceived that the outcome of appraisals were a reasonable reflection of employees' actual performance and contribution. The interview data indicates that the impact of supervisors' favouritism on appraisals in City Bank was weaker than in State Bank, but somewhat more prevalent than in Foreign Bank.

Perceived fairness in relation to appraisal outcomes
There were also differences in perceptions of appraisal-related outcomes between the three banks. For instance, in State Bank, little importance was attached to appraisals by either supervisors or employees because there was a tacit understanding that the supervisor would make decisions on rewards which were related to an employees' guanxi connections and the desire to 'give face' (mianzi) to more senior employees. Consequently, those without guanxi connections were not motivated by appraisals, considering them unfair and a waste of time. As one employee said: Both the supervisors and employees need to spend a lot of time on the appraisal process with great pressure and tension . . . but we all know they have no impact on our personal development or income. So, I do not think my colleagues and I could be motivated to positively participate in, or pay more attention to, the appraisal. The supervisors do not take the appraisal seriously, because a large number of employees do not really care about the results. I would suggest that the bonuses and promotion should be linked to the appraisal outcome and that the outcome should be based on a fair and unbiased assessment of our performance. Then, both employees and supervisors would take the appraisal seriously and the employees would actually be highly motivated to improve their performance. Interviewee #4 A different perspective was found in Foreign Bank because, as noted above, there was a strong emphasis on profitability in the business departments. This created an organizational culture where appraisals in the business departments were taken seriously. Moreover, the outcome of appraisals was clearly linked to rewards, such as bonuses and promotions. The employees who received better performance ratings also received higher bonuses. According to informants, because the appraisals actually reflected performance, they worked better as a motivational tool. One employee said: How well you perform equals how much you can get. The result of performance appraisals certainly determines how much bonus you obtain at the end of the year and it increases your possibility of gaining a promotion. This means we all take the appraisal system seriously. Interviewee #14 In City Bank, performance ratings were also perceived as an important influence on employees' bonuses, promotions and opportunities for training, regardless of guanxi connections. However, while there are clear rewards based on performance, there were also occasional instances where those with influential guanxi were given preferential promotion opportunities. City Bank certainly faced more pressures than State Bank to be profit-orientated, but it nevertheless retained traces of an organizational culture based on personal connections. To quote one informant: I believe that the performance ratings can determine the bonuses. However, I do not really think performance ratings could have great impact on promotion. Performance rating is just one influential factor in promotion. I guess the supervisor needs to consider factors such as communication skills, team-building skills, leadership and so on. But I have seen that guanxi can sometimes be important too, it is not surprising to me that a colleague whose performance is not that good can still get promoted. Interviewee #21 So, on the basis of our interviewees, it appears that the reward system in City Bank partially reflects employees' performance ratings and actual contributions. In this sense it offers a more equitable appraisal system than in State Bank. Nevertheless, our interviews indicate that the favouritism and bias associated with guanxi operates to a greater extent in City Bank than in Foreign Bank.

Procedural justice
Employee involvement in goal-setting procedures As well as differences between the banks in employee perceptions of distributive justice, we also found variation in perceptions of procedural justice. One key area of difference was the extent of employee involvement in the goal-setting process. In State Bank, our participants were not satisfied with the goal-setting process. First, they were not given clear performance objectives in advance of the appraisal. This meant they were confused about performance standards and did not know how to go about achieving their goals. Second, employees were not given opportunities to participate in the objective setting process and had to passively accept objectives set by their supervisors. One interviewee in State Bank said: The supervisor sets the objectives for each of us at the beginning of the year and passes the objectives to HR department. Then, the HR department contacts us and gives us the objectives. We can only accept it, even if we are not satisfied with it, or we do not know whether we can fulfil it, or we do not know how to realize it. Interviewee #6 Interestingly, in both Foreign Bank and City Bank, neither of these problems arose. Informants in both banks noted that they had opportunities to participate in the objective setting process, were able to express their own opinions to their supervisor and had a clear understanding of performance objectives. One informant in City Bank noted: My supervisor will hold separate meetings with each of us at the beginning of a year. In the meeting, he introduces the team objective of the year and discusses this with me regarding my personal objective. Finally, we can always achieve a mutual agreement on this. I think I've been basically satisfied with my personal objectives over the past several years. Even if sometimes the objective was a little higher than my expectation, I totally understand why my supervisor allocates it to me and I am willing to strive to achieve this with his support and help. Interviewee #20 Supervisors in Foreign Bank and City Bank allowed employees to participate in the goal-setting process, ensuring they had clear performance objectives each year. This differed significantly from State Bank where joint participation was not encouraged. Our interviews indicate that this factor is key to understanding organizational differences in employee perceptions of procedural justice.

