Transformational leadership focuses on real-time problems, defines new benchmarks, builds understanding, and motivates and shapes the behavior of subordinates to achieve organizational goals effectively (Manshadi et al., 2014; Nagy and Edelman, 2014; Middleton et al., 2015; Jiang et al., 2017; Matwally and El Zarka, 2017; Arif and Akram, 2018). Studies suggest that the role of every manager in the organization is to be a leader instead of only a manager (Hall et al., 2015). Organizational success improves through the enhanced effects of TL (Sun and Henderson, 2017; Maaitah, 2018). TL also enhances employee performance in groups/teams (Amin et al., 2016). Past research has shown that TL plays a critical role in the success of Chinese SMEs (Lin and Sun, 2018). Chinese employees prefer leaders who exhibit the traits of TL (idealized consideration, IS, LI, and II), for instance, acting as a role model, selflessness (Farh and Cheng, 2000), avoiding the use of abusive power, setting a good example, and working for employee well-being (Dunfee and Warren, 2001; Cheng et al., 2004; Xiaoxia and Jing, 2006; Chen et al., 2012; Lin and Sun, 2018; Su et al., 2019).

During the past decade, there has been extensive research on TL and its relationship to multiple outcomes. TL significantly affects OCB (Rodrigues and Ferreira, 2015; Sarwar et al., 2015; Ismaeelzadeh et al., 2016; Saif et al., 2016; Cofie, 2018; Hassi, 2018). Additionally, employees exceed their assigned duties when a TL style is used (OCB, 2018). Within SMEs, all four dimensions of TL play a vital role in cultivating OCB (Jiang et al., 2017). Past research examined the positive effect of TL traits on OCB (Emami et al., 2012; Pickford and Joy, 2016; Majeed et al., 2017; Cofie, 2018). Bass (1985) explained that the qualities of transformational leaders such as individualized consideration, IS, inspirational motivation, and individualized influence enhance OCB in employees; for instance, Jiang et al. (2017) examined positive impacts of TL dimension on OCB. The following relationship has been established for this study:

There is a considerable amount of literature on ER, highlighting its importance for all types of firms. The cost of losing employees is much higher than retaining them through compensation plans (Carter et al., 2019). To investigate this issue, many factors have been considered, such as the control variables of age, education, experience, sex (Deshwal, 2015), peer support (Ali et al., 2017), recruitment and selection, job preview, organizational culture, employee relations, awards and recognition, work–life balance, and training and development (Olendo and Muindi, 2017). Leadership equally benefits employees and organizations, and specifically, TL affects ER (Amankwaa and Anku-Tsede, 2015), both directly and indirectly (Khan, 2015; Nohe and Hertel, 2017). Transformational leaders improve subordinates’ performance by achieving organizational goals (Sow et al., 2016) and implementing a reward system to retain employees (Adekanbi, 2016). TL increases ER (Abouraia and Othman, 2017; Gyensare et al., 2017; Jiang et al., 2017) and reduces turnover intention (Maaitah, 2018).

Transformational leadership influences the retention choices of employees (Sulamuthu and Yusof, 2018). Furthermore, the theory of transformational leaders strongly supports the relationship between TL and ER (Amankwaa and Anku-Tsede, 2015; Khan, 2015). This study proposes that employees show higher levels of retention when leaders exhibit individualized influence, IS, inspirational motivation, and individualized influence. Past studies have examined the relationship between TL and ER through TL theory (Adekanbi, 2016; Sow et al., 2016). Therefore, we have proposed the second hypothesis, as follows:

The concept of OCB first appeared in the early 1980s and initially described the specified behavior of employees within the organizations. Scholars described “organizational commitment and individual traits” as factors to enhance OCB (Emami et al., 2012). Voluntary behaviors of employees to prove themselves as good citizens of the organization are called OCB (Tambe, 2014). Similarly, an organization’s success is critical without OCB (Obiora and Okpu, 2014). The positive effects of OCB’s three dimensions, namely, public benefits, sportsmanship, and self-sacrifice, on employee well-being increase ER (Tambe and Shanker, 2015; Yurcu et al., 2015). Additionally, OCB refers to the behaviors that employees exhibit outside of their formal responsibilities. Leaders can help employees enhance OCB and benefit the organization (Pickford and Joy, 2016; Yaylaci, 2016; Zeyada, 2018). Furthermore, OCB refers to discretionary behavior, which is not directly or explicitly recognized by the formal reward system. However, such behaviors promote the effective functioning of the organization (Majeed et al., 2017). OCB enhances both individual and team performance (Mehdizadeh et al., 2018; OCB, 2018). Few studies have also examined a negative relationship between OCB and turnover intention (Islam et al., 2012). A higher level of OCB significantly affects ER (Dash and Pradhan, 2014; Paillé et al., 2015; Pivi and Hassan, 2015; Anvari et al., 2017; Olendo and Muindi, 2017; Mittal and Kaur, 2018). This relationship will be analyzed in Chinese SMEs with the following hypothesis:

