Talent management practices and turnover intention: The role of perceived distributive justice and perceived organizational support

Abstract This study aims to analyze the effect of talent management practices on turnover intention either directly or through perceived distributive justice and perceived organizational support as mediation. The research was conducted using a quantitative approach with total sample of 397 Directorate General of Taxes employees. Analysis of research data using SMART-PLS with the result that there is an influence of talent management practices on turnover intention through perceived distributive justice and perceived organizational support. Talent management practices have been shown to increase perceived distributive justice and perceived organizational support. Turnover intention decreases because of perceived distributive justice and perceived organizational support by employees. The results of this study make a theoretical contribution to support social exchange theory, equity theory, and organizational support theory as well as make a practical contribution to HR managers of public organizations. The novelty of this research is in the form of a comprehensive model with perceived distributive justice and perceived organizational support as mediating the influence of talent management practices on turnover intention in public organizations. The results of this study answer the scarcity of studies on public organizations in Asia. The results of this study also add to the repertoire of empirical research which can be used as a reference in drawing generalizations or conclusions regarding the influence of talent management practices on turnover intention.

Talent management is an important part of human resource management which focuses on the process of selecting, developing, and retaining high potential employees.Initially, talent management was mostly applied in private organizations, but recently public organizations have begun to adopt this practice.The application of talent management is expected to provide positive attitudes and effective work behavior, including reducing employees' desire to look for alternative jobs outside the organization.This research helps outline the impact or influence of talent management practices on turnover intentions in public organizations both directly and through perceived distributive justice and perceived organizational support.The results of this research can add to the body of empirical research regarding the influence of talent management practices in public organizations because most research is conducted in the private sector.

