Talent management and performance in the public sector: the role of organisational and line managerial support for development

Abstract

While research shows that talent management (TM) practices are linked to individual and organisational outcomes, little is known about the mechanisms through which these processes unfold, especially in the public sector of emerging market economies. This exploratory study investigates TM in the public sector context of Bahrain addressing an important research gap. Drawing on Organisational Support Theory (OST) and strategic human resource management (SHRM) literature, we test the role of line managerial support for development (LMSD) and organisational support for development (OSD) as mechanisms through which TM practices influence individual talent performance. Results from a survey based on a talent pool of a Bahraini governmental administrative institution show that LMSD fully mediates the relationship between TM practices and each of the individual talent performance indicators, apart from task performance. We also find that the indirect relationship between TM and individual talent performance was significant at low levels of OSD. The study contributes to unlocking the TM black box in the public sector of developing countries by highlighting the key role of line managers in the implementation of TM practices revealing a substitution effect between LMSD and OSD, as interchangeable sources of developmental support that can enhance talent performance in this context.

Keywords:

• Public sector talent management
• Bahrain
• organisational support for development
• line managerial support for development
• talent performance

Introduction

After debating talent management (TM) mainly in the context of private-sector multinational companies (MNCs) (Collings et al., 2019), scholars have responded to calls to broaden studies of TM to a wider range of contexts (Gallardo-Gallardo et al., 2020; Kravariti et al., 2021). TM in the public sector is of growing interest (Tyskbo, 2019), and offers fresh insights into building human capacity in this context, as ‘governments and their societies have also joined the talent race’ (Khilji & Schuler, 2017, p. 400). Public sector TM refers to the application of key human resource (HR) practices that can facilitate staff, who possess both the required dexterities and the context’s values, in fulfilling the sector’s overall goal for the common good (Kravariti & Johnston, 2020). Reflecting the complex public sector context, public sector organisations have a wide range of stakeholders with divergent interests, rules, norms and logics and operate under different conditions from private organisations (Delbridge & Keenoy, 2010), which highlights the importance of understanding contextual factors in relation to TM in the public sector context (Thunnissen & Buttiens, 2017).

The growth of interest in public sector TM reflects the more intense competition for talent (Boselie & Thunnissen, 2017) and the focus on New Public Management (NPM) which promotes managerial and efficiency logics reflecting increasing pressures for accountability and the need to provide value for money services (Culie et al., 2014). The development of agile leaders who can manage effectively in volatile public sector environments (Barkhuizen, 2014) and the need to manage talents effectively to meet strategic objectives (Tummers & Knies, 2013) requires moving beyond traditional bureaucratic approaches to TM (McFarlane et al., 2012). Public sector TM implementation is more challenging due to tensions between traditional public sector values of fairness and equal opportunities and the new managerialist agendas of NPM (Boselie & Thunnissen, 2017). Also goals in the public sector are more diverse, less tangible and more difficult to measure making TM practices more difficult to implement (Blom et al., 2020). In addition, resource constraints limit flexibility in the implementation of TM practices in the public sector (Grant et al., 2020).

Yet, there remains a lack of empirical research on TM in the public sector (Clarke & Scurry, 2017) and little is known about how key actors affect individual and organisational outcomes (Gallardo-Gallardo et al., 2020). Scholars have called for further studies which examine the TM ‘black box’ and explore mechanisms that link TM practices to key outcomes (Mensah et al., 2016). Line managers, often assigned with the delivery of TM agendas in the public sector (Reilly, 2008), may constitute such a key mechanism (Thunnissen, 2016).

Drawing from the Strategic Human Resource Management (SHRM) literature and the concomitant ‘black box’ (Messersmith et al., 2011), as well as Organisational Support Theory (OST; Eisenberger et al., 2002), we explore the role of line managerial support for development (LMSD) as a mediating mechanism in the TM and individual talent performance relationship in the Bahrain public sector context. The SHRM literature has examined the devolution of HRM responsibilities to line managers and their role in effective implementation is linked to their support for talent development, promoting skill enhancement and career progression (Kehoe & Han, 2020). To what extent is the impact of TM practices on individual talent performance in the public sector, similar to private sector organisations, affected by LMSD? Hence, our first research objective is to assess the extent to which LMSD mediates the relationship between TM practices and individual talent performance in the Bahrain public sector.

SHRM research highlights that line managers are dependent on organisational resources and support to fulfil their HRM responsibilities (Purcell & Hutchinson, 2007) while employees are also affected by the support they receive from the organisation. Organisational support for development (OSD) is a major source of support comprising systematic efforts to develop talents for key positions in the organisation (Kumar et al., 2018). Organisations can provide formal support for development through the allocation of resources for learning and skills development including building a learning culture (Maurer & Lippstreu, 2008). Hence, our second research objective is to assess the extent that OSD moderates the relationship between TM practices and LMSD in the Bahrain public sector context.

