Coaction Interrupted: Logic Contestations in the Implementation of Inter-Organisational Collaboration around Talent Management in the Public Sector in Scotland

Our study explores how Human Resources (HR) actors from 24 public sector organisations in Scotland interpret and address multiple and competing institutional logics in the context of the implementation of an inter-organisational collaboration within the Public Sector to develop and implement talent management (TM). Our findings reveal that HR actors encountered day-to-day difficulties in blending different versions of TM, the need to have professional autonomy versus the requirement to develop some shared TM practices, and the fit of TM practices with public sector organisations. We therefore found that institutional logics impacted in different ways on efforts to collaborate and the types of TM practices developed and implemented. Theoretically, our study findings reveal that HR actors have agency at the practice level and that powerful actors can advocate particular logics to protect their professional and organisational interests.


Introduction
Talent management (TM) has moved beyond its traditional multinational and industrial settings and made its way into the public sector (van den Brink et al., 2013).Public sector organisations increasingly make use of TM practices including exclusive talent approaches, high potential programmes and talent development (MacFarlane et al., 2012).A particular example of the public sector is local government (Prosser et al., 2019), comprised of organisations involved in the delivery of health, education, policing and local authority services (Perlman, 2016).
These organisations emphasise professional expertise, responsiveness to the needs of the public and demonstration of value to the taxpayer (McCracken et al., 2012).
Public sector TM is unique in a number of respects.First, public sector organisations are characterised by goal ambiguity whereas strategic TM in the private sector (Collings et al., 2017) emphasises goal clarity and strategic alignment.Goals in public sector organisations are less tangible, more diverse and harder to measure (Blom et al., 2018), thus making TM practices more difficult to design and implement.Second, the public sector context is characterised by institutional logics that adhere to the values of bureaucracy, expertise, autonomy and egalitarianism (Alvhus, 2018).Therefore, TM tends to be perceived differently by employees who operate within these logics (Asplund, 2019).Third, general resource and Human Resources (HR) expertise constraints are prevalent in public sector organisations, thus impacting the types of practices selected and their implementation.There tends to be less flexibility and freedom around the implementation of TM practices (Fernandez and Moldogaziev, 2010).
It is against this context that public sector organisations look at different ways to utilise and gain benefits from HR and TM practices.One such approach is through a process of interorganisational collaboration driven by the need to achieve cost-reductions, drive efficiency and enhance service delivery (Sigismund Huff, 2005).Selden and Wooters (2011) proposed the concept of 'sharing' where HR services are consolidated into inter-organisational units and involve a horizontal partnership with other government entities within a specific geographical area.However, while these initiatives are on the surface attractive, they are in reality complex to implement (Cunha et al., 2017).Public sector HR actors often struggle to address competing priorities, cope with change, and respond to the requirements of powerful institutional actors (Pollitt and Bouckaert, 2011).
Competing institutional logics therefore play an important role in how HR practices are understood and enacted in organisations day to day (Alvehus, 2018).In particular, TM practices such as those aimed at attracting, developing, deploying and retaining talent are often contentious and complex within the public sector (Boselie and Thunnissen, 2017), suggesting that competing logics are likely to play a significant role in facilitating or inhibiting the extent to which they are successful (Tyskbo, 2019).As new initiatives are introduced, well established ways of working and the institutional logics that underpin them may become contested and challenged, resulting in resistance to new ideas, conflict and outcomes that were not anticipated.The presence of competing logics results in tensions and contradictions and creates dissonance for actors who have to manage these changes on the ground (Cunha et al., 2017).McDermott et al. (2013), for example, illustrated that the implementation of public policy can differ greatly from that which is mandated thus resulting in major gaps between the intended policy, the change initiative and its actual implementation.TM is a relatively new phenomenon in public sector organisations and comprises a complex and transformational set of HR practices to implement (Schuler et al., 2011).Questions arise concerning the fit of a market logic-driven TM in the public sector (Boselie and Thunnissen, 2017;Thunnisssen and Buttiens, 2017).For example, Pemer and Skjølsvik (2018) highlighted that conflicting institutional logics created incompatibilities regarding preferred approaches to getting things done, resulting in increased institutional complexity.Hargrave and Van de Ven (2006) found that implementation is more complex when old institutional logics are not fully replaced by the new strategic initiative, but continue to impact the way actors operate.However, Cunha et al. (2017) found that organisational actors make sense of competing logics and live through them.
Compounding these challenges are issues related to inter-organisational collaboration initiatives (Berger et al., 2005).The notion that public sector organisations should collaborate to achieve TM goals challenges existing ways of doing things, and institutional logics that emphasise autonomy, independence and territory (Norris-Tirrell and Clay 2010; Vangen et al., 2017) present another layer of complexity in addition to the issues posed by TM practices.
Horizontal inter-organisational collaboration, which is the focus of this study (Sorensen, 2018), can increase efficiency and result in new practices.However, there may be a lack of goal consensus (Sarapuu et al., 2014), low trust (Longva and Osland, 2007) and responsibility avoidance (Pell, 2016), with little achieved due to power and political issues.
Our paper seeks to contribute to these debates by exploring the influence of multiple competing logics of public sector HR practitioners in how they implement TM practices, how they collaborate with other public sector organisations in that endeavour and how they respond to private sector notions of TM.Scholars are interested in how different logics influence implementation actors, or what we describe as actors involved in day-to-day implementation (Nath, 2019;Schedler and Ruegg-Sturm, 2014).Institutional logics refer to organising principles that provide guidelines to actors as to how to behave and shape cognition and action (Meyer and Hammerschmid, 2006;Saz-Carranza and Longo, 2012).Institutional logics have framing capacities that are used by implementation actors to answer 'why' and 'how' questions and to label their actions.These actors are embedded within institutional contexts which guide their rationalities and identities.
The study context involved multiple, co-existing and potentially conflicting logics, including the 'state logic' focusing on equity, legality and correctness, and the 'market logic' focusing on profit, performance, competition, effectiveness, efficiency (Meyer and Hammerschmid, 2006).Professional logics emphasise egalitarianism, expertise, autonomy and integrity, and may elicit scepticism towards TM (Liebermanand and Friedrich, 2010).
Conflicting logics refer to situations which require implementation actors to reconcile different logics to achieve implementation.We respond to calls by Bévort and Suddabay (2016) and focus on frontline actors who were centrally involved in making the collaboration work on a day-to-day basis.This frontline context represents a gap in the literature and is important because, as Alvehus (2018) points out, actors on a day-to-day basis will seek to balance and reconcile institutional logics.Bertels and Lawrence (2016) also highlighted that individual actors will construct 'the meaning of those logics in ways that reflect and facilitate the aims and resources of individuals and their organisations' (p.368).
We adopt a temporal lens to explore the salience of particular institutional logics at each stage of implementation (Schultz and Hernes, 2013) and how conflicts are resolved.Interorganisational collaborations typically proceed through different stages (Pettigrew et al., 2019) involving voluntary or mandated formation (McDermott et al., 2013), with the engagement of implementation actors varying during the process of implementation.They involve stages of formation, initial scoping, development and implementation, scaling up and sustainability (Huxham and Vangen, 2000;Kumar et al., 2007).A temporal perspective views actors as operating in an ongoing present (Schultz and Hernes, 2013), suggesting that the operation of institutional logics will continually emerge and re-emerge and be shaped by past and future goals and priorities.Therefore 'the contingencies of the moment' (Emirbayer and Mische, 1998, p.962) influence institutional logics when implementing strategic change.
The remainder of the article is structured as follows: The theoretical background is presented, followed by a description of the research design.Our results are then presented, followed by a discussion of the theory and practice implications of the research, its contributions to the literature, the study limitations and future research avenues.

