The entrepreneurial environment of start-ups is chaotic and the responsibilities of team members are often unclear (Freeman and Siegfried Jr, 2015). Leaders must incentivize their employees to make the entire organization run in an orderly manner. EL plays the role of a “guiding light” for start-ups, providing support and vision to team members (Freeman and Siegfried Jr, 2015). The entrepreneurial leader has excellent job skills and personal charisma to accomplish the specific tasks and responsibilities of the start-up. Employees can establish achievable and innovative goals and strive for them with the guidance of such a leader. An entrepreneurial leader is also adept at seizing entrepreneurial opportunities and avoiding risks to achieve entrepreneurial goals by influencing the actions of their followers (Musara and Nieuwenhuizen, 2020). EL can greatly motivate employees and develop organizational commitment (Cai et al., 2019). Organizational development requires the continuous efforts of the entrepreneurial leader and their employees, as people in an organization are the drivers of creativity (Tóth et al., 2020).

EP is the extent to which a firm achieves a specific task or goal during the entrepreneurial process and is a key factor in explaining entrepreneurial outcomes and the enterprise's overall competitive advantage. EP can be interpreted as entrepreneurial willingness, the ability to identify opportunities, or business success (Baron, 2004). One study examined the impact of the interaction between EL and the creativity of entrepreneurial team members on the innovative capacity of new firms using patent creation as a measure (Chen, 2007). Previous studies point to the positive impact of entrepreneurial orientation on EP (Rauch et al., 2009), and the effect of authentic leadership on EP has been confirmed (Shirey, 2006). In an unpredictable environment, a company's survival depends on managers' entrepreneurial spirit and leadership (Demartini and Beretta, 2019). EL is particularly important in the start-up phase of a new business, focusing on empowering employees, which helps to enhance their self-efficacy and entrepreneurial abilities (Guberina and Wang, 2021). EL will motivate employees to be more proactive in pursuing organizational goals, thus improving organizational performance and value creation (Alsharif et al., 2021). Therefore, this study posits that there is a positive correlation between EL and EP, hypothesized as follows.

An organization's ability to innovate and sustain a competitive advantage is inextricably linked to knowledge sharing (Terhorst et al., 2018). Knowledge sharing among individuals is the key to creating value, gaining a competitive advantage, and improving overall organizational performance (Obrenovic et al., 2020). Knowledge sharing can be classified as either explicit and tacit (Zhao et al., 2020). Tacit knowledge refers to the skills accumulated by employees during their tenure and is abstract expertise (Obrenovic et al., 2021) which facilitates the development of employees and companies. Tacit knowledge is highly personalized, based on the accumulation of personal experience, and influenced by personal subjective factors (Wang et al., 2006). The cost of building tacit knowledge for individuals is high (Kogut and Zander, 1992), so employees need motivation to share tacit knowledge. Leadership can motivate employees to actively provide and share their ideas (Edú-Valsania et al., 2016). A study on small craft enterprises in Bali demonstrated a positive association between EL and knowledge sharing (Riana et al., 2020). EL is characterized by flexibility, patience, persistence, adventure, tenacity, self-confidence, motivation, and initiative (Renko et al., 2015), improving employees' trust in their team and leaders and making them more willing to share tacit knowledge. Knowledge sharing contributes to innovation (Fernandes Crespo et al., 2021), and is a key part of knowledge management; this helps organizations achieve a sustainable competitive advantage (Han et al., 2020), and contributes positively to organizational performance (Orpipath, 2018). Leaders can positively contribute to employee tacit knowledge sharing (TKS) by providing incentives and empowerment (Obrenovic et al., 2021). A recent study explored the role of tacit knowledge in driving business model innovation among team members and team leaders through an Italian engineering consulting firm (Castellani et al., 2021). Therefore, this study argues that TKS mediates the positive influence of EL on EP. The following hypothesis is proposed.

Job embeddedness (JE) refers to the combination of forces that prevent individuals from leaving their jobs (Mitchell et al., 2001). Specifically, it includes a set of social, economic, and psychological factors that influence individuals to stay in their jobs (Kiazad et al., 2015). JE is important in organizational settings because it refers to the factors which motivate employees to continue working (Karatepe, 2013) and to stay in their current organizations. JE is often used to explain employees' turnover intention (Mitchell et al., 2001). Scholars have found that JE can also be used to predict other job-related results (Ghosh et al., 2017).