Bahrain, as an emerging economy, is an interesting research context for two main reasons. First, context plays a significant role in understanding the enactment of TM practices and their outcomes (Gallardo-Gallardo et al., 2020) and significant differences exist between developed and developing countries (Vaiman et al., 2019). Due to lack of research on the latter (e.g. van den Brink et al., 2013) there is the need to contextualise TM in developing countries by examining a range of cultural and institutional factors. Second, public sector TM in emerging markets is often problematic and in the early stages of development, e.g. governments in the Middle East have introduced initiatives to enhance the development of national talent, but in practice the region’s collectivistic culture constrains the enactment of exclusive approaches to TM (Metcalfe et al., 2021). While Bahrain has taken some initiatives (e.g. establishment of several training institutions) to support talent development (Metcalfe et al., 2021), a key feature of this study is to investigate TM challenges faced in the Bahrain public sector in a context where the evolution of TM practices is in its early stages.

To summarise, this study contributes to the TM literature in three ways. First, we offer empirical evidence from a governmental administrative institution in Bahrain, which provides new insights on TM in an unexplored context. Second, we shed light on the TM ‘black box’ in the public sector (Thunnissen, 2016) by investigating the role of line managers in effective TM implementation and their impact on talent performance. We further postulate that OSD is a key moderator in this process, and we test a conceptual model that maps the key relationships described above, contributing to developing conceptual frameworks. Third, the study provides insights on individual talents’ perceptions in the public sector, which enhances scarce knowledge in this area (Farndale et al., 2014).

Theoretical background and conceptual model

TM has grown in importance as organisations seek to respond to the unprecedented uncertainty and complexity of today’s global environment (Collings et al., 2019). However, while the importance of management and leadership are increasingly recognised in the public sector, TM in this context is a relatively new development (Boselie & Thunnissen, 2017). Our approach is consistent with scholars (e.g. Lockwood, 2006) which sees talent as key people (Gallardo-Gallardo & Thunnissen, 2016), and we see TM in the public sector as primarily concerned with the attraction, training, development, and retention of key employees (Kravariti & Johnston, 2020).

A key debate in the area focusses on whether approaches to public sector TM should be inclusive (i.e. all employees as potential talents) or exclusive (i.e. focus on talents with valuable and unique skills (Collings & Mellahi, 2009). It has been argued that a more inclusive TM approach may fit with public sector values of equality and fairness (Thompson, 2017), but studies on public sector TM have highlighted that a hybrid approach, involving both inclusive and exclusive approaches is often employed in practice (e.g. Poocharoen & Lee, 2013). Exclusive approaches are used in the public sector for strategic recruitment of professional and leadership positions as illustrated by Tyskbo’s (2019) recent study of TM in a Swedish hospital, where an exclusive approach was utilised despite the highly egalitarian and collectivist context.

Public sector TM is important for several reasons. It can support NPM initiatives and increase staff commitment (Dougherty & van Gelder, 2015) and can help implement wider managerial reforms (Thunnissen & Buttiens, 2017). It also enhances the attraction, development and retention of talents and addresses talent shortages (Clarke & Scurry, 2017). TM research on the private sector has established the positive impact of TM on talent performance (e.g. Mensah, 2019) and has determined individual talent performance as a key outcome of TM implementation (Luna-Arocas & Morley, 2015). Research shows that TM conveys positive messages to talents who respond with positive performance behaviours (Dang et al., 2020). Empirical studies have explored some of the underlying mechanisms through which TM impacts on individual talent performance including person-job fit (Mensah & Bawole, 2020), job satisfaction (Wickramaaratchi & Perera, 2020) and affective commitment (Mensah et al., 2016), recognising the important role of line managers and organisational support (Bratton & Watson, 2018).

By contrast, empirical evidence linking TM in the public sector and individual talent performance is very limited. In Sri Lankan public banks, for example, investments in TM led to increases in talented Generation’s Y individual performance (Wickramaaratchi & Perera, 2020). However, evidence from TM in the emerging market economies of the Arab Middle East is limited to Qatar (see Masri & Suliman, 2019). Given the limited attention placed on the link between TM and individual performance in public sector contexts, as well as the mediating and moderating variables affecting this relationship, the present study addresses the research gap by investigating the role of LMSD and OSD as mediating and moderating forces (Figure 1).

Figure 1. Conceptual framework.

The mediating role of line managerial support for development

In linking public sector TM practices and individual performance, our primary focus is on line managers whose role in the implementation of HR practices has been transformed over the last decades (Purcell & Hutchinson, 2007). Research highlights how line managers enact HR practices in different organisational and institutional contexts (Kehoe & Han, 2020), and how they influence the implementation of intended policies (Wright & Nishii, 2013). Consistent with OST theory and literature on perceived supervisor support employees would respond favourably to intended policies to the extent that line managers effectively help them in this process, and their effectiveness in HRM implementation also shapes employee procedural justice perceptions (Sikora & Ferris, 2014). In summary, line managers’ implementation of HR practices influences a range of attitudinal and performance outcomes including employee affective commitment and job satisfaction (e.g. DeWettinck & Vroonen, 2017), turnover intentions (Sikora et al., 2015), individual and team performance (Pak & Kim, 2018), and organisational performance (Azmi & Mushtaq, 2015).