Theoretical background
The public sector and institutional logics Scholars conceptualise institutional change as taking place as a result of shifts in institutional logics (Ocasio et al., 2017).Institutional logics are normative concepts that regulate individual and organisational behaviour (Besharov and Smith, 2014;Smets et al., 2015).They are socially shared, and deeply held assumptions and values that provide organisational actors with a set of criteria for reasoning and legitimacy (Thornton and Ocasio, 2008).Kroezen and Heugens (2019) emphasise the collective nature of institutional logics, their taken-for-granted nature and influence on embedded actors.They are ideational in that they provide day-to-day actors with both cognitive and normative tools to make sense of the environment and justify action (Ocasio et al., 2015).
Public sector organisations are an appropriate setting in which to study multiple institutional logics because they are characterised by institutional complexity and sites of contestation between multiple logics (Currie and Spyridonidis, 2016).They are highly fragmented organisations (Pache and Santos ,2010) resulting in greater institutional complexity.Many different logics operate in public sector organisations (Ferry et al., 2019).
Professional logics emphasise expertise and solving the problems of clients, with van den Broek et al. (2014) noting emphasis on technical quality, prestige and autonomy and development of formal rules.Second, market-oriented logics vested in the emergence of new public management emphasise external performance, differentiation, benchmarks, efficiency, proficiency and easily measurable outcomes.Third, a legalistic-bureaucratic logic (Meyer and Hammerschmid, 2006) and is concerned with procedures, directives, governance and serving the public interest.