Introduction
Talent management is one of the most popular HR management practices applied to many organizations after the publication of the talent war introduced by McKinsey around 1997.Talent management is an important element for modern organizations (Lewis & Heckman, 2006) and (Collings & Mellahi, 2009).Increased attention to talent management is also driven by the notion that the resources owned by an organization are available in limited quantities, including human resources.The success of an organization is largely determined by the human resources in it, especially the human resources who occupy key or strategic positions.In today's vibrant, dynamic, and competitive era, attention to talent management is increasing because talent management is one of the most important keys in human resource management (Raja et al., 2021).Attention to the field of talent management has been shown by the existence of research interest in the field of talent management.Some researchers see talent management as a new technique in human resource management that focuses on recruitment, retention, development of leadership competencies, career paths, succession plans, and others (Collings & Mellahi, 2009;Gallardo-Gallardo et al., 2015;Thunnissen et al., 2013).While other opinions say that talent management is futureoriented and fulfills strategic goals (Lewis & Heckman, 2006;Vaiman & Collings, 2013).
Along with the importance of implementing talent management, research interest in talent management has soared over the last two decades and empirical evidence has started to increase since 2012.On the other hand, it is known that despite the increasing scientific attention in the field of talent management over the last few years, questions remain, in particular, those related to how and why talent management is structured, implemented, and developed in organizations, not to mention its results or effectiveness (Gallardo-Gallardo et al., 2020) and especially in public organizations.The implementation of talent management in the public sector or government agencies is intended to encourage organizational performance both in terms of providing public services and other performance.Dhanabhakyam and Kokilambal (2014) revealed that talent management allows public service to be maintained and attracts skilled leaders.Strong and progressive talent management will better equip public services to have the right people, in the right place, at the right time.The implementation of good talent management is expected to have a positive impact on both employees and organizations.
A review of empirical research conducted by Gallardo-Gallardo et al. (2017) shows that only a few talent management publications focus on public sector organizations.Academic attention regarding the impact of talent management on public sector organizations is still very limited.The results of Barkhuizen (2014) confirm previous research showing that talent management practices do not receive sufficient attention in the public sector (see Barkhuizen et al., 2014;Masibigiri & Nienaber, 2011;Theron et al., 2014).All talent management practices were poorly implemented in the local government institutions investigated.Some of the most problematic talent management practices including poor ones were workforce planning and talent retention strategies.Bui and Chang (2018) revealed that empirical research related to talent management is more focused on the private sector, resulting in a lack of theoretical developments and empirical research examining how talent management works in public organizations, especially in developing countries.
Employee reactions to the implementation of talent management refer to all the positive attitudes and effective work behaviors shown by employees identified as talents in response to their organization's talent management (De Boeck et al., 2018).Turnover intention is a reaction that is considered important along with the increasing phenomenon of employee turnover.Turnover intention is considered an important reaction because turnover intention is one of the best predictive methods of employee turnover (Barrick & Zimmerman, 2005) which is easier to measure than measuring actual turnover (Firth et al., 2004).There is still a research gap regarding the influence of talent management practices on turnover intention.Research conducted by Kaur (2013), Barkhizen et al. (2015), Latif et al. (2019) found a significant effect of talent management practices in reducing the intention to leave.Likewise, Dayeh and Farmanesh (2021) found a significant negative effect of talent management on turnover intention in the banking sector in Jordan.Another study conducted by Rana & Abbasi (2013), Kekgonegile (2014), Hoogervorst (2009) found different results that there was no significant effect of talent management practices on turnover intention.Bui and Chang (2018) found that hard and soft talent management had a positive effect on turnover intention.Research by Ogbeibu et al. (2021) revealed the influence of green hard talent management on encouraging turnover intention, as well as green soft talent management positively influencing turnover intention.The inconsistency of the effect of talent management practices on turnover intention has encouraged this research to try to close the research gap by proposing a mediating variable.The need for a mediating variable in this study is also driven by the condition that so far scholars have implicitly assumed that talent management leads to employee attitudes, but the unilateral causal and effect method may not be satisfactory (Swailes et al., 2016) and (Thunnissen et al., 2016).
Referring to equity theory (Adams, 1963), perceived distributive justice is a variable that is seen as appropriate as mediation because in principle if an employee feels the distribution of resources is unfair, then there will be an intention to leave voluntarily (Joseph, 2018).Thurston and Mcnall (2010) stated that talent management procedures that are implemented fairly will reduce the impact of distributive justice felt by employees.Employees who feel that their contributions are not reciprocated by the organization will tend to see distributive justice that is less favourable (Gelens et al., 2014) and conversely, those who are selected as talents will feel greater distributive justice.Gelens et al. (2013) also encourage human resource practitioners to assess the impact of their talent management practices on employees instead of focusing solely on organizational results.The recommendation given is that researchers can study how large the differences in perceptions of distributive justice are for unequal allocation of resources to employees with high potential and not high potential in various organizations.In addition, employee perceptions regarding fairness and equality depend on the national culture and also the talent management strategy implemented (Agarwal, 2016).
Employee reactions are a form of social exchange where employees will give positive reactions such as dedication and loyalty when they feel that the organization provides appreciation and attention (Meyers et al., 2017).One model of social exchange according to Cropanzano and Mitchell (2005) is perceived organizational support.Perceived organizational support is seen as the next mediating variable because concerning organizational support theory as stated by Orpen (1994) that employees tend to develop stronger bonds and trust their organization if they are aware that their efforts and contributions are valued by the organization (Nguyen et al., 2022).With these exchanges, employers and employees navigate a two-way street of social exchange where both parties can mutually benefit (Dawley et al., 2010).Organizations that are involved in HR practices in the form of investment and support for employees indicate that the organization seeks to continue social exchange relations with its employees (Allen et al., 2003).
This study aims to analyse the influence of talent management practices on turnover intention either directly or through mediation.This research was conducted in line with the phenomenon of the importance of managing and retaining employees to survive considering that the Directorate General of Taxes has very strategic duties and responsibilities to collect tax revenues.This research is important to do because as stated Dayeh and Farmanesh (2021) several studies related to the talent management process that lead to outcomes are still limited, and Ogbeibu et al. (2021) also revealed that referring to the literature related to talent management and turnover intention cannot be concluded because the research findings are still diverse.The results of this research will answer the scarcity of talent management studies in public organizations and add to the repertoire of empirical research because various empirical studies have limitations in generalizing to various sectors, so this research responds to the suggestions of Gelens et al. (2015) to conduct multilevel studies in different types of organizations and across cultures.This research has originality or novelty that is different from previous studies in the form of a comprehensive model of the influence of talent management practices on turnover intention through perceived distributive justice and perceived organizational support as mediating in public organizations.The research conducted at the Directorate General of Taxes has never been done before and is an interesting matter because the Directorate General of Taxes is one of the pioneers in implementing talent management in public organizations in Indonesia.The results of this study will at the same time answer the scarcity of talent management studies on public organizations in Asia.Dessler (2017) defines talent management as a holistic, integrated, and result-oriented process and goals of planning, recruiting, selecting, developing, managing, and compensating employees.Talent management practices are very different in every company with different approaches to talent management in different ways (Collings & Mellahi, 2009;Lewis & Heckman, 2006).Two approaches are often used in talent management, namely the exclusive and inclusive approaches (De Boeck et al., 2018).The exclusive approach believes in the difference between employees who are included in the talent and non-talent categories with the number of talent employees being only a small number of all employees.The exclusive approach refers to a resource-based view, that is, with limited resources, organizations pay more attention to talented employees than non-talented employees.An inclusive approach views that all employees have their talents and can contribute according to their respective talents (Gallardo-Gallardo & Thunnissen, 2016;Van den Brink et al., 2013).