In the TM field, line managers are the ‘link pin’ between intended policies and practice (Gallardo-Gallardo et al., 2020). While senior public sector managers are pivotal in orchestrating the alignment of TM, organisational culture and strategic objectives, line managers are the ‘owners of talent’, and make key human capital decisions (Kock & Burke, 2008). Effective implementation of TM depends on how effective line managers are in delivering their HRM responsibilities (Reilly, 2008), yet they often lack the soft skills to motivate and support their employees with mentoring and coaching (Kock & Burke, 2008). In the Gulf Cooperation Council (GCC) region, for instance, line managers often support ‘sheikhocracy’ in that they expect and accept no challenge from their subordinates, which limits TM implementation (Finlay et al., 2005). Public sector line managers face challenges of aligning TM with the strategy of the organisation, together with the need to develop TM practices which fit with the internal culture (Harris & Foster, 2010). Tensions related to equal treatment may arise as introducing a version of exclusive TM for a minority of employees is not a good fit with the egalitarian organisational culture of most public sector organisations (Harris & Foster, 2010). Research suggests that line managers should be involved early in the development of TM practices, so that their perspective and concerns are taken on board (Thunnissen & Buttiens, 2017).

Public sector research indicates that TM practices are generally underdeveloped, poorly applied, ad hoc, fragmented and reactive (Culie et al., 2014) which emphasises the key role and challenges facing line managers in TM implementation. Thunnissen’s (2016) study on the mediating role of line managers in the TM and employee performance relationship in public sector contexts showed that the implementation of TM strategies in Dutch academic departments was influenced by academic line managers and that the power of HR departments was limited due to informal political processes (van den Brink et al., 2013). Some talents differentiated between promises at the organisational (HR) level and at the unit level, and suggested their development could be either supported or limited by line managerial practices (Clarke & Scurry, 2017).

In this article, we focus on LMSD as a mediating mechanism, as opposed to the general role of line managers, because line managers assume multiple responsibilities in terms of developing talent. These include identifying internal talents who have the potential to fill key positions (Hirsh, 2015). Line managers may have the additional responsibility of supporting the development needs of individuals and coaching has emerged as a key line manager role whose effectiveness was dependent on several individual-, dyadic- and organisational- level factors (Nyfoudi & Tasoulis, 2021). This lends support to the notion of line managers as ‘talent developers’ and LMSD is likely to affect the performance not only of those who are selected for the talent pool (Dries, 2013), but also those who are excluded. Finally, responding to the challenge of skills shortages in the public sector due to resource constraints including limitations on buying talent externally, line managers have an important role in ‘making’ talent internally, with a focus on developing professional skills through mentoring and coaching (Kock & Burke, 2008).

In conclusion, consistent with SHRM literature (Kehoe & Han, 2020) and OST theory (Eisenberger et al., 2002), we respond to calls to examine the public sector TM black box (Gallardo-Gallardo et al., 2020), and argue that employees will respond favourably to TM practices to the extent that line managers effectively support their development in this process. Research has shown that line managers play a key role in implementing HRM practices by influencing employee perceptions (Kuvaas & Dysvik, 2010). In addition, managerial support for development can influence a range of employee outcomes including work engagement, turnover intentions, and job satisfaction (Kumar et al., 2018). Sidani and Al Ariss (2014) underscored the role of line managers in the GCC region in particular, suggesting that the region’s local line managers often hold rigid and traditional attitudes that limited the effective enactment of TM practices. Hence, we suggest that employees who believe their line managers support their development, in the context of TM practices in the Bahrain public sector, are likely to reciprocate by engaging at higher levels of task, contextual, and adaptive performance, and lower levels of counter-productive performance.

Hypothesis 1a: LMSD mediates the relationship between TM practices and task performance.

Hypothesis 1b: LMSD mediates the relationship between TM practices and contextual performance.

Hypothesis 1c: LMSD mediates the relationship between TM practices and adaptive performance.

Hypothesis 1d: LMSD mediates the relationship between TM practices and counterproductive work behaviour.

The moderating effect of organisational support for development

Line managers and employees do not operate in a vacuum; a variety of internal and external factors affect TM implementation (Thunnissen & Buttiens, 2017), including the support offered by the organisation. Research in OST has established that various sources of support, most notably the organisation and supervisors, as well as co-workers, are drivers of employee attitudes and performance (Eisenberger et al., 2002). These sources of support are perceived as distinct and independent, as employees maintain different relationships with these constituents, and empirical evidence indicates that supervisory support is not simply one facet of organisational support (Shi & Gordon, 2020). For example, research shows that when supervisors provide support they influence employee turnover cognitions without the mediation of organisational support, and when supervisors provide high support, organisational support becomes a less important predictor (Maertz et al., 2007). Experimental studies have examined scenarios under which employees experience high levels of organisational support and low levels of managerial support, and vice versa, finding that such imperfect scenarios lead to lower levels of work engagement (Shi & Gordon, 2020).

However, organisational and supervisory support are not necessarily mutually exclusive, but likely to be complementary, and many studies have shown that they tend to be moderately to strongly correlated (e.g., Caesens et al., 2014). Hence, although OST theory has generally supported a multi-foci perspective in terms of these distinct sources of support with potentially complimentary or conflicting goals, little is known about the interaction effects and how different configurations of support impact on employees outcomes, with recent research showing that employees experience a generalised climate of support across sources but can also perceive dissimilar levels (Caesens et al., 2020).