Organizational leadership has been proven to positively impact employees' JE (Holmes et al., 2013). JE is considered a beneficial state to both the organization and the individual. Employees with high JE tend to have higher work ethic, superior productivity, and excellent innovative behavior (Norouzinik et al., 2021). According to JE theory, a meaningful connection and good fit between the employee, the organization, and the job will motivate good performance (Lee et al., 2004). JE reflects the relationship between leaders and followers, and the behavioral characteristics of leaders influence employees' JE (Mitchell et al., 2001). High JE employees go the extra mile to stay in the organization by performing better and achieving their personal career goals (Al-Ghazali, 2020). They develop more positive qualities and create value for the organization (Norouzinik et al., 2021). JE mediates between job factors and key organizational results (Holtom et al., 2006). Therefore, it is believed that JE mediates between EL and EP.

The more connected people are to others, their workplace, and their community, the less likely it is that they will leave their jobs (Lee et al., 2004). Employees improve their JE when they feel socially connected and when their skills and knowledge are enhanced (Karatepe, 2013). When employees are in an atmosphere where they are willing to share tacit knowledge, they will inevitably strengthen their attachment to the organization and will not leave it. For knowledge sharers, when the organization or their colleagues recognize the knowledge they contribute, it enhances their reputation, gains them respect, and makes them more enthusiastic about their work. As was mentioned earlier, information exchange is a type of social exchange. When employees feel that the tacit knowledge shared by their colleagues is beneficial to them, they have a reason to stay in the organization. In return, they will contribute more to the organization. This study investigates a link between TKS and JE when mediating EL and EP, and proposes the following hypothesis.

Career growth opportunities (CGO) are opportunities provided by managers for employees to enhance their job competence and expertise, allowing the employees more room for development and responsibility at work. They are a kind of stimulation and motivation for employees (Weer and Greenhaus, 2020). Employees can keenly perceive the CGO of their organizations (Weng and McElroy, 2012). Career-focused employees are more likely to pursue performance and develop a desire for advancement than less career-focused employees. Leaders or internal incentives strongly influence these employees, and corporate leaders should strive to take advantage of such strong motivation (Weer and Greenhaus, 2020). From the perspective of self-determination theory (Deci et al., 1994), CGO is an informational environmental factor. When employees acquire or perceive better CGO, they are more enthusiastic about their work, more comfortable with their work, and capable of controlling their work resources. Specifically, they are actively engaged, energetic, and committed at work (Schaufeli and Bakker, 2004).

In general, the more CGO employees perceive in the organization, the higher the organization's care and support for the employee (Lu et al., 2016). The occurrence of employees' TKS comes from a combination of personal and external factors. The factors that influence unique behaviors in a given situation result from the interaction of individual characteristics and external conditions. Personal factors are generally inherent and unshakable, but external factors with motivational effects have room for improvement. Incentive policies offered by organizations can effectively promote knowledge sharing (Orpipath, 2018). CGO provides a unique environment for organizations that may influence the effectiveness of leaders through the attitudes and behaviors of subordinates (Bashir et al., 2020). When employees perceive that the organization provides more CGO, they develop positive work emotions and form psychological attachments (Wang et al., 2014). A Chinese proverb says, “A drop of water should be repaid with a spring.” When employees are aware of the “water” provided by the organization, they will naturally be more willing to reciprocate and are more inclined to share tacit knowledge with their colleagues, thus promoting their sense of belonging to their jobs and improving their performance. This is also reflected in the reciprocity principle of SET: when individuals feel support from the organization, they have a stronger sense of belonging and a deeper sense of identity. Furthermore, when employees feel motivated by the organization's supportive measures, an employee will gladly perform TKS to promote the organization's progress. Employees will feel obligated to remain loyal to the company (Bashir et al., 2020) and stay embedded in the organization. Subsequent research reaffirms this (Lu et al., 2016), and additionally exemplifies the “links” in JE theory (Mitchell et al., 2001). Bashir et al. (2020) confirmed that CGO moderates the relationship between procedural justice and organizational commitment. Therefore, this study makes the following hypotheses.

To sum up, our conceptual model is shown in Figure 1.

The survey data was collected through the online survey website “Wenjuanxing”. Participants were employees of Chinese start-ups established in the past 5 years. Several efforts were made to reduce common method variance. Firstly, to weaken the chance of comparison among respondents and improve the data quality of this study, the data of no more than ten employees per company was collected. Secondly, before data collection, participants were informed that filling out the questionnaire is anonymous and that the information they provide will be used for research only. Thirdly, Internet protocol access was restricted to avoid repeated surveys. Employees were asked about their experiences with EL and their shared work experiences, including TKS, JE, CGO and EP. This survey was closed after collecting 500 responses. After excluding questionnaires with responses that were too long or too short and contradictory answers, 376 valid questionnaires were obtained, representing a validity rate of 75.2%. Of these, 136 respondents (36.2%) were male, and 240 (63.8%) were female. Respondents were mainly between 21 and 40 years old (96.5%). The majority of respondents had a bachelor's degree or higher (79.5%), and most respondents had a monthly salary of between ¥3,000 and ¥8,000 (60.9%). Respondents came from various companies, most of which had fewer than 100 employees (75.5%).