Scholars have argued that OCB plays a critical role in SMEs’ success in China (Farh et al., 2004). TL affects ER in Chinese SMEs, both directly and indirectly (Sun and Wang, 2017). Furthermore, OCB was found to partially mediate the relationship between internal corporate social responsibility and intention to quit (Rashid et al., 2018). Jiang et al. (2017) found that OCB mediated more than half of the effects of TL on sustainable employee performance. Saoula and Johari (2016) studied the positive mediation of OCB between the relationship of perceived organizational support and turnover intention.

Similarly, Khokhar and Zia-ur-Rehman (2017) examined the positive and significant mediating role of OCB between TL, creativity, and ER. Selamat and Ran (2019) found that OCB significantly mediates the relationship between organizational justice and performance within SMEs in China. Chiang and Hsieh (2012) found that OCB partially mediated the relationship between perceived organizational support and job performance. The four traits of Bass’s (1985) theory, including individualized influence, IS, inspirational motivation, and individualized consideration, develop specific leadership skills in leaders. Transformational leaders help to develop OCB in their followers (Emami et al., 2012; Majeed et al., 2017; Cofie, 2018), further leading to ER (Adekanbi, 2016; Sow et al., 2016). However, this relationship with the perspective of Chinese SMEs has not been extensively explored. We propose that OCB plays a mediating role in the relationship between TL and ER. Therefore, we develop the following relationship to check the mediation of OCB:

Communication is critical to motivate employees, measure success, deliver products and services to customers (Conrad, 2014; Nwata et al., 2016), and enhance the performance of employees (Femi, 2014). Effective communication between leaders and employees significantly increases employee output and efficiency (Hills, 2015; Luthra and Dahiya, 2015; Sadia et al., 2016) and employee commitment (Marchalina and Ahmad, 2017). Therefore, through effective communication, employees feel valued, motivated, and rewarded for their efforts toward organizational success (Kukla, 2017), and individual and organizational betterment (Sergeeva, 2018). According to scholars, “communicators have finally started putting a greater focus on the development of leadership and management communication” (Gatehouse, 2018).

Scholars have argued that effective communication by leaders increases OCB (Yildirim, 2014; Herfina and Rubini, 2015). Diebig et al. (2017) studied the positive moderation effect of direct communication on TL and daily team cooperation. Garnett et al. (2008) studied communication as a moderator between organizational culture and public organizations’ performance. Pettit et al. (1997) studied the moderation of communication between job performance and satisfaction. Similarly, Villegas and Cerveny (2004) also found communication to mediate between job satisfaction and absenteeism positively. On the basis of the discussion above, we propose that communication could be considered a moderator between the relationships of TL, OCB, and ER. Bass’s (1985) theory supports the idea that leaders must effectively convey the vision and mission they have for the organization. This study explores the relationship in SMEs in China. All the constructed relationships are presented in Figure 1.

Employees of manufacturing SMEs in Jiangsu Province Mainland China participated in this study. A sample size of 505 was obtained using random sampling technique. A total of 600 employees were contacted personally and online to distribute the questionnaires. Out of the 516 questionnaires received, 11 were rejected owing to missing information/incomplete responses. A total of 505 (84.16%) response rates were recorded for further examination. A total of 406 male and 98 female participants accounted for 80 and 20%, respectively. Participants in the 20–30 (210), 31–40 (231), 41–50 (42), and 51–60 (21) age groups accounted for 41.5, 46, 8, and 4%, respectively. Similarly, 56 participants had intermediate (high school) education, 210 participants had bachelor’s degrees, 210 participants had master’s degrees, and 28 participants had Ph.D. degrees, accounting for 11, 41.5, 41.5, and 5.5%, respectively. Twenty-eight participants had less than 1 year of work experience, 196 participants had 1–5 years, 182 participants had 6–10 years, 77 participants had 11–15 years, and 21 participants had more than 15 years of work experience, accounting for 5.5, 39, 36, 15, and 4% of the total participants, respectively (Table 1).