Talent management practices and turnover intention
Talent management practices are a series of processes, programs, and cultural norms that are integrated within an organization that is designed and implemented to attract, develop, place, and retain talent to achieve strategic goals and future business needs (Silzer & Dowell, 2010).Every organization has the challenge of recruiting, developing, and retaining employees (Raja et al., 2021).Barkhizen et al. (2015) identified talent management practices consisting of management commitment, talent commitment, talent development, performance management, talent acquisition/recruitment, staffing, workforce planning, talent review process, and talent retention.Zhang (2014) identified talent management practices consisting of attracting talent, developing talent, and retaining talent.
Various outputs resulting from the implementation of talent management include commitment, satisfaction, attachment, perceived superior support, and motivation (Collings & Mellahi, 2009).Barkhuizen et al. (2014) suggested that the direct results of talent management are employee satisfaction, motivation, commitment, and perceived organizational support.De Boeck et al. (2018) identified employee reactions to the implementation of talent management which included affective reactions, namely commitment, satisfaction, motivation, involvement, trust, well-being, stress, insecurity, and identity struggles; cognitive reactions namely belief in knowledge skill attitude (KSA), development, psychological contract beliefs, employer's obligations, fulfillment of psychological contracts, breach of psychological contracts; and behavioral reactions, namely turnover intention, transfer of work, performance efforts, organizational citizenship.
The implementation of talent management has developed widely, not only in private organizations but also in public sector organizations.Talent management enables public services to be maintained and attracts skilled leaders and strong and progressive talent management will further complement public services to get the right people, in the right place at the right time (Dhanabhakyam & Kokilambal, 2014).The influence of talent management practices on turnover intention has been proven in several previous studies such as Kaur (2013) with the results that almost all respondents were satisfied with the results of the talent management program and almost all employees agreed that talent management programs created a good work culture and increased employee value and reduce employee turnover.Research by Barkhizen et al. (2015) at the Botswana National Education Commission produced several findings, namely the existence of teacher talent management practices that were poorly implemented regarding workforce planning, performance, and talent retention, and talent management practices were inversely related to the intention to quit teachers in Botswana.Dayeh and Farmanesh (2021) found a significant negative influence of talent management on turnover intention in the banking sector in Jordan.Latif et al. (2019) tested the effect of talent management on turnover intentions with the result that talent management was proven to influence employee turnover intentions in the Malaysian ICT industry.Referring to the social exchange theory and supported by the results of empirical research, the hypotheses built in this study related to the relationship or influence of talent management practices on turnover intention is: H1: Talent management practices have a significant effect on turnover intention.