Based on the discussion above, indicating the complex interactions between organisational and managerial support, in this study we focus on OSD as a moderator in the relationship between talent management, LMSD and talent performance. Conceptualising OSD as comprising of the systematic organisational efforts aimed at helping employees grow (Kumar et al., 2018), we posit that OSD is likely to be a catalyst that undermines or reinforces the role of line managers in the effective implementation of TM practices. While LMSD captures the responsibilities of line managers, such as talent attraction and development, OSD captures the organisational orientation including investing in opportunities for development, and support for a developmental culture. We perceive that OSD is a moderator in the public sector, because line managers may be unwilling or unable to perform people management responsibilities due to lack of time, lack of key competencies, or lack of socio-emotional and tangible resources provided by the organisation (Thunnissen, 2016).

In support of this proposition, studies reveal that organisations which acknowledge employee contributions and care about employee well-being experience more effective implementation of HR practices’ and enhanced employee performance (Eisenberger et al., 2002). SHRM studies indicate that when line managers were able to draw on several sources of support they rated themselves as more effective in implementing HR practices (Op De Beeck et al., 2018). Purcell and Hutchinson (2007) showed that lack of organisational support undermined the effectiveness of ward line managers in the NHS in the implementation of HR practices, and the effect of line managers’ people skills on employee satisfaction was stronger when line managers received support from HR (Bos-Nehles et al., 2013). Hence, we anticipate that the extent to which public organisations support talents to develop moderates the mediated effect of TM practices on individual talent performance in the Bahrain public sector context, and postulate the following:

Hypothesis 2a: The indirect positive effect of TM practices on talent task performance, via LMSD, is conditional on OSD, so that the indirect effect is stronger when OSD is high.

Hypothesis 2b: The indirect positive effect of TM practices on individual contextual performance, via LMSD, is conditional on OSD, so that the indirect effect is stronger when OSD is high.

Hypothesis 2c: The indirect positive effect of TM practices on adaptive performance, via LMSD, is conditional on OSD, so that the indirect effect is stronger when OSD is high.

Hypothesis 2d: The indirect negative effect of TM practices on counterproductive work behaviour, via LMSD, is conditional on OSD, so that the indirect effect is stronger when OSD is high.

Talent management in Bahrain

Bahrain is an emerging market economy and part of GCC. In this region, the majority of nationals work in the public sector whereas private sector organisations rely heavily on expatriates (The World Bank, 2019). In Bahrain, the government occupies a central role in the country’s economy and is the main employer of nationals (Sidani & Al Ariss, 2014). A key challenge is local talent scarcity (De Bel-Air, 2015) and a localisation strategy has been developed to address this strategic talent issue with a focus on developing national talent (De Bel-Air, 2015).

Talent development plays a key role in Bahrain in developing the human capital and leaders needed for national competitiveness (Al Mannai et al., 2017). The Bahraini Society for Training and Development in collaboration with the Supreme Council for Women and the Ministry of Education are the responsible authorities for TM planning (Metcalfe et al., 2021). The government’s 2030 vision launched in 2008 encourages Bahrainis to reach their full potential via participation in various talent development programmes. For example, development programmes for civil servants and high potentials in the government sector (BIPA, 2020) and in the municipalities have been developed (Bahrain Ministry of Tourism, 2020). Compared to other GCC countries, Bahrain seems to perform well in terms of inclusively encouraging talented nationals to upgrade their skills (Sidani & Al Ariss, 2014), even though different developmental programmes indicate that an exclusive approach to TM might also be applicable.

Bahrain’s support for the development of diversified talents is, however limited by several factors. For example, civil disturbances –driven by Arabs’ reluctance to adhere to westernised and perceived individualistic practices– along with the traditional central role of the region’s governments in developing human capital decelerated the development of a concrete TM framework for national talent (Metcalfe et al., 2021). Major challenges remain for the country to develop and sustain a talented workforce (Biygautane & Al Yahya, 2014), including the limited capacity of public sector line managers to enact TM practices. In theory, public sector line managers are expected to deal with a range of administrative issues (BIPA, 2019), but more recently they have assumed greater responsibilities (Abou-Gamila et al., 2015) including setting targets and taking initiatives (Wauters, 2017). The blurred nature of line managers’ authorities and responsibilities has contributed to the inefficient enactment of TM practices such as infrequent performance reviews and ambiguity over the link of performance appraisals to rewards (Al-Alawi et al., 2019). In addition, since 2014 and the drop of oil prices and increased political uncertainty, Bahrain’s civil service has been challenged by significant budget deficits (Joyce & Al Rasheed, 2017) which constrained the implementation of TM practices and the enactment of the Vision 2030 (Abou-Gamila et al., 2015). An important feature of this study is that it goes beyond formal TM policies to provide empirical evidence on the role of line managers in the enactment of TM practices within the Bahrain public sector context.

Method

Sample

The study was conducted in a Bahraini governmental administrative institution, which employs 179 individuals, 120 of whom participated in the study representing the organisation’s talent pool due to their holding core positions. Fifty-nine employees directly employed by the government have peripheral responsibilities (e.g. maintenance or caretaking duties) and were not part of the talent pool. All 120 participants were Bahrainis and came from a range of responsibility levels with the majority (66.9%) being employees with no managerial responsibilities followed by 17% professionals (e.g. accountants), 10.2% line managers and 5.1% senior managers.