Voice and a fair hearing
Participants also differed in their perceptions of how easy it was to express themselves in appraisals and ensure they had a fair hearing. As noted above, in State Bank, participants without guanxi claimed they did not have the opportunity to participate in goal setting and felt generally unable to express their feelings about appraisals, or to influence their supervisors' decisions. Informants noted that this was due to the hierarchical and formal relationship between supervisor and employee. This was especially difficult for those without guanxi connections as they did not have an open channel of communication with their supervisors. Some felt that they didn't even want to talk with their supervisors about appraisals because they viewed the supervisor as unhelpful and uncaring. One informant in State bank said: I can feel the distance between my supervisor and me. Actually, I am a little afraid of him. I do not even know whether he is aware of what I do at work. My colleagues who have good guanxi with their supervisor can readily show their achievements to the supervisor, obtain more help from them and can even influence the supervisor's decision. But I can't do this. I just passively accept the objectives given and the final performance rating, even if I am not satisfied with them. Interviewee #5 In both Foreign Bank and City Bank, participants did not experience any of these difficulties and noted that the organization actively encouraged open two-way communication between supervisors and employees. Employees in both banks noted that they had the opportunity to share their ideas and feelings about their appraisals with their supervisors and were able to talk openly about their contributions and performance. They were able to discuss openly their strengths and weaknesses and tell supervisors about emerging problems. As one participant in City Bank said: If I have any problem in work, it is easy for me to ask my supervisor for suggestions. My supervisor is always ready to help us. I will try my best to let my supervisor know of my achievements and contributions at work and I make sure he has the full picture of what I have actually done. Interviewee #21 Formal justification of the performance ratings to employees Finally, there were notable differences in the post-appraisal procedures used by supervisors in each bank. Participants from State Bank felt that supervisors did not provide enough evidence to justify their assessments. As one informant reported, his supervisor quickly told him the result and provided some brief general suggestions for future improvements before bringing the formal interview to a close. He felt unable to directly challenge his supervisor because of the perceived need to maintain 'face' (mianzi) and harmony within the bank. As he noted: In the formal meeting, my supervisor tells me the result of appraisal and explains to me why he allocated this rating to me. His explanation isn't convincing to me. But I won't argue with him or even complain about it to higher managers, because this will break my relationship with the supervisor and even the harmony within the department. All I can do is to accept the supervisors' decisions with smiles and ask for suggestions for future work.

Interviewee #8
This situation contrasts with the experience of participants in both Foreign Bank and City Bank. They felt that supervisors provided sufficient evidence to justify their appraisal decisions and were able to offer employees constructive feedback which demonstrated a familiarity with their work. As one participant in Foreign Bank noted, his supervisor put a lot of effort into the appraisal, providing very detailed feedback on the performance rating: Although it is not strictly necessary, my supervisor informed me of my colleagues' appraisal outcomes and compared their achievements with mine, in order to justify his decision on my performance rating. My supervisor explained his reasons in detail with great patience. I appreciate this because the appraisal outcome is directly linked to our yearly bonuses which really matters to me. Interviewee #14 So, on the basis of our interviewees, it appears that employees in both Foreign Bank and City Bank were satisfied with the performance standards used in appraisals, felt able to voice their opinions and received adequate explanations of their appraisal outcome. In State Bank, however, favouritism, the need to maintain face, and biases associated with guanxi relationships effectively disempowered and silenced those employees without good connections.