Talent management practices, perceived distributive justice, and turnover intention
Distributive justice is rooted in equity theory (Adams, 1965) which states that social exchange underlies the relationship between workers and employers (Biswas et al., 2012).Distributive justice is part of organizational justice which refers to the allocation of outcomes such as promotion opportunities or financial rewards.Organizational justice itself consists of several sub-dimensions including distributive justice.O'Connor and Crowley-Henry (2014) identified 3 (three) forms of organizational justice, namely distributional justice which is defined as distribution and allocation justice, procedural justice which is perceived justice and allocation process (Gelens et al., 2014), and interactional justice as the quality of interpersonal treatment people receive when procedures are applied.Furthermore, O'Connor and Crowley-Henry (2014) stated that career development interventions are vulnerable to perceptions of distributive, procedural, and interactional justice.Another important input noted is how various forms of fairness moderate and mediate the effects of organizational practices (Gelens et al., 2014), such as when employees who are not considered "talented" perceive procedural and appropriation processes to be fair and thus the possible negative impact of not being selected as a high potential employee is diminishing.Perceived distributive justice is an employee's response to the results received from the organization based on the efforts that employees have made to the organization compared to other employees to determine the level of fairness (Gelens et al., 2014).
To understand how employee reactions to talent management are shaped, perceived organizational justice theory offers a suitable framework (Gelens et al., 2013).Perceived distributive justice shapes the behavioral, cognitive, and emotional reactions of employees (Cropanzano & Mitchell, 2005).Gelens et al. (2013) stated that employees who are categorized as low potential employees (not talented enough) will compare their efforts and rewards with high potential employees and if they see that they have been treated unfairly, this can hurt their efforts and job satisfaction.Based on the study of distributive justice conducted by Folger and Konovsky (1989), McFarlin and Sweeney, (1992), and Sweeney and McFarlin (1993) found that distributive justice have good individual-level consequences as a predictor of individual reactions to certain outcomes such as job satisfaction, salary satisfaction, and intention to stay with the organization.
Talent management practices will be seen by employees whether these practices are fair practices or not.Talent management practices that are carried out with due regard to the principle of justice are expected to encourage employees to have a good perception of distributive justice.Perceptions of talented and non-talented employees may be different.Schumann (2001) argues that employees must have equal opportunities to demonstrate their potential to contribute.Employees who don't have such an opportunity can get the impression that the contributions and results are invalid and based on unrepresentative information.Talent management practices that are implemented properly are expected to encourage all employees, both talented and nontalented, to feel that there is distributive justice.
The influence of perceived distributive justice on turnover intention was also stated by Cropanzano and Mitchell (2005) where perceptions of distributive justice shape the behavioral, cognitive, and emotional reactions of employees.Talented and non-talented employees can have different perceived distributive justice as research by Timinepere et al. (2018) who revealed the results of differences in perceptions of organizational justice between medical doctors, paramedics, and support staff in teaching hospitals in Nigeria.These differences in employee status drive differences in perceived distributive justice.Research by Rahim et al. (2001) revealed that distributive justice has the strongest negative relationship with the intention to change jobs.This is in line with the results of research by Khan et al. (2013) which states that high perceived distributive justice can help reduce employee turnover intentions.Likewise research by Raza et al. (2017) found that distributive justice had a significant negative relationship with employee intention to leave.Some of these studies can be used as a reference for the influence of perceived distributive justice on employee turnover intentions.Referring to social exchange theory and equity theory and supported by empirical research results, the hypotheses built in this study are: H2: Talent management practices have a significant effect on perceived distributive justice.

