The purpose of this study was to investigate the structural relationships between coaching leadership, psychological empowerment, innovative behavior, and leader-member exchange (LMX) among white-collar workers in large manufacturing corporations. To ...
The purpose of this study was to investigate the structural relationships between coaching leadership, psychological empowerment, innovative behavior, and leader-member exchange (LMX) among white-collar workers in large manufacturing corporations. To achieve this, the following research objectives were established: First, to verify the direct effects between coaching leadership, psychological empowerment, and innovative behavior. Second, to analyze the mediating effect of psychological empowerment in the relationship between coaching leadership and innovative behavior. Third, to verify the moderated mediating effect of LMX on the path from coaching leadership to innovative behavior through psychological empowerment.
The population of this study consisted of white-collar workers in large manufacturing corporations in South Korea. The criteria for large manufacturing corporations were set as manufacturing affiliates of the top 10 groups among the 92 business groups designated for disclosure in 2025. The target population was limited to regular employees performing functions such as business management/support, R&D, sales/marketing, manufacturing/production planning, and other office jobs within these affiliates. Based on the indirect estimation of the target population, the sample size required for hypothesis testing was set at 450.
The research instrument was a questionnaire consisting of 52 items to measure innovative behavior, coaching leadership, psychological empowerment, LMX, and demographic characteristics. All variables were measured using existing instruments whose reliability and validity have been verified in previous domestic and international studies. To ensure the suitability of the tools, content validity was verified by two experts in Human Resource Development (HRD),and face validity was tested by three white-collar workers in the manufacturing sector. Data collection was divided into a pilot study and a main study. The pilot study was conducted through an online survey for 10 days from August 18 to August 27, 2025, and 95 responses were used for analysis. The main survey was conducted for approximately three weeks from September 8 to September 26, 2025. To ensure a balanced sample across job functions, more than 50 responses were secured for each job category, and a total of 484 responses were used for the final analysis. Data analysis was conducted using the R 4.3.2 program. First, descriptive statistics such as frequency, percentage, mean, and standard deviation were analyzed. Second, difference analysis using t-tests and ANOVA was performed to identify demographic variables affecting innovative behavior. Third, correlation analysis was conducted to identify relationships between variables, and multicollinearity was verified through the Variance Inflation Factor (VIF) and tolerance. Fourth, common method bias was checked using Harman's single-factor test. Fifth, hierarchical regression analysis was used to confirm direct effects, and Hayes’ (2013) PROCESS Macro Models 4 and 7 were utilized to verify mediating and moderated mediating effects. The significance level was set at .05, and mediating effects were considered significant if the bootstrapping confidence interval did not include zero.
The main research results are as follows: First, the direct effects of coaching leadership on psychological empowerment and innovative behavior, as well as the direct effect of psychological empowerment on innovative behavior, were all statistically significant and positive. Second, sychological empowerment partially mediated the relationship between coaching leadership and innovative behavior. Third, the indirect effect of coaching leadership on innovative behavior through psychological empowerment was significantly moderated by LMX. The following conclusions were drawn from the results: First, the level of innovative behavior among white-collar workers in large manufacturing corporations varies depending on the leader's coaching leadership and the employee's psychological empowerment. When leaders provide direction, growth-oriented feedback, and support for autonomous decision-making, employees perceive higher levels of meaning, self-efficacy, self-determination, and impact. This suggests that innovative behavior is maximized when a leader acts as a facilitator who strengthens intrinsic motivation. Second, employees become psychologically empowered through coaching, which leads to innovative behavior by discovering meaning in their work. This indicates that coaching leadership promotes sustainable and voluntary innovation by changing the psychological state of employees. Third, even with the same coaching behavior, the effects differ depending on the quality of the relationship. The mediating effect was stronger in groups with high LMX, confirming that mutual trust and emotional bonds are essential for coaching to effectively stimulate psychological empowerment and innovation.
Based on these results, the following practical suggestions are provided: First, a multi-layered leadership development system should be established to systematically develop coaching competencies, including diagnostic tools and cyclical training. Second, an integrated HR system should be designed to strengthen psychological empowerment through job redesign, growth-oriented performance management, and competency development platforms. Third, differentiated development strategies should be established based on the quality of LMX, including 1:1 dialogue systems and informal communication channels. Fourth, a data-driven HR system should be introduced to monitor and manage key variables and implement customized programs. Suggestions for future research are as follows: First, comparative studies are needed to identify differences in coaching leadership effects according to industry and job types. Second, multi-level modeling research is required to consider the multi-layered impacts of individuals, teams, and organizations. Third, longitudinal research designs should be employed to verify causal relationships and temporal changes between variables over multiple time points.