A micro perspective on institutional logics and the introduction of TM
An important focus of our study concerns the way in which HR practitioners on a day-to-day basis respond to different institutional logics in the context of implementing TM in the context of inter-organisational collaboration.Day-to-day or on the ground actors continuously combine, configure and manipulate institutional logics to suit their purposes (Jarzabkowski et al., 2013;McPherson and Sauder, 2013) and change occurs through their deliberate actions (Dalpiaz et al., 2016) and institutional work (Lawrence et al., 2011).These actions and institutional work may lead to reinterpretations of established logics (Smets et al., 2012) or reinforce the status quo.TM may be perceived as a business practice supported by a strong market logic thus challenging traditional HR practices.Thornton et al. (2012) suggests that the market logic will supress professional logics.
HR actors in public sector organisations have considerable autonomy and are the guardians of professional HR practice (Bruns, 2014).They place value on standardised guidelines when making HR decisions and may perceive TM as something that falls outside their remit (Jacobson and Sowa, 2016).They orient towards professional rather than strategic interests and are likely to reject the strategic intent of TM practices (Perlman, 2016).In addition, powerful HR actors may invoke particular logics as a way of promoting or protecting their self-interests and these actors will be influential in determining what is implemented (Currie and Spryidonidis, 2016).Public sector HR actors place emphasis on equality, development and steady career progression (Perlman, 2016) and may reject TM practices that emphasise differentiation and exclusive talent approaches (Collings et al., 2017).State logics potentially inhibit HR actors to consider performance metrics so TM practices that involve performance metrics and objective measurement of employee abilities are likely to be rejected.
Inter-organisational collaboration is something that HR actors may find challenging because the state logic, for example, gives emphasis to bureaucracy, accountability, correctness and autonomy of individual organisations to act (Skelcher and Smith, 2015).Public sector organisations are characterised by formality, centralised decision making, procedural correctness, legality, neutrality, the need to serve the interests of the public, and the status quo (Norris-Tirell and Clay, 2010).There may also be a confused sense of authority, poor delegation, lack of ownership and lack of consensus about what should be done (McGuire, 2006).Therefore, inter-organisational collaborations to realise TM objectives may stumble because individual actors often participate in these initiatives with concern about territory, protecting their patch and an unwillingness to divulge sensitive information.
An important gap in the literature concerns the role of institutional logics during each stage of implementing inter-organisational collaboration around TM and how conflicts were resolved, which is something we address in this study.For example, HR actors may bring to the table different logics concerning approaches to collaboration and the criteria for success.Johnson et al. (2003), for instance, differentiated between the champions of the collaboration and those responsible for its day-to-day operation.De Lancer Julnes and Holzer (2001) found that conflicts between individual actors often occur when it comes to developing practices and the identification of performance metrics.Inter -organisational collaboration was less successful where there was a requirement for actors to focus on service measurability or where performance was measured in a rigorous manner (Amirkhanyan, 2009).They will initially coalesce around ambiguous goals and see less value in implementing specific performance measures (Huxham and Vangen, 2000).They may also develop new practices in response to the institutional context through what Lander et al. (2013, p.132) describe as 'blending experimentation'.HR actors will respond differently to conflicting institutional logics (Andersson and Liff, 2018).Reay and Jones (2016), for example, highlighted how individual actors reinterpret and rearrange institutional logics that surround their professional roles.McPherson and Sauder (2013) suggested that individual actors can hijack an entire professional logic to suit their interests.Actors can work within competing logics where, for example, they create a common identity that balances different logics, or cherry pick the logics that best suit their situation (Pache and Santos, 2013).Higher status actors may have greater impact over what logics are used and interpreted and the purposes for which they are used (Greenwood et al., 2011).There may also be negative consequences with logics such as those focused on legalistic-bureaucratic values and cognitions resulting in actors behaving in ways that are rigid and focused on turf wars (Ward et al. 2018).

Research setting
In line with calls for advancement of research concerning the 'sense-making activities of individual agents in organizations' ( (Dalpiaz et al., 2016, p.913) and the enactment of multiple institutional logics by powerful agents in particular settings (Spicer and Sewell, 2010), we chose the Scottish public sector as the empirical setting for our study.Many public sector organisations are now mandated by law to work together to deliver services collaboratively.
The introduction of the Community Empowerment Act (2015) in Scotland and major developments in health and social care integration have provided an impetus for public services to better integrate and collaborate.For example, the Public Bodies (Joint Working) (Scotland) Act 2014 provided a framework for integrating adult health and social care services, creating Integration Authorities with statutory responsibilities to coordinate local health and social care services, with a view to breaking down barriers to joint working between NHS Boards and local authorities.This resulted in significant challenges around workforce planning and development, with Audit Scotland recommending that 'there is a pressing need for workforce planning to show how an integrated workforce will be developed ' (2015, p.33).The participating organisations were excellent exemplars operating under strong state, bureaucratic and professional logics.Therefore, the implementation of TM practices with a strong market logic made the context an ideal one.