Research suggests that a decade ago the Bahraini workforce was male dominated (Unicef, 2011). However, both genders were equally represented in our sample (i.e. male = 50.9% and female 49.1%) suggesting some advances in gender equality reflecting the Vision 2030, which aims to enhance women’s engagement at all levels of the public sector (UNDP, 2021). The majority of the respondents (47.5%) were between 25 and 34 years old and well educated with 52.5% holding a Bachelor’s degree and an additional 26.2% possessing a postgraduate degree.

Study participants were requested to voluntarily complete an anonymous online questionnaire whose return would be taken as evidence of informed consent. The study was approved 10^th July 2019 under the ethical review procedures of the Faculty of Business and Law Ethics Committee of the University of Portsmouth. The organisation’s HR department helped in distributing the questionnaire’s link to employees that constitute the organisation’s talent pool, in line with previous research (e.g. Mamman & Al Kulaiby, 2014). We received 118 usable responses resulting in approximately 98% response rate.

Measurement

To investigate the role of LMSD and OSD as vehicles through which TM practices impact on individual talent performance, we adopted an exploratory approach. All items were measured on a five-point Likert scale (1 = strongly disagree, 5 = strongly agree).

Talent management practices

To move TM research forward, we need to discuss the measurement of various TM constructs (intended, actual, perceived) and at various levels of analysis (organisational, team, individual-level; Meyers et al., 2020). While a multi-level measurement of TM practices in the particular context is beyond the scope of this study, we explicitly focus on perceived TM practices for three reasons. First, if TM practices are to influence individual performance, they must first exist as cognitions in employees’ minds (Wright & Nishii, 2013). Second, even if there may be official and formal TM practices, employees will infuse them with their own meaning depending on their beliefs, experiences, and perceived organisational intentions (Meyers et al., 2020). Third, employees may develop idiosyncratic views regarding TM practices which often differ within or across employee groups (Liao et al., 2009).

The specific measurement of perceived TM practices was developed based on prior work in this field (El Dahshan et al., 2018). It includes nine items concerning employee perceptions of TM practices related to talent recruitment (e.g. ‘The organisation is attracting and recruiting key people’), talent development (e.g. ‘The organisation is growing future managers/leaders’), and talent retention (e.g. ‘The organisation is retaining key employees’). This scale employs evaluative measures of perceived TM practices, i.e. it is focused on employees’ assessment of TM constructs. This perspective is drawn from HRM research, which has identified three approaches to measuring employee perceptions on HR practices, namely measures investigating content (what), strength (how), and attribution (why; Wang et al., 2020). Studies focusing on ‘what’ measure the content of practices as subjectively experienced and are further classified into descriptive, evaluative, observation-based, and experience-based (ibid.). Although this classification system has not been systematically applied in the measurement of employee perceptions of TM practices, we suggest that the scale used in our study is focused on the evaluation of the content of TM practices. Additionally, following studies conducted in the GCC region viewing talents as individuals that hold key positions (see Sidani & Al Ariss, 2014), and the view of talent as key employees (Gallardo-Gallardo & Thunnissen, 2016), we worded the TM scale’s items accordingly. The scale’s reliability coefficient was .72.

Line managerial support for development

LMSD was measured with six items adapted from Maurer and Lippstreu (2008). The items included questions capturing employee perceptions on the received support from their line managers related to their development. A sample item is ‘My line manager encourages me to participate in activities which promote skill improvement’. The reliability coefficient was .85.

Organisational support for development

OSD was measured with six items adapted from Kraimer et al. (2011). The items captured employee perceptions on the extent to which their organisation supports their capabilities’ advancement through various developmental opportunities (Kumar et al., 2018). A sample item is ‘My organisation provides opportunities for employees to develop their specialised functional skills’. The scale’s reliability score was .87.

Individual talent performance

We relied on self-reported perceptions to assess individual talent performance related to task performance, contextual performance, adaptive performance, and counterproductive work behaviour. We adopted the 23-item scale of Koopmans et al. (2014) that consists of individual performance indicators for each of the four dimensions addressing individual talent performance. A sample item for task performance is ‘I am being result-oriented’, for contextual performance is ‘I am cooperating with others’, for adaptive performance is ‘I am adjusting work goals when necessary’, and for counterproductive work behaviour is ‘I am purposely making mistakes’. The composite alpha reliability here was .88.

Control variables

In accordance with published HRM scholarship, we controlled for the following variables because they often influence talent perceptions: gender (Björkman et al., 2013), age (Newburry et al., 2014), educational level (Sonnenberg et al., 2014), responsibility level (Tymon et al., 2010), and length of service (Rathi & Lee, 2015).

Results

First, we ran exploratory factor analysis (EFA) with Direct Oblimin on all scale items to establish their retention. Given that items were translated from English into Arabic, and the independent variable was developed, EFA was preferred (Kuvaas, 2006). Items loaded either low or in more than one factors were deleted from the final measurement model.

Moreover, we ran Pearson correlations coefficients for the research variables (see Table 1). Preliminary statistical measurements renounced the probability of multicollinearity (Tolerance > 0.10 and VIF < 10.00) and allowed the inclusion of limited extreme cases (M of Mahalanobis Distance = 1 and of Cook’s Distance < 1). TM practices significantly and positively correlate with LMSD (r = .51, p < .01). LMSD significantly and positively correlates with task performance (r = .23, p < .05), contextual performance (r = .30, p < .05) and adaptive performance (r = .33, p < .01) while it is not significantly correlated with counterproductive work behaviour (r = −.15, p > .05). In line with Maertz et al. (2007) and Caesens et al. (2020), we found a statistical and positive correlation between LMSD and OSD (r = .71, p < .01).