The statistical software SmartPLS was used to analyze data. First, measurement model techniques were used to test the Cronbach alpha, heterotrait–monotrait (HTMT) ratio, composite reliability (CR), and average variance extracted (AVE). Second, this study used discriminant validity and correlation to analyze the theoretical model. Third, the study assessed the structural model by analyzing collinearity/common method bias [variance inflation factor (VIF)], coefficient of determination (R²), F², predictive relevance (Q²), and standardized root mean square residual (SRMR). Finally, this study performed structural equation modeling (SEM) to test the hypothesis.

The reliability of the scales was determined by Cronbach’s alpha (CA) test. The validity of the measurement scales was found to be significant, with values of 0.895 for C, 0.926 for ER, 0.897 for OCB, and 0.931 for TL. Adequate CR or internal consistency reliability measured in the present study ranged between 0.923 and 0.948 (equal or above 0.7, as suggested by Bagozzi and Yi, 1988; Hair et al., 2011). Moreover, the present study met the threshold of convergent validity (AVE) of at least 0.50 (Fornell and Larcker, 1981; Chin, 1998; Table 2).

According to scholars, the HTMT, to assess multicollinearity within the data, should not be higher than 0.9 (Gold et al., 2001; Teo et al., 2008). The study met the standard, as results were found in the range of 0.343 to 0.736 (Table 3). The discriminant validity results are presented in Table 4, which shows a significant value of 0.84 for C, 0.879 for ER, 0.842 for OCB, and 0.886 for TL. Moreover, the results show that TL has a positive correlation with OCB (0.533), TL and ER (0.557), and OCB and ER (0.457); communication and TL, OCB, and ER have positive correlations with values of 0.59, 0.659, and 0.547, respectively.

This study measured collinearity and common method bias issues through the VIF. VIF is defined as the reciprocal of tolerance. As suggested by the scholars Kock (2015) and Hair et al. (2011), this study was considered bias-free with no values equal to or lower than 3.3 (Table 5). Furthermore, Harman’s single factor test (suggested by Podsakoff et al., 2003) indicated that the maximum variance that is explained by a single factor is 38.4%. Henceforth, we conclude that this dataset does not suffer from common method bias (Kock, 2015).

According to previous studies, R² measures the model’s predictive power (Sarstedt et al., 2014). The value of 0.418 indicates that 41.8% of variations in ER occurred because of independent variables (0.75 = substantial, 0.5 = moderate, and 0.25 = weak, as suggested by Henseler et al., 2009; Hair et al., 2011). Additionally, Cohen (2013) noted that the values of 0.02, 0.15, and 0.35 represent small, medium, and significant effects, respectively. If the value of f² is <0.02, it indicates that there is no effect. The results of the study shown in Table 5 show that there was an effect.

Predictive relevance is an indicator of the model’s out-of-sample predictive power or predictive relevance given by Stone and Geisser’s Q² value (Geisser, 1974; Stone, 1974). In the SEM, Q² values larger than zero for a specific reflective endogenous latent variable indicate the path model’s predictive relevance for a particular dependent construct. The results of this study show medium predictive importance (0.02 = small, 0.15 = medium, and 0.35 = immense, as suggested by Geisser, 1974; Stone, 1974).

Standardized root mean square residual is the absolute measure of fit, and a value of zero indicates the perfect fit. SRMR is defined as “the root mean square discrepancy between the observed correlations and the model-implied correlations.” The results show a significant value of 0.065 (Table 5), and if the value of SRMR is less than 0.08, it is generally considered a good fit (Hu and Bentler, 1998). This study satisfies and ensures the goodness of fit.