Discussion
Our quantitative results indicate that employee distributive and procedural justice perceptions of appraisal processes are significantly lower in State Bank than in both Foreign Bank and City Bank. Analysis of our qualitative data indicates that, in State Bank, employee guanxi connections are perceived as one of the most important influences on supervisors' appraisal-rating decisions. This is perceived as demotivating and unfair by those without guanxi connections. Similarly, it appears that the results of appraisals in State Bank are not fairly linked to the distribution of bonuses and promotions. This contrasts with employees in both Foreign Bank and City Bank who felt that appraisals were a reasonable (though by no means prefect) reflection of work performance. Rewards were usually distributed based on performance ratings perceived as fair, although this was somewhat more ambiguous in City Bank.
In brief, our study indicates that theoretical models of organizational justice developed in the West are relevant to managers and employees in China (DeNisi and Murphy 2017). We would argue that this is especially so for those employed in the professional services, such as banking, where there is a growing need for graduate-educated, urban-based labour (Cousin 2011). In relation to the influence of guanxi on perceptions of fairness, our findings show that it has varying influence which may be dependent on organizational, and even departmental level goals (Sheldon and Sanders 2016). Of particular note is that where the profit motive is stronger and the market competition harder, the influence of guanxi in appraisal diminishes (Gu and Nolan 2017). Fairness in appraisal clearly matters to our participants, most of whom wish to be judged on merit. Our qualitative study indicates that the equity norm (Thurston 2001) is especially important to those without guanxi. If China's policy initiative to move up the value chain continues (Nolan 2018b), then there will inevitably be a growing demand for professional services employees, many of whom will not hold traditional guanxi connections. It is inevitable that this group may wish to be fairly evaluated on the basis of their skills, experience and education, not on their personal networks (Dello Russo, Miraglia, and Borgogni 2017).
Our findings also show that explanations for differences in employee justice perceptions occur partly because of contextual variation at the organizational level. While those in State Bank did not feel able to challenge their supervisors' decision-making processes, this cannot be explained solely by reference to the catch-all concept of 'national culture', and the assumed need of Chinese employees to maintain harmony at all costs (Hofstede 2001;Warner 2014). We have demonstrated that participants in Foreign Bank and City Bank did feel their supervisors' decision-making was fair (though not perfect) and that they were given opportunities to engage with, and influence, the appraisal process. As such, our research indicates that particularist cultural attitudes, such as guanxi, are less influential outside the state-owned sector. Furthermore, our research shows that supervisors clearly have different personal goals in different types of bank, and may even have different goals within different departments within the same bank. As noted above, in Foreign Bank, which was clearly profit-orientated, the use of fair, performance-based appraisals were especially important in the business departments where the fate of both supervisor and employee were tied to financial targets. However, in the HR department of Foreign Bank the goals were 'softer' and so guanxi continued to play some role in appraisals in that division. In brief, while guanxi is influential in moulding justice perceptions in performance appraisal processes, it is not homogenously so. This questions some of the core assumptions of the national culture approach outlined earlier, especially the assertion that it is a stable and predictable influence on HRM practices, immutable to change across time and space (Hofstede 2001;Warner 2014). Rather, we would argue that it is highly context specific and should be understood as a dynamic, shifting process (Child 2009;Nolan 2018a).

Implications for practitioners
Our findings have several implications for practitioners. The process of implementing performance appraisal needs careful planning in order to be effective in Chinese organizations. This requires management's careful consideration of the relevance of the 'best practice' model of performance appraisal for their particular organization. If senior managers are not committed to ensuring that performance appraisals are administered in a way which allows employees to feel they have received fair treatment, then they are likely to lose talent. This is an enduring problem for the state sector (Guo, Huy, and Xiao 2017). Another implication is the need for greater professionalization of all those involved in the HR function, including managers and supervisors involved in appraisals (Zhao and Du 2012;Saha and Rowley 2015;Nankervis, Rowley, and Salleh 2016). Investment in quality training, management development and cross-cultural awareness programmes is important, not just in relation to understanding the many factors involved in influencing the appraisal process itself, but also in relation to the cultural diversity within which this process is grounded and practiced. There is also a greater need to improve employee participation (Kernan and Hanges 2002), adopt a wider variety of appraisal procedures, such as 360 evaluations (Church et al. 2019), and to address organizational climate, as individuals are likely to learn justice evaluations from their peers (Roberson and Colquitt 2005).

Directions for future research and limitations of the study
Further research could address some of the limitations of our study. Our study is cross sectional and is limited in geographical scope (based on cities in Eastern China) and in sectoral focus (it only examines the banking sector). Further research could be longitudinal, examining the introduction and operation of appraisal processes over time. We also need to know more about how individual characteristics such as age and gender influence appraisal outcomes. More qualitative and in-depth case approaches could investigate in more detail some of the themes we have identified here. As well as confirming our findings, such studies are likely to broaden and deepen our understanding of the appraisal process in different organizational contexts. As Sheldon and Sanders (2016) have noted, too many studies focus on the consequences of HRM practices for firm performance and overlook the daily lived experience of those who are actually exposed to such procedures. In this regard, we would encourage others to adopt our own, emic, 'inside out', approach to the study of Chinese organizations (Tsui 2006).

Conclusion
Our survey data shows that employees' distributive and procedural justice perceptions in State Bank were significantly poorer than in Foreign Bank or City Bank. Our interviews indicated several reasons for this. These included: supervisor bias towards those with guanxi connections, unclear links between appraisal outcomes and rewards, poor feedback and lack of employee voice. Our qualitative data also indicates that employees appreciated management's attempts to establish more meritocratic appraisal processes in Foreign Bank and City Bank. Indeed, we would conclude by suggesting that younger, educated professionals may even be resisting guanxi cultures, especially when they are not well-connected themselves. Clearly this claim needs further investigation, but our study does highlight the limitations of over-reliance on the national culture approach when investigating performance appraisal in China. The assumption that national culture features, such as guanxi, are extremely stable over time and universal in influence across all sectors and all organizations lacks nuance and needs to be challenged.