Table 1. Means, standard deviations and correlations among study variables.

Variables	Mean	SD	1	2	3	4	5	6	7	8	9	10	11
(1) TM practices	25.60	6.26
(2) OSD	22.43	6.49	.61**
(3) LMSD	16.80	5.02	.51**	.71**
(4) Task performance	14.32	4.96	.20*	.23*	.23*
(5) Contextual performance	20.19	7.45	.15	.23*	.30**	.63**
(6) Adaptive performance	16.54	6.58	.29**	.32**	.31**	.68**	.80**
(7) Counterproductive work behaviour	17.69	2.76	.30	-.08	-.15	−19*	-.31**	-.28**
(8) Gender	1.49	.50	-.07	-.01	.08	-.01	-.06	-.03	-.04
(9) Age	2.51	1.15	-.06	-.07	-.05	.00	.07	-.00	.16	-.40**
(10) Educational level	2.97	.97	.13	.06	.07	.05	-.02	.10	.05	-.02	.00
(11) Responsibility level	4.44	.93	-.00	.06	.09	-.01	.07	-.3	-.04	-.01	.08	.19*

*p < .05; **p < .01; N = 118.

We used the PROCESS Version 3 macro developed by Hayes (2018) to statistically test our hypothesised relationships. PROCESS was preferred over a latent variable approach (e.g. structural equation modelling) because it uses ordinary least squares (OLS) and without a path diagram measures indirect effects in both unmoderated and moderated mediated models with one or more moderators and mediators (Hayes, 2018). PROCESS also allows bootstrapping and to concurrently investigate mediation and moderation effects (ibid). A latent variable approach to testing moderated mediation models was not appropriate, because of the sample size and the conditions related to the degree of freedom for the independent variable (Einarsen et al., 2018).

As illustrated on Table 2, the findings of PROCESS mediation analyses (Model 4) with 10,000 bootstrap samples and 95% confidence interval (Einarsen et al., 2018) reveal that the confidence intervals for the indirect effect of TM practices on contextual performance (bootstrapping estimate = .17; 95% CI [.04, .32]) and on adaptive performance (bootstrapping estimate = .12; 95% CI [.01, .25]) through LMSD is positive and significant, while on counterproductive work behaviour is negative and significant (bootstrapping estimate = −.05; 95% CI [-.10, −.01]). The direct effect estimates of TM practices on contextual performance (effect = .03; p = .83), on adaptive performance (effect = .19; p = .08), and on counterproductive work behaviour (effect = .06; p = .21) establish that LMSD fully mediates each of the aforementioned indicators. The confidence intervals for the indirect effect of TM practices on task performance is positive but includes zero (i.e. bootstrapping estimate = .07; 95% CI [-.03, .16]). Therefore, Hypothesis 1a is rejected, whereas Hypotheses 1b, 1c and 1d are confirmed.

Table 2. Results of the moderated-mediation analysis.

Predictor	Coeff.	SE	t	p	Boot LLCI	Boot ULCI
	Mediator variable model: LMSD
Constant	−8.01	4.04	−1.98	.05	−16.02	.00
TM practices	.59	.18	3.24	.00	.23	.96
OSD	1.01	.19	5.35	.00	.64	1.38
TM practices X OSD	-.02	.01	−2.85	.01	-.04	-.01
	Dependent variable model: task performance
Constant	9.02	2.04	4.43	.00	4.99	13.05
LMSD	.16	.10	1.53	.13	-.05	.37
TM practices	.10	.08	1.20	.23	-.07	.27
	Indirect effect of X on Y
	Effect	Boot SE
LMSD	.07	.05			-.03	.16
	Dependent variable model: contextual performance
Constant	12.38	3.00	4.12	.00	6.43	18.33
LMSD	.42	.15	2.74	.01	.12	.73
TM practices	.03	.12	.21	.83	-.22	.27
	Indirect effect of X on Y
	Effect	Boot SE
LMSD	.17	.07			.04	.32
	Dependent variable model: adaptive performance
Constant	6.76	2.61	2.60	.01	1.60	11.93
LMSD	.29	.13	2.19	.03	.03	.56
TM practices	.19	.11	1.75	.08	-.03	.40
	Indirect effect of X on Y
	Effect	Boot SE
LMSD	.12	.06			.01	.25
	Dependent variable model: counterproductive work behaviour
Constant	18.12	1.15	15.81	.00	15.85	20.39
LMSD	-.12	.06	−2.01	.05	-.23	-.00
TM practices	.06	.05	1.27	.21	-.03	.15
	Indirect effect of X on Y
	Effect	Boot SE
LMSD	-.05	.02			-.10	-.01

Notes: CI = confidence interval; LL = lower limit; UL = upper limit; 10,000 bootstrapping samples were used; SE = standard error. N = 118.