Data collection
Direct access to HR actors involved in the inter-organisational collaboration ensured richness of data.We relied on two data sources in this study: fieldwork in the form of semi-structured interviews with 24 day-to-day actors, or what we describe as actors who were on the ground when it came to the implementation of the initiative, and document analysis of key public documents related to the need for inter-organisational collaboration within the Scottish public sector.

Interview:
We used semi-structured interviews and informal conversations with key informants (HR actors) to gain insights into their reasoning and reflections about the introduction of TM and their experiences of inter-organisational collaboration.Reay and Jones (2016) emphasise that qualitative methods such as semi-structured interviews in particular help researchers to investigate institutional logics.The study participants were key informants, or what Gioia et al. (2012, p.17) refer to as 'knowledge agents', in that they initiated interorganisational collaborations, and they were active participants and recipients of the new TM practices developed.The first and third authors collected the interview data.We used interview guides with a flexible structure of questions, allowing for deviation from the question sequence in order to follow particular and interesting lines of inquiry or to delve deeper into particular topics.We adopted a 'naturalistic' approach (Covaleski et al. 1998, p.306) in which questions were designed to elicit responses on issues related to the collaboration that were relevant to participants.In selecting our sample, we followed the advice of O' Reilly and Parker (2013) who noted that 'in qualitative enquiry, the aim is not to acquire a fixed number of participants rather than gather sufficient depth of information as a way of fully describing the phenomenon being studied ' (p.191).In this case, our interview process followed an earlier questionnaire on TM, which was issued to 91 HR leads from across Scotland's public service organisations.
Contact details for the 91 HR leads were held on a database owned by the Improvement Service.The questionnaire generated 31 responses (34% response rate).Participants were asked within the questionnaire if they would consent to participating in a qualitative data collection exercise and our final sample comprised all those who consented.Interviews typically lasted around 90 minutes, while a few lasted up to two hours.Due to the informed consent of participants, 14 interviews were recorded and transcribed verbatim and detailed hand-written notes taken during the 10 interviews that were not recorded.Interviews followed a protocol comprising six questions exploring issues related to collaboration with emphasis on how and why inter-organisational collaborations emerged, and their operation to date.An important initial stage of the interview involved questions that explored study participants' understanding of TM.We did not provide a definition of TM but instead were interested in eliciting the views and perspectives of HR actors.We reached, in line with other studies, data saturation about two-thirds of the way into our studythat is after interviewing 16 HR actors (see Saunders and Townsend (2016) for a comprehensive review on saturation).(The interview composition and schedule can be viewed at the appendix).

Documents:
Interviews are a valuable way of generating deep insights into the underlying processes and practices of an empirical phenomenon such as inter-organisational collaboration and have been used to investigate network dynamics and inter-organisational collaborations in previous studies (e.g.Fortwengel and Sydow, 2018;Majchrzak, et al., 2015).
However, as pointed out by Eisenhardt and Graebner (2007), they are subject to a number of weaknesses including subjectivity and retrospective sense-making.Therefore, it is important to obtain data from other sources in order to achieve some form of triangulation.For these reasons, we collected archival data including brochures highlighting collaborative initiatives, websites, details of talent development practices, official reports, policy documents, research reports and working papers.(The list of documents is provided in the appendix).These documents provided important insights into the political, policy and contextual factors influencing collaborations, indications of the progress and outputs from these initiatives and how collaborations are sustained or otherwise.They also provided insights into the dominant institutional logics and how they had changed over time.