H3:
Perceived distributive justice has a significant effect on turnover intention.H4: Perceived distributive justice mediates the effect of talent management practices on turnover intention.

Talent management practices, perceived organizational support, and turnover intention
Perceptions of organizational support refer to employees' perceptions of the extent to which the organization values their contribution and cares about their well-being.Perceived organizational support is defined as employees' beliefs about the extent to which their organization values their contributions and shows concern for employee welfare (Rhoades & Eisenberger, 2002).Employees' perceptions of the organization providing support for HR practices that signify investment in employees and recognition of their contributions (i.e.participation in decision-making, fairness of rewards, and growth opportunities) contribute to the development of perceived organizational support.Based on the theory of organizational support (Eisenberger et al., 1986), Rhoades and Eisenberger (2002) suggested 3 (three) aspects of perceived organizational support, namely justice received, superior support, and rewards and organizational working conditions and the consequences of perceived organizational support include commitment organization, work-related influences, job involvement, performance, strain, desire to stay, and withdrawal behavior.Collings and Mellahi (2009) state that talent management practices can trigger perceived organizational support because they are adopted as a result of deliberate efforts to attract, select, develop, and retain the best employees in the most strategic roles.Talent management is a supportive human resource practice that can influence employees' perceptions of their organizational support (Mensah, 2015).Research by Gelens et al. (2015) concluded that employees who were designated as high potential employees felt greater perceived organizational support than those who were not designated as high potential.However, Gelens et al. (2015) revealed that the results of this research cannot yet be generalized to sectors other than finance and companies with different organizational and national cultures or different talent management approaches.Likewise, the research conducted by Plessis et al. (2015) and Mensah (2019) needs to be continued with further research to ensure the causality of the relationship and influence.
Another study related to the effect of talent management practices on perceived organizational support was conducted by Isa et al. (2018) who found the effect of talent identification and talent development on perceived organizational support.Talent identification and talent development are talent management practices that are often carried out by many organizations.Sari et al. (2022) also revealed research results that support that there is a significant effect of talent acquisition and talent pool on perceived organizational support.Talent management practice is basically a form of organizational support and encouragement to employees in career development.The practice of talent management is expected to increase the organizational support felt by employees with both talent and non-talent status.
Research by Eisenberger et al. (2002) revealed that an increased sense of reciprocity and perceived organizational support encourages employees' willingness to stay.Research by Allen et al. (2003), Dawley et al. (2008), Dawley et al. (2010) yielded findings consistent with the perceived organizational support has a significant negative relationship with the intention to leave.Another research conducted by Islam et al. (2013) and Takaya and Ramli (2020) confirmed the results of previous research regarding the effect of organizational support on the intention to leave.Some of these empirical studies state that they cannot be generalized due to limitations, so this research needs to be carried out to respond to these conditions.Based on the results of this empirical research, the logic built in this study is that the existence of support from the organization will turnover intention.The opposite condition will occur if employees do not feel organizational support, this will encourage an intention to leave.Referring to social exchange theory and organizational support theory and supported by empirical research results, the hypotheses built in this study are: H5: Talent management practices have a significant effect on perceived organizational support.

H6:
Perceived organizational support has a significant effect on turnover intention.H7: Perceived organizational support mediates the effect of talent management practices on turnover intention.
An integrated conceptual framework for the relationships between variables in this research is presented in Figure 1.

Participants and procedure
The population of this study is all employees of Directorate General of Taxes who have not participated in the talent management program and employees who have participated in the talent management program but have not been selected as talent ready now.The selection of this population is based on the consideration that talent management practices must have a good impact on all employees, both talented and non-talented.This focus on non-talented employees is also an answer to the scarcity of empirical studies because previous research tends to avoid the reactions of non-talented employees.The sample was calculated using the proportionate random sampling method with the Slovin formula which resulted in a total sample of 397 employees from all work units of the Directorate General of Taxes.The survey was carried out from mid-October to the end of November 2022 via online media.Submission of the link survey was assisted by the human resource manager in each respondent's work unit.