The maturity scales published in national and international journals were used in this study. All English scales were translated using a standard translation back-translation procedure to avoid semantic ambiguity to the greatest extent possible. Second, given the sensitivity and covert nature of corporate surveys, all variables in this study were appropriate for use in an informal setting and were measured by anonymous self-assessment. All items were assessed using Likert 7-point scales, ranging from 1 (strongly disagree) to 7 (strongly agree).

The data analysis process for this study was as follows. First, validation factor analysis was performed using Amos 23.0 to test the discriminant validity between EL, TKS, JE, CGO, and EP. Second, the relationship between the variables was analyzed using SPSS 21.0. Third, hierarchical multiple regression analysis (Baron and Kenny, 1986), structural equation modeling, and SPSS PROCESS Macro (Hayes, 2013) were used to test the hypotheses.

Because the EL, TKS, JE, CGO, and EP data come from the same source, the possibility of a common method variance existing between the five constructs could not be disregarded. Firstly, the Harman single factor method was used to test common method variance (Podsakoff et al., 2003). After unrotated exploratory factor analysis for all items of the study variables, the total variance explained by all factors with characteristic roots greater than 1 was 59.6%. The various explanation of the first principal component was 30.6%. The variance explained by this factor did not exceed 40%. There is no serious common method variance problem in which a single factor explained most of the variance of all variables. In addition, to estimate the homoscedasticity bias and measure the discriminant validity of the variables, a confirmatory factor analysis was performed using Amos 23.0 (Table 1). Based on the results of the validation factor analysis and the corrected indicators for each indicator variable, some items were removed so that all questionnaire items had factor loadings greater than 0.5 (Hair et al., 2014). According to the standard for model fit: χ ²/df < 3 (Kline, 2015); GFI > 0.80 (Doll et al., 1994); AGFI > 0.80 (MacCallum and Hong, 1997); RMSEA < 0.08 (Hu and Bentler, 1999); TLI > 0.90 (Hu and Bentler, 1999); CFI > 0.900. Table 1 shows that the five-factor model outperformed the other models (χ ²/df = 1.757; GFI = 0.931; AGFI = 0.909; CFI = 0.966; RMSEA = 0.025; TLI = 0.960).

Table 2 shows the mean, standard deviation, correlation coefficient, average variance extracted (AVE), and composite reliability (CR) of all variables involved in this study. There is a significant positive correlation between EL and TKS (r = 0.541, p < 0.001), JE (r = 0.569, p < 0.001), CGO (r = 0.587, p < 0.001), and EP (r = 0.606, p < 0.001). The results of variable correlation analysis preliminarily verify the research hypotheses of this paper. The AVE from all variables (except EP) is >0.5, which is an acceptable value (Fornell and Larcker, 1981), and the CR is >0.7 (Fornell and Larcker, 1981) (Table 2). Although the AVE of EP is 0.495, its CR is 0.796. The convergence validity of this variable meets the requirements (Fornell and Larcker, 1981).

Following the suggestion of Baron and Kenny (1986), hierarchical regression analysis was applied to test hypothesis 1 (Table 3). The results showed a significant positive association of EL on EP (M6, β = 0.866, p < 0.001). Furthermore, a structural equation model was used to test this hypothesis. The main path coefficient (γ) of this model is shown in Figure 2, and EL has a significant positive association on EP (γ = 0.443, p < 0.001). Therefore, hypothesis 1 is accepted.

Table 3 shows that there was a significant positive association of EL on both TKS and JE (M1, β = 0.821, p < 0.001; M3, β = 0.794, p < 0.001). When both independent (EL) and mediating variables (TKS and JE) were added to the analysis of EP, both EL (M7, β = 0.609, p < 0.001), TKS (M7, β = 0.106, p < 0.05) and JE (M7, β = 0. 214, p < 0.001) showed a significant positive association on EP. Hayes' (2013) SPSS PROCESS Macro (model 6) was used to obtain a preliminary idea of the direct vs. indirect associations in the model. Table 4 again validates hypothesis 1 and points to the presence of a mediating association of EL in affecting EP [b = 0.257, SE = 0.058, 95% CI = (0.144, 0.372), excluding 0].