As illustrated on Table 3, the findings of PROCESS moderated mediation analyses (Model 7) with 10,000 bootstrap samples and 95% confidence interval (Kawai & Mohr, 2020) demonstrate that the index of moderated mediation for counterproductive work behaviour is positive and significant (index = .00; 95% CI [.00, .01]), yielding support for Hypothesis 2d. The indices of moderated mediation for contextual and adaptive performance are negative but significant (index = −.01; 95% CI [-.02, −.00]; index = −.01; 95% CI [-.02, −.00]), thus providing support for Hypotheses 2b and 2c. The index of moderated mediation for task performance is negative and non-significant (index = −.00; 95% CI [-.01, .00]), rejecting Hypothesis 2a. The indirect effect from TM practices to contextual performance, adaptive performance and counterproductive work behaviours through LMSD is significant only at low levels of OSD (bootstrapping estimate = .10; 95% CI [.02, .22]; bootstrapping estimate = .07; 95% CI [.00, .18]; bootstrapping estimate = −.03; 95% CI [-.07, −.00]). These results indicate that TM practices are less effective in supporting individual talent performance via LMSD in contexts of medium and high OSD. Hence, Hypotheses 2b, 2c and 2d are partially confirmed as the moderation effect of OSD is supported but the conditional relationships do not comply with the expected level of the moderator. Figure 2 illustrates the moderating effect of OSD via simple slopes.

Figure 2. Moderating effect of OSD on the TM practices- LMSD relationship.

Table 3. Conditional indirect effect & index of moderated mediation (PROCESS macro, Model 7).

			Bootstrapped CI [95%]
Conditional effects (via LMSD)	Coeff.	SE	LL	UL
	Task performance
OSD (-1 SD)	.04	.03	-.02	.10
OSD (M)	.01	.01	-.01	.04
OSD (+1 SD)	-.01	.01	-.04	.02
	Index	SE	LL	UL
Index of moderated mediation	-.00	.00	-.01	.00
	Contextual performance
OSD (-1 SD)	.10	.05	.02	.22
OSD (M)	.03	.03	-.02	.10
OSD (+1 SD)	-.02	.03	-.09	.05
	Index	SE	LL	UL
Index of moderated mediation	-.01	.00	-.02	-.00
	Adaptive performance
OSD (-1 SD)	.07	.05	.00	.18
OSD (M)	.02	.02	-.01	.08
OSD (+1 SD)	-.01	.02	-.06	.04
	Index	SE	LL	UL
Index of moderated mediation	-.01	.00	-.02	-.00
	Counterproductive work behaviour
OSD (-1 SD)	-.03	.02	-.07	-.00
OSD (M)	-.01	.01	-.03	.01
OSD (+1 SD)	.00	.01	-.02	.03
	Index	SE	LL	UL
Index of moderated mediation	.00	.00	.00	.01

Notes: CI = confidence interval; LL = lower limit; UL = upper limit; 10,000 bootstrapping samples were used; SE = standard error. N = 118.

Discussion and conclusion

This study addresses an important research gap by advancing our limited knowledge of TM in the public sector and responds to calls to pay greater attention to contextual factors influencing TM (Gallardo-Gallardo et al., 2020). The study designed and tested a conceptual model which links TM practices with individual talent performance, and which explores underlying mechanisms (i.e. LMSD) and boundary stipulations (i.e. OSD). Our contribution is threefold. First, the present study enhances the literature on public sector TM in an emerging market context with rare empirical evidence from a Bahraini governmental administrative institution that seeks to support local talent development through the enactment of TM initiatives. Although inclusive TM may in theory better fit the public sector context (Thompson, 2017) and the collectivistic culture of the context under investigation (Metcalfe et al., 2021). Bahrain has initiated talent development schemes for public sector employees inclusively but variations exist based on talents’ key positions (e.g. Tamkeen). In line with public sector TM studies (e.g. Poocharoen & Lee, 2013; Tyskbo, 2019), we suggest that a hybrid TM approach is adopted by the Bahrain public sector reflecting the government’s approach to address talent scarcity.

Second, we provide evidence on the key role of line managers as an enabler of TM practices and a driver of talent performance, which enhances the limited empirical research on this area (Thunnissen, 2016). Although TM scholars have theorised the role of line managers in TM processes (Kock & Burke, 2008), few studies have tested how key actors are engaged in translating TM practices into actual individual outcomes, particularly in the public sector (Gallardo-Gallardo et al., 2020). This study highlighted that in the Bahrain public sector, TM practices positively affect individual talent performance, with regard to contextual performance, adaptive performance and counterproductive work behaviour via LMSD. In line with prior SHRM studies (e.g., Wright & Nishii, 2013), our findings indicate that the supportive-developmental role of line managers is a key mechanism through which TM practices affect individual talent performance in the public sector. As predicted by OST (Eisenberger et al., 2002), employees favourably respond to intended practices based on the developmental support they receive from their line manager, reflected in improved individual performance (Pak & Kim, 2018; Sikora et al., 2015). This study established the fully mediating role of LMSD in the effective implementation of TM practices and supports the argument that line managers bring TM policies to life (Gallardo-Gallardo et al., 2020). Low correlations among the study variables were somewhat expected as additional internal and external to the public sector factors affect TM implementation and outcomes including talent performance (see Kravariti & Johnston, 2020 for a review). For instance, Garrow and Hirsh (2008) suggest that TM may not bring about desired employee outcomes if it does not conform with an organisation’s culture or does not fit within current employee management practices.