Data analysis
The data analysis process comprised three steps.While two of the authors engaged in the fieldwork, the first and second author analysed interview transcripts and documentation and discussed emerging themes.Previous research highlights that particular institutional logics are very nuanced and that policy documents have a tendency to blend ideal type institutional logics in a way that proves challenging when it comes to categorisation (Jarzabkowski, et al., 2009).
Given these observations, we conducted the data analysis in an iterative way, continually sensitive to ambiguity and contestation as they were interpreted on the ground by HR actors.
In the first step of the analysis we read both the interview transcripts and the documents closely to identify stages of the inter-organisational collaboration, interpretations of institutional logics and to gain insights into the implementation of inter-organisational collaboration to date.The starting point for our analysis focused on identifying chronological developments in the form of events, stages and decisions using Langley's (1999) temporal bracketing approach.To illustrate the evolving nature of the inter-organisational collaborations and the events to which our study participants made reference to when discussing the process that unfolded, we identified three stages: conceptual and initiation phase, development and implementation phase, and scaling up and sustainability phase (Figure 2).Within each phase we identified events, issues and logics.We then started the process of analysing the data to identify HR actors' interpretations of institutional logics.The two authors discussed these first order codes and in particular focused on areas where there were differences in their analysis.
They then reached agreement on the first order codes.
In the second step of data analysis the authors focused on the relationships between the first order codes and their synthesis into broader more theoretical codes, and to move the data analysis from the descriptive to the interpretative.(Figure 1 illustrates the codes generated form the data).The aggregated themes provide the structure for the presentation of the study findings.

Findings
Three aggregated themes emerged from the data analysis (see Figure 1).The first aggregated theme focused on the salience of institutional logics throughout the inter-organisational collaboration process.The second aggregated theme focused on making inter-organisational collaboration happen around TM.The third aggregated theme focused on the implementation of inclusive, less market-driven approaches to TM and consisted of two secondary codes: (a) rejection of formal approaches to collaboration around TM, and (b) the development of TM practices more aligned to existing logics.To recap and provide some perspective on how these issues were important as the inter-collaboration process evolved, we conceptualise the implementation process as consisting of three phases: concept and initiation phase, scaling up and initial implementation phase, and scaling-up and sustainability phase, which are illustrated in Figure 2.

Challenges of blending the market logic of TM with existing logics
This secondary code consisted of two primary codes which we now describe here as they evolved over the three phases of implementation.

Institutional context of inter-organisational collaboration:
During the conceptual and initiation phase (phase 1, Figure 2) HR actors on the ground had a strong sense of what was required by national mandates.The task that they faced to translate them into practice and they acknowledged that these mandates emphasised a market logic to managing human resources.
They referenced these mandates and policy documents and how these required that long established approaches to HR be replaced by new more strategic HR practices.Study participants made frequent reference to the articulation of the key policy goal of the Scottish Government in its response to the Christie Commission (Scottish Government, 2011) that 'Scotland's public services are being challenged as never before by huge financial cuts from Westminster and by changing patterns of demand' (p.2).These sentiments regarding the challenges facing actors arising from budgetary measures were reflected in the responses of study participants in respect of the need "to do more with less" (Iv14) and "to deliver leaner, better, cheaper and more efficient services…closer to the customer or service user" (Iv6).One study participant emphasised that: "I think they [senior managers] have a good awareness of the general principles of Christie.If you are asking do they understand how to apply the principles, I would say yes in theory, but there's no real appetite to get on with it and get it done" (Iv8).
Study participants were very conscious that what they were asked to do challenged the logics of existing practices around professionalism and more egalitarian approaches to HR.
Challenges involved in translating the national agenda to the ground: During the conceptual and initiation phase the challenges involved in translating the national agenda to the ground level were particularly salient.These challenges in many cases focused initially on the need to collaborate rather than on the implementation of TM.Issues related to fragmentation and insularity emerged, which made collaboration initially difficult.These challenges were reflected in the comments of a participant from higher education who emphasised that "Public sector organisations have been quite insular in the past, quite comfortable, and the drive for a more creative and collective approach wasn't necessary" (Iv10).This view is echoed within Audit Scotland's reports on health and social care integration (2015,2018) and was highlighted by a participant from within health and social care, who explained: "We have got to let people be creative and try out new ideas, use their initiative and have autonomy so that leadership will shine through and they will be able to develop without the confines of being constrained by fear" (Iv1).
The well-established cultural norms acted as a barrier to getting collaboration up and running.
For example, a participant from the NHS reflected that "We have good intentions, but are inherently competitive…currently we don't transfer skills across departments, never mind across organisations" (Iv11).

Difficulties of managing different agendas, organisational boundaries and different TM challenges
This secondary code consisted of three primary codes which we describe here as they evolved over the three phases of implementation.