Model 83 was used in SPSS PROCESS Macro (Hayes, 2013) to test the mediating (H2, H3, and H4) and moderating hypotheses (H5, H6, and H7). The effect of EL on TKS was observed under two different CGO according to the procedure recommended by Cohen et al. (2014). The positive effect observed between EL and TKS was more significant when employees were in a high CGO environment (Figure 3). Therefore, hypothesis 5 was supported.

Table 5 shows the mediating association of TKS and JE separately, and the serial mediating association of TKS and JE. The mediating association of TKS between EL and EP was low [b = 0.050, SE = 0.026, 95% CI = (0.004, 0.104)], medium [b = 0.069, SE = 0.031, 95% CI = (0.008, 0.131)] and high [b = 0.087, SE = 0.040, 95% CI = (0.010, 0.168)] (Table 5). The mediating association of JE between EL and EP was significant with a 95% bootstrap confidence interval of 0.088 to 0.269 [b = 0.172, SE = 0.046, 95% CI = (0.088, 0.269)]. The serial mediating association of TKS and JE between EL and EP was low [b = 0.025, SE = 0.011, 95% CI = (0.007, 0.049)], medium [b = 0.034, SE = 0.012, 95% CI = (0.014, 0.060)] and high [b = 0.043, SE = 0.016, 95% CI = (0.017, 0.079)]. None of these confidence intervals include 0, so hypothesis 2, hypothesis 3, and hypothesis 4 were confirmed. Furthermore, Table 5 shows that the mediating association of TKS and the serial mediating association of TKS and JE were strengthened with the elevation of CGO. Therefore, hypothesis 6 and hypothesis 7 are supported.

Many current discussions on authentic leadership (Edú-Valsania et al., 2016) and ethical leadership (Bavik et al., 2018) have taken place, while there have been insufficient discussions of EL. There is still a lack of research analyzing the effects of EL on EP, but this paper starts to fill this gap. First, this study establishes a moderated serial mediation model of EL and EP. This study also validates the mechanism by which EL influences EP from multiple paths, remedying the fact that the extant literature has mostly examined the influence of leadership style on EP from a single level, and provides empirical support for the improvement of EP in start-ups. The initial exploration of EL as an example in Chinese start-ups should complement and extend to the current EP field. These findings broaden the range of predictors of EP and suggest the significance of leadership behavior on employee behavior and psychology. Second, the study highlights the role of TKS. Knowledge sharing is critical to high performance environments (Riana et al., 2020). Third, it expands the research on JE. Most of the previous studies on JE have focused on intention, while ignoring the critical role of JE in analyzing employee job outcomes (Lee et al., 2004; Holtom et al., 2006). Fourth, in analyzing the relationship between EL and EP, this study introduces CGO as a moderating variable in the first stage based on SET, which is a significant innovation. This study clarifies the theoretical model of boundary conditions by using CGO as a moderating variable for the entire research framework. It also enhances the contextual features of the study (Lu et al., 2016). Finally, in the extant literature on EL, case studies have been common (Wang et al., 2012). This paper is a quantitative study, which improves the persuasiveness of the causal relationship between the variables of interest.

This study points out the importance of EL for start-ups. A creative atmosphere is a work environment that should be cultivated (Edú-Valsania et al., 2016), and EL provides this environment for companies. EL can encourage employees to do more useful things for the organization and make them more trusting in the organization, more willing to perform TKS, and more dependent on the work and the company. Managers should exercise their ability to actively promote TKS and integrate knowledge management goals into corporate strategic objectives to improve organizational effectiveness (Castellani et al., 2021) and corporate performance. Firms need to provide employees with opportunities to build and protect job resources, such as employment security, social support, and career development opportunities, to increase their JE and subjective well-being (Ampofo et al., 2017). Employees who have a high sense of belonging to their jobs make more significant contributions to the organization. Business leaders need to focus on non-material incentives to keep employees active. This study points out the importance of CGO, which stimulates positive psychological effects for employees such as increased self-confidence and optimism, and increases job satisfaction (Kahn, 1990). This can allow companies to motivate their employees to hold positive work attitudes, creating an appropriate organizational climate and strong team cohesion.

Although this study contributes significantly to the study of EL, it still has the following shortcomings. First, this study is a cross-sectional study, which cannot determine causality. A longitudinal study should be designed to explore this. Second, the subjects of this study were all employees, and leaders could be included as participants in future research. Third, although this study adopted a series of methods to weaken the homoscedasticity bias of the questionnaire, it still cannot be completely avoided. Fourth, other relevant variables may be selected as moderating variables in future research. Finally, this study only discusses the factors that play a positive role in EL's influence on EP, and subsequent studies may choose to discuss the negative factors affecting this mechanism.