Measurement
This research questionnaire was developed from previous research using a Likert scale of 5 (1 = strongly disagree to 5 = strongly agree).Talent management practices were measured by 26 statement items consisting of 5 statement items for the talent review process from Lerotholi (2021) with an example of a specific item "The Directorate General of Taxes knows the bestperforming employees", 10 statement items for talent development from El Nakhala ( 2013 Perceived distributive justice was measured by three item statement from Gelens et al. ( 2014) and Loi et al. (2009).An example of a statement item for perceived distributive justice is "I feel that the award I received from the Directorate General of Taxes has reflected the efforts I have made".To measure perceived organizational support we adapted eight statement items from Gelens et al. (2015) and Eisenberger et al. (2002) with the example of a statement item is "The Directorate General of Taxes appreciates the extra effort made by employees".Turnover intention was measured by three statement items from Bui and Chang (2018) and Rahman and Nas (2013) with an example of a statement item such as "I often think about quitting from the Directorate General of Taxes".All statement items have a Cronbach's alpha value > 0,6 as follows, Cronbach alpha talent management practices was 0,977, perceived distributive justice was 0,958, perceived organizational support was 0,958, and turnover intention was 0,898.Therefore all statement items are declared reliable.

Outer model analysis
Data analysis in this study included descriptive statistic analysis and inferential statistical analysis using PLS software tools version 3.4.1.Inferential statistical analysis was carried out in two stages, namely evaluating the outer model and evaluating the inner model.Evaluation of the outer model for a reflective construct consists of four evaluations, namely: internal consistency (composite reliability), indicator reliability, convergent validity (average variance extracted), and discriminant validity (Hair et al., 2021) with the results as presented in Table 2.
Construct reliability testing was done by looking at the value of composite reliability and Cronbach's alpha.Composite reliability values > 0.7 and Cronbach's alpha > 0.6 indicate that all research variables have reliable indicators/items.The reliability indicator was evaluated from the outer loading value of each measurement item.An outer loading value > 0.7 indicates a reliable indicator (Hair et al., 2017), while an outer loading between 0.4-0.7 can be used if the constructed variable formed has convergent validity (AVE) > 0.5 (Hair et al., 2014).The test results show that the outer loading value for each variable is > 0.5.Evaluation of convergent validity was seen from the value of average variance extracted (AVE).The AVE value > 0.50 indicates that the average construct explains more than half of the variance of the latent variable, so it can be concluded that the measurement items can explain the constructed variable well.The test results show that all variables have an AVE value > 0.50.
The discriminant validity was evaluated by comparing the outer loading of each item with the cross-loading of the item with other variables in the structural model.The test results show that the value of the outer loading variable was greater than the cross-loading of the item with other latent variables.For example, the outer loading of talent management practice was greater than cross-loading items with perceived distributive justice, perceived organizational support, turnover intention and so on.

Inner model analysis
Inner model analysis in this study was carried out by evaluating the corrected coefficient of determination (adjusted R2), predictive relevance (Q2), and the goodness of fit of the structural model.Adjusted R2 and Q2 values are presented in Table 3.Based on the R2 value, it shows that talent management practices can explain the variable perceived distributive justice by 62.5%, perceived organizational support by 77.7%, and turnover intention by 23.4%.The Q2 values are all positive which means that all endogenous variables can be well predicted by exogenous variables.The goodness of fit evaluation of the structural model is seen from the SRMR value with the reason being to avoid model specification errors in PLS-SEM (Henseler & Sarstedt, 2013).The test results show an SRMR value of 0.052 where the value is < 0.08 so it can be concluded that the structural model of this study is in good condition (fit).Multicollinearity problems do not occur in this study because all variables have a variance inflation factor (VIF) value of less than 10 (Hair et al., 2010).