Third, OST predicts that when line managers are supported by their organisation, they can have a much more positive impact on employee performance (Bos-Nehles et al., 2013), yet this has not been tested in public sector contexts. Somewhat consistent with our predictions, our findings suggest that when OSD is low, then the indirect impact of TM on talent performance (with the exception of task performance) is mediated by LMSD. In other words, low levels of OSD accentuate the role of line managers in TM processes. This is in line with evidence from the public sector contexts indicating the enhanced role of line managers who assume some of the middle managers’ responsibilities (Abou-Gamila et al., 2015). By contrast, when OSD is medium or high, the mediating role of line managers is insignificant. This finding suggests that when organisations fully support talents with well-established policies, fair decisions, and investments in TM, with concomitant support from senior managers and HR departments, then line management’s contribution to enact practices becomes less significant. Taken together, findings arguably reveal a substitution effect between line managers and the organisation as supportive resources, suggesting that employees might seek support from either the former or the latter during the TM process, which is in line with previous research showing that when supervisory support is high, organisational support is less important, and vice versa (Maertz et al., 2007).

Implications for practice

This study highlighted the role of line managers as key actors in the process of enacting TM practices in a public sector context and suggested that talent performance would be enhanced by investing in capable line managers. The findings of the study point towards a substitution effect between OSD and LMSD suggesting that public sector organisations may alternatively invest in high levels of OSD, alleviating line managers’ talent developmental responsibilities and enabling them to focus more on other duties. This may be achieved through an emphasis on the role, capabilities and engagement of HR professionals in TM design and implementation in the public sector, and support from senior managers. Clearly, talent performance is not a mere outcome of TM design, but requires configurations of support from line managers, HR and senior managers to secure effective implementation.

Limitations and future research

Responding to calls for greater context pluralism in TM research (Gallardo-Gallardo et al., 2020), we provide empirical evidence from an underexplored public sector context. While our small research sample limits the generalisability of our findings, we suggest that insights from this exploratory study may be applicable to GCC countries which share with Bahrain cultural, societal, and economic similarities (Sidani & Al Ariss, 2014). Future research should explore public sector organisations from different contexts and collect data from multiple sources as well as satisfy sample size requirements for testing moderated mediation models by using factor-based SEM techniques (Sarstedt et al., 2020). Multi-perspective approaches would provide insights into the views of a larger number of stakeholders. Due to the complexities of managing talent in an increasingly volatile public sector external environment (Boselie & Thunnissen, 2017), public sector TM research should adopt dynamic perspectives to chart how the TM agenda changes in relation to shifts in the strategies of public sector organisations (Clarke & Scurry, 2017). Our study complements the work on the well-researched BRIC emerging economies and highlights that emerging markets vary in terms of stages of development, regulatory environments and national cultures (Vaiman et al., 2019). Our research suggests that TM is more complex in the public sector and that future research could pay more attention to the role of competing logics in TM implementation (Tyskbo, 2019).

Methodologically, the scales adapted from well-established instruments provide us with an initial platform of how to measure TM practices and individual talent performance within the public sector context, yet need improvement. Specifically, scholars have argued that we need to examine various TM constructs (e.g. intended, actual, perceived) and at various levels of analysis (Meyers et al., 2020). There is also a level issue in our study in that we measure organisational-level antecedents but individual-level outcomes, and common method bias include self-assessed performance outcomes. Future studies should address those methodological issues by further developing those scales and their research designs. Qualitative studies could address the limitations of surveys by providing in-depth insights into a diverse range of issues, such as the tensions faced by line managers in implementing public sector TM in a rapidly evolving context. Developing our knowledge of the links between TM and performance could enhance understanding of the main constraints to the effective implementation of TM in this context, and shed light on issues related to the integration of TM with other HRM areas. A fruitful area for future research is exploring the links between diversity management and TM in the public sector. For example, research could examine changes in female labour force participation and experiences in the public sector context. Also, future research could examine more generally the barriers to developing effective TM in the Bahraini and GCC contexts.

Additionally, our analyses did not provide support for Hypotheses 1a and 2a. This suggests that task performance’s measurement might be problematic because it usually includes job-related elements such as volume of workload (Koopmans et al., 2011). However, it is challenging to identify attitudinal items that can reflect task performance as with the other three indicators of employee performance (Koopmans et al., 2014). Future studies could adopt alternative scales for evaluating individual talent performance which overcome this challenge.

Finally, this study focused on examining LMSD and OSD as the mediator and moderator respectively. We are in line with general literature (e.g., Caesens et al., 2020) that there is complexity between the two concepts and thus future research should look at different configurations of employee support. Drawing on research which suggested that individual talent performance is an outcome of TM practices (e.g. Luna-Arocas & Morley, 2015) and that both organisational and line managers’ support play a key role in this relationship (e.g. Bratton & Watson, 2018), this study provides a roadmap which suggests that TM practices impact on individual talent performance through LMSD and contingent on OSD. Future research can explore other potential mediating and/or moderating factors when examining TM’s effect on individual talent performance such as demographic data, the nature of talent’s work, organisational culture, and additional contextual factors, as well as identify optimal configurations of support in the TM black box.

Data availability statement

The data that support the findings of this study are available from the corresponding author upon reasonable request.

Disclosure statement

No potential conflict of interest was reported by the authors.