This study conducts the PLS-SEM to test the theoretical model. The findings show (H1) that TL had a positive and significant direct impact on OCB (β = 0.169, t = 4.737, p < 0.000). The direct effects of TL on ER (H2) were also positive and significant (β = 0.356, t = 6.479, p < 0.000). Similarly, the direct impact of OCB or ER (H3) was positive and significant (β = 0.179, t = 2.203, p = 0.033). OCB as a mediator (H4) had a positive and significant direct impact on the relationship between TL and ER (β = 0.030, t = 2.169, p = 0.000). Similarly, the impacts of communication as moderator on the relationship of TL and ER (H5), OCB, and ER were found to be positive and significant (β = 0.183, t = 5.035, p = 0.000), and H6 was noted to be negative but statistically significant (β = −0.181, t = 3.373, p = 0.001) (Table 6 and Figure 2). The results were also supported by previous studies (Jiang et al., 2017; Khokhar and Zia-ur-Rehman, 2017; Majeed et al., 2017; Olendo and Muindi, 2017; Cofie, 2018; Maaitah, 2018; Mittal and Kaur, 2018; Sulamuthu and Yusof, 2018).

Figure 3 shows the interaction of communication on the relationship between ER and OCB. The lines on the graph show that if there is highly effective communication in SMEs, the moderation effect of communication will be higher, and the ER will be increased by OCB. Similarly, the Figure 4 shows the moderation effect in the relation of OCB and TL. ER will be increased in SMEs with effective communication.

First, the findings extend the research of TL. This study confirms that the ability of the leaders to articulate a compelling vision, skills at expressing confidence in goal achievement, innovative perspectives on problem solving, spending time on the training and development of the team, and specifically having a strong sense of purpose (Bass and Avolio, 1995) impact OCB (Rodrigues and Ferreira, 2015; Ismaeelzadeh et al., 2016; Saif et al., 2016; Majeed et al., 2017; Cofie, 2018) and ER (Amankwaa and Anku-Tsede, 2015; Khan, 2015; Jiang et al., 2017; Maaitah, 2018). This study illustrates the complexity of the relationship between OCB and ER. Results show that employees of SMEs express genuine courtesy toward coworkers, even under the most tiring business or personal situations, which helps to enhance OCB. Moreover, this study elaborates that if employees willingly help coworkers, defend the organization’s name, and express loyalty toward the organization, ER increases (Pickford and Joy, 2016; Olendo and Muindi, 2017; Mittal and Kaur, 2018).

Second, the results suggest that OCB has a decisive mediation role between TL and ER. The results show that transformational leaders can help increase ER more effectively if employees exhibit OCB (Jiang et al., 2017; Khokhar and Zia-ur-Rehman, 2017). This study has examined that communication, including the importance (to employees) of moving upward in the organization, the streams of information they communicate to their immediate boss, and their feelings about how their boss can help their career growth. This study has uniquely examined the effects of communication as a moderator between the relationships of TL, OCB, and ER. Unlike other studies, this study reveals the positive impacts of OCB as a mediator in TL and ER. The primary implication of this study is the crucial mediating role of OCB and the moderating component of communication; these findings contribute significantly to the existing literature.

This study offers a few practical implications on how TL facilitates the increase in ER. Specifically, it is essential to understand that the traits of TL develop skills in managers to retain employees. SMEs should train leaders to develop TL characteristics in them. Accordingly, leaders should try to implement the skills of TL such as developing a compelling vision for employees, focusing on goal achievement, having problem-solving techniques, having a sense of purpose, and spending time on the training and development of the team to enhance ER.

Furthermore, leaders should pay attention to develop OCB in subordinates. Importantly, when leaders want to increase OCB, they should practice traits of TL to help and support the employee in achieving their goals, practice different problem-solving methods and train employees to enhance OCB. Moreover, leaders should not neglect the importance of communication with their employees to communicate the responsibilities clearly, to listen to the problems and issues in task performance, and to manage them accordingly.

This study has a few limitations associated. First is the time limit bounded to obtain the maximum number of responses. This study is based on cross-sectional data, and more longitudinal studies are required to develop in-depth knowledge and to capture the relationships between variables as well as to check for differences in results if longitudinal data are used instead of cross-sectional data. The data for this study were gathered from 505 employees of SMEs in China, the sample size can be increased, and comparative analysis of the same model in private and public sector organizations can be checked. Moreover, the present study is exclusively focused on SMEs in China. Applying these results to different cultural contexts and populations may require appropriate alterations. Future research may include the study of other mediating variables, such as job satisfaction, deviant workplace behavior, and supervisor conflicts. Furthermore, there are multiple approaches to analyzing the relationships between the direct and indirect paths of the model.