Hypothesis testing
The results of the research hypothesis test are divided into two, namely direct testing and indirect testing.The results of the direct test are presented in Table 4. Based on the test results at a significance level of 95% as presented in Table 4, it shows that is rejected because the p-value is > 0.05 (0.670; 0.045).This means that talent management practices do not significantly affect turnover intention.H2 is accepted because the p-value <0.05 (0.000; 0.791), which means that talent management practices have a significant effect on increasing perceived distributive justice.H3 is accepted as shown by the p-value <0.05 (0.000; −0.289), which means that perceived distributive justice has a significant effect on reducing turnover intention.Furthermore, H5 is accepted, marked with a p-value <0.05 (0.000; 0.882), so it is concluded that talent management practices significantly increase perceived organizational support.Finally, H6 is accepted as shown by the p-value <0.05 (0.021; −0.271), which means that perceived organizational support has a significant effect on reducing turnover intention.
The results of testing the indirect effect through mediation are presented in Table 5 with the results of H4 being accepted because the p-value <0.05 (0.000; −0.228), which means that talent management practices affect reducing turnover intention through perceived distributive justice.The last hypothesis, namely H7, is also accepted because the p-value <0.05 (0.022; −0.239), so it is concluded that talent management practices have a significant effect on reducing turnover intention through perceived organizational support.

Discussion
In general, this study aims to analyze the influence of talent management practices on turnover intention either directly or through perceived distributive justice and perceived organizational support as mediation.The results of this study reveal that talent management practices have a significant negative effect on turnover intention through perceived distributive justice and perceived organizational support.Perceived organizational support has a stronger mediating effect than perceived distributive justice, although when viewed from the direct effect on the turnover intention, it indicates that perceived distributive justice has a stronger influence than perceived organizational support.
The results of this study provide novelty in the form of the perceived role of distributive justice and perceived organizational support as mediation.The results of this study answer the gap stated by Luna-Arocas & Morley (2015) that the implementation of talent management may not have a direct relationship with employee attitudes but instead, they produce effects through several mediation mechanisms.The results of this study also answer some of the limitations of empirical research, for example, related to the type and scale of the organization, the characteristics of the respondents, as well as the small number of samples.With this research, the literature will have references to empirical research conducted in the tax administration sector.This research provides relevant and very useful theoretical contributions and practical implications.

Theoretical implications
This research provides theoretical contributions related to talent management, perceived distributive justice, perceived organizational support, and turnover intention.The contribution made, namely the first, is consistent with social exchange theory (Blau, 1964) which states that the relationship between employees and the organization is reciprocal.In the context of exclusive talent management, it pays more attention to talented employees than non-talent, so employees who are not yet talented feel a little impact from talent management practices.The results of this study are in line with the research of Rana & Abbasi (2013), Kekgonegile (2014), and (Hoogervorst, 2009).
Second, support for the equity theory is shown from the results of research that proves the influence of talent management practices on turnover intention through perceived distributive justice.The significant influence of talent management practices on perceived distributive justice in this study is consistent with the research of Gelens et al. (2014), while the significant effect of perceived distributive justice on turnover intention is in line with the results of research by Khan et al. (2013) and Raza et al. (2017) which also found no effect of talent management on turnover intention.
There is no influence of talent management practices on turnover intention, presumably because the respondents are non-talent employees who have not been directly involved in the talent management program.The condition of respondents who have never participated in a talent management program directly is thought to influence respondents' perceptions.Perceptions of talent management practices will certainly be more in line with actual conditions when respondents are directly involved as talent employees.The absence of the influence of talent management practices on turnover intention is a logical thing because non-talent employees do not directly experience the existence of a talent management program.This finding is theoretically Third, support for organizational support theory is evidenced by the influence of talent management practices on perceived organizational support, and then with perceived organizational support, it encourages a decrease in turnover intention.The significant effect of talent management practices on perceived organizational support in this study supports previous research from Gelens et al. (2015) and Swailes et al. (2016) and the effect of perceived organizational support on turnover intention is in line with the results of research by Dawley et al. (2008) and Dawley et al. (2010).
Finally, this study contributes to the development of empirical research related to talent management practices on turnover intention in public sector organizations.The widespread implementation of talent management in the public sector needs to be followed by an evaluation of the effectiveness of its implementation.The results of this study provide answers to criticisms or issues that often arise regarding the implementation of talent management, namely the issue of fairness over the distribution of resources to employees (Schumann, 2001).Talent management practices that are implemented properly encourage all employees to feel that there is justice and that this perceived distributive justice reduces the intention to switch to another organization.The results of this study also add to the repertoire of empirical research which can be used as a reference in drawing generalizations or conclusions regarding the influence of talent management practices on turnover intention because as stated by De Boeck et al. (2018) that it is not permissible to make generalizations about the influence of talent management practices before knowing the reactions of non-talent employees.