Visualising what inter-organisational collaboration looked like at a local level:
The data analysis reveals that during phase 1 (Figure 2 This logic led some organisations to perceive collaboration as just another initiative that would not achieve anything of significance.A local authority participant remarked that "In the past there have been [collaborative TM] initiatives which haven't gone anywhere, so there's less appetite now" (Iv2).Lack of time was also regularly cited as a barrier to working collaboratively.An emergency response service participant suggested that "Everybody is too busy getting on with the day job" (Iv13); a point also reflected by a justice administration participant, who added: "I know what I want to do but I hardly have time in the working day to draw breath, so how could I get involved without letting my day job slip?" (Iv8).
The difficulties with the logics of TM: Throughout phases 1 and 2 of the collaboration process, HR actors experienced difficulties with the concept of TM.For example, they reported challenges in engaging with the performance and competitive advantage rhetoric associated with private sector notions of TM.For some, the very terminology itself created a dissonance which resulted in suspicion and scepticism.A government agency participant suggested that "The language [of TM] is meaningless in the public sector" (Iv3) and, as a consequence, they will avoid versions of TM that emphasise that some people are more or less talented than others.
Despite a clear strategic need to engage in collaborative TM, the difficulties arose from the challenges of untangling the conceptual nature of TM and trying to adopt its managerial logic within the context of much stronger professional, state and bureaucratic logics.Consequently, participants described the implementation of collaborative TM as something that is "aspirational" (Iv3; Iv8; Iv14).They viewed TM as something that might involve training and development initiatives, however were sceptical and dismissive of approaches to TM that involved segmentation of workforces on the basis of performance and potential.Arguably, this contrasted with the desires outlined in the Christie Commission (2011), where it recommended "a systematic and coordinated approach to workforce development" (p.39) including brining leadership and management development and production of shared competency framework.
The types of practices demanded of TM, such as competency frameworks, structured performance assessment and difficult discussions about talent differentiation were considered intrusive and did not necessarily align effectively with the strong professional and community logics found in public sector organisations.For example, performance assessment was viewed as "not yet fair or transparent" (Iv14).Participants also viewed competency frameworks with scepticism.A local authority participant highlighted that: "I just don't think the one-size-fits-all approach can work due to the complexities of big organisations…It's difficult to have generic role profiles …it isn't practical…there's disconnect between the corporate view and the view of those who deliver operationally" (Iv5).
The challenges of sharing knowledge and practices: HR actors encountered conflicts of logics related to sharing of knowledge and practice during phase 2 (Figure 2).Collaboration and the requirement to share knowledge conflicted with logics related to organisational autonomy, which are central to the legalistic-bureaucratic logic.A participant from the NHS emphasised that: "We're having to talk and understand each other at a depth we've never had to before.I can't tell you how much effort it's been trying to get my head around how different services work and the politics surrounding this" (Iv9).A participant from an emergency response service also described something that reflects the legalistic-bureaucratic logic: "I try to avoid the words 'empower' and 'innovate' as we are so risk driven… unfortunately we don't have the luxury of empowerment as command control is the only way.I think it's important to recognise that not all organisations across the public services are the same and do have different needs" (Iv13).
Perhaps for this reason, the notion of sharing TM practices, as opposed to co-producing them, was more attractive.For example, a participant from justice administration explained that "We have done some work recently on sharing best practice, along the lines of 'you show me yours and I'll show you mine'" (Iv12).This view was supported by a local authority participant who highlighted: "I would be very wary of collaborating but I think there is an opportunity for sharing learning and understanding what other organisations are doing" (Iv6).

The challenges of making inter-organisational collaboration happen around TM
This aggregated theme consisted of two secondary codes: (a) the priorities of individual HR actors and their organisations, and (b) the influence of powerful actors and their dominant logics.

The priorities of individual HR actors and their organisations
This secondary code consisted of two primary codes which we now describe as they evolved over the three phases of implementation.
The auditing of existing practices and approaches: In terms of events and activities, the conceptual and initiation phase began with efforts to audit current TM practices and to question whether there is need for TM at all.Conflicts around these issues prevailed during phase 2 (Figure 2).This involved examining culture and practices and how they might facilitate or inhibit the process of collaboration around the development and implementation of TM.
Mirroring an important and enduring concern reflected within Audit Scotland's progress report into health and social care integration (2018), study participants highlighted major differences in terms of cultural characteristics and priorities.For example, a participant from the NHS explained that "I've gone to workshops for people to share their perspectives, but they are using the same words with completely different meanings" (Iv9).However, the potential for collaboration in education was highlighted by a local authority participant, describing it in these terms: "At the moment people train in a particular area, e.g. a teacher, social worker, nurse…and you wonder if there's another approach, a more merged qualification that crosses over.What we need are new roles, and new types of more flexible qualifications" (Iv2).

The exploration of opportunities and constraints to collaboration:
The data analysis reveals that during both phases 1 and 2 HR actors had mixed cognitions about the possibilities and constraints of collaborating around the implementation of TM.Some actors understood at a cognitive level that there were merits in collaborating.For example, they described how they were trying to achieve TM approaches similar to those adopted by private sector organisations.
A participant from an emergency service organisation suggested that "People are looking more out with their service to others…I think there's a real interest in seeing what's out there" (Iv13).This position is echoed in the documents analysed.For example, the framework and plan for public health leadership and succession planning (2017) highlights extensive mapping to career and training pathways, as well as skills and competencies.Similarly, the Scottish Government's strategic plan on working collaboratively for a better Scotland (2018) recognised the need for increased partnership working between public and private sector organisations.Participants also highlighted the need to attract, recruit and develop high quality talent, and the subsequent imperative to share available skills to get them through difficult times.In particular, they highlighted the need to develop leadership capability.Some actors expressed the desire to be viewed as exemplars of best practice TM, however they were continually challenged by the desire of some organisations to pursue their own agenda.Similarly, Audit Scotland's progress report ( 2018) flags a number of barriers to be overcome in order to speed up change, including 'a lack of collaborative leadership and strategic capacity…disagreement over governance arrangements; and an inability or unwillingness to safely share data' (p.5).Within our interviews, a local authority participant described it this way: "I so see evidence of localism…Sometimes the challenge isn't that we don't understand the [collaborative] concepts and aspire to be working in these ways, it's that we don't have the skills and capabilities, and even the right people in post" (Iv5).

The influence of powerful HR actors and their dominant logics
This secondary code consisted of two primary codes which we now describe as they evolved over the three phases of implementation.
The ways in which powerful HR actors use institutional logics: Throughout the second phase (Figure 2), actors also encountered a combination of organisational politics, power, dominance and control, reflective of a legalistic-bureaucratic logic.Actors experienced situations where larger organisations were more focussed on addressing their own issues surrounding autonomy, rather than concentrating on collaborative opportunities across public services.For instance, a local authority participant put it this way: "We are still working at the boundaries of our own organisation and are at the very early stages [of collaboration]" (Iv6).
Actors frequently had long discussions about TM initiatives, but these did not necessarily lead to a common understanding of what would actually happen.A participant from a government agency suggested that: "People are trying to get their own backyard sorted first before collaborating with others.I need to better understand the end product as collaboration for collaboration's sake are empty words" (Iv3).
Desire to retain control over the HR agenda: While some actors expressed the desire to be viewed as exemplars of best practice TM, they were continually challenged by the desire of other organisations to pursue their own agenda.In the case of smaller organisations, interorganisational collaboration was difficult, especially when having to compete with larger organisations.A participant from justice administration stated: "I think if you asked our chief executives if there's an appetite to develop collaborative TM initiatives, the answer would be 'yes', as long as it's ours and we run it" (Iv12).Actors found it difficult to create expectations, identify ways of working together, resource allocation and who would carry the greatest burden in terms of resources.A local authority participant described it this way: "It's the organising of it all [collaboration] that's the biggest challenge; who's going to contribute what and who's going to pay?" (Iv2).Participants therefore reported that there is often more talk than action, and initiatives often do not get off the ground.

Implementation of inclusive, less market driven approaches to TM
The third aggregate theme consisted of two secondary codes: (a) rejection of formal approaches to collaboration around TM, and (b) the development of TM practices more aligned to existing logics.

Rejection of formal approaches to collaboration around TM
This secondary code consisted of two primary codes which we now describe as they evolved over the three phases of implementation.

Discussion around what types of TM could be implemented and what was unacceptable:
There was much discussion across the three phases of implementation about what types of TM practices were acceptable and what were not.For example, a higher education participant emphasised the types of TM that might fit with current logics, suggesting that "It's probably quite easy to sell the notion of 'talent' if it's inclusive and holistic as opposed to exclusive and elitist" (Iv10).There was much discussion of the notion of inclusive talent.It was described by a government agency participant this way: "It's inclusive in that everyone has got talent, and tied up in routine learning and development…Some people have more stretching objectives than others because they are more capable, or that they know that might place them better for the future.There's no differentiation of levels of talent, but I'm potentially interested in this for the future.This is currently aspirational" (Iv3).
Challenges in aligning TM with existing HR systems: HR actors during phases 2 and 3 ( Figure 2) emphasised the challenges of aligning market logic TM with the existing bureaucratic HR systems currently in operation For example, a local authority participant highlighted the challenges involved in aligning TM with existing practices: "It concerns me, yes, but we have another project coming up looking at our approach to 'career pathing' or 'career development', so there should be something for everyone giving people the opportunity to develop their career regardless of their role" (Iv6).Another local authority participant described it this way: "I just don't think the one-size-fits-all approach can work due to the complexities of a big organisation.The Council has 18,000 staff so to think that you could actually come up with a single approach has failed in a number of areas.I think the best we could do would be to come up with a set of agreed principles and then based on specific circumstances leave it to the directorates as to how they are going to apply them" (Iv5).

Development of TM practices more aligned to existing logic
This secondary code consisted of two primary codes which we now describe as they evolved over the three phases of implementation.
Emphasis on a development focused approach to TM: In the third phase of interorganisational collaboration (Figure 2), things began to develop at a slow pace and not necessarily in the way anticipated by powerful organisational actors.While the Scottish Government in its response to the Christie Commission acknowledged that 'growing mutual understanding and shared purpose breaks down barriers between organisations' (2011, p.13), the reality on the ground was something different.As noted earlier, in the language of various reports between 2011 and 2018, emphasis continued to be on developing more sophisticated cross-sector approaches to workforce development, the extent of collaboration remained fragmented.Nevertheless, some actors, for instance, through Workforce Scotland,1 actively sought to up-scale collaborative activity and continued to develop publicity materials advertising various talent development initiatives.These included leadership development, communities of practice, and cross-organisational coaching, mentoring and facilitation.
Throughout this period, a form of informal, or voluntary approach, to collaboration also emerged with the establishment of networks and sharing of knowledge at senior levels.
The extent of knowledge sharing and co-production of TM practices: During phase 3 the data reveal some evidence of progress, however the reality of inter-organisational collaboration around TM occurred at a very slow pace, and not in a form that powerful organisational actors anticipated.Authorising environments were often complex and ambiguous.Participants reported that the HR systems required did not exist to track success and determine the quality of outcomes.They found it difficult to engage with evidence-based approaches to the implementation of TM practices.There was increased evidence of more informal and fragmented collaborative interactions, such as some organisations offering out places on internal training courses.However, due to competitive legalistic-bureaucratic logic conflicts, there was less scope for bottom-up collaboration between organisations leading to the co-production of TM practices.A participant from the NHS described it this way: "We don't get many people asking 'how do you do that?'It's more 'this is how we do it'" (Iv11).This observation again foregrounds the dominance of legalistic-bureaucratic logics.Where organisational politics, issues of power, and autonomy competition come to the fore, these hampered inter-organisational collaboration.In this phase, there was modest evidence that new practices had emerged due in the main to the difficulty of overcoming arguments and ideas from professional logics.Actors had not reached the point where they could express ideas, leverage practices and implement them collaboratively.They frequently fell back on debates about autonomy and that the practice would not work in their organisation.As a result, interorganisational collaborations did not always take on an independent life and flourish within the organisational field.

Discussion
The implementation of TM in public sector organisations requires the integration of new institutional logics by HR managers into their daily functioning.However, the literature acknowledges that this is not an easy task given the prevalence of professional and state/bureaucratic logics in these organisations.Studies that utilise institutional logics research have primarily focused on empirical setting characterised by one unitary entity rather than in contexts with institutional complexity in the case of inter-collaborations between many public sector organisations (Battilana et al., 2015).We therefore seek to contribute to these debates over how HR actors on the ground deal with these conflicting logics by examining how they respond to the implementation of TM through inter-organisational collaborations.
We first explored the role of institutional logics in shaping both the processes of interorganisational collaboration and the development and implementation of TM.We found that HR actors were continually preoccupied with the interpretation, blending and reinterpretation of institutional logics.actors viewed both TM and inter-organisational collaboration as ambiguous concepts, thus providing scope to decide how they were enacted.Other actors viewed TM as market driven and inter-organisational collaboration as problematic because it impacted their autonomy and decision making scope within their individual organisations.
Day-to-day actors sought to maintain TM practices associated with their professional and bureaucratic logics.Professional logics presented challenges to engaging with the market logic of TM.HR actors found difficulties with dimensions of the market logic of TM such as talent differentiation, segmentation, the measurement of performance and high potential talent (Collings et al., 2017;Schuler et al. 2011).They also found difficulty in blending legalisticbureaucratic and state logics with a strategic view of TM.These logics emphasised a version of TM focused on well-defined managerial and specialist roles and equality in the application of TM practices, something which conflicted with a strategic view of TM.
Similar challenges emerged in the context of inter-organisational collaboration.The inability to blend and integrate different institutional logics impacted the pace of implementation and the types of collaborations that emerged.HR actors' experience of working within a strong legalistic-