Practical implications
The results of this study provide practical contributions related to the influence of talent management practices on public organizations as follows: first, the importance of creating distributive justice and providing support to all employees, both talented and non-talented.The implementation of talent management needs to be followed by other human resources management practices to encourage perceived distributive justice and perceived organizational support, for example by awarding employees according to contribution, level of effort, and level of employee performance.
Second, there is a need to evaluate the talent management approach used by the organization to see which approach is appropriate whether it is an exclusive or inclusive approach.Implementation of inclusive talent management can be an alternative to developing employees according to their respective talents.Finally, it is necessary to evaluate the implementation of talent management in the public sector periodically considering that talent management is currently widely applied to various public organizations.
The results of this research are relatively in line with the results of research conducted in private organizations.For example, related to the effect of talent management practices on turnover intention, which does not show a significant effect, it also occurs in the mining sector because talent management is implemented poorly (Kekgonegile, 2014).Even research by Ogbeibu et al. (2021) in the manufacturing sector shows that the influence of green hard and soft talent management drives turnover intention.The role of distributive justice which is felt to be very significant in encouraging the influence of talent management practices on turnover intention at the Directorate General of Taxes also occurs in research in the financial sector conducted by Gelens et al. (2014) and the telecommunications sector conducted by Raza et al. (2017).Likewise, the role of organizational support is felt to be significant in the context of the Directorate General of Taxes as happened in the banking sector (Islam et al., 2013;Mensah, 2019), the financial sector (Gelens et al., 2015), and the manufacturing sector (Dawley et al., 2010).It can be said that the Directorate General of Taxes has implemented modern human resource management practices since the introduction of bureaucratic reform, one of which was marked by the implementation of talent management.The Directorate General of Taxes adopts talent management practices that occur in private organizations so that this can encourage similarities in impact or influence.

Limitations and future research suggestion
This study still has some limitations that need to be followed up with other studies in the future.First, the research uses a quantitative approach so that discussions and conclusions are drawn from the results of quantitative data analysis.To find out more about the impact of talent management practices on turnover intention in public organizations, it is necessary to deepen the qualitative data from the results of in-depth interviews with respondents.
Second, this research was conducted at just one institution in Indonesia to support drawing conclusions or generalizations regarding the effect of talent management practices on public sector organizations, it is necessary to increase the scope of research on other public organizations both in Indonesia and other Asian countries because an exclusive talent management approach is necessary validated for Asian Countries (Bhatia & Baruah, 2020).Finally, future research is suggested to add the dependent variable as the impact of talent management practices or add a moderating variable on the effect of talent management practices on turnover intention.

Table 5 . Results of the indirect effect test
by the resource-based view which is a reference for exclusive talent management.Exclusive talent management practices encourage more positive reactions from talented employees than nontalent employees because organizations are more focused on developing talented employees.For this reason, talent management practices do not have a direct influence on turnover intention, but through the role of perceived distributive justice and perceived organizational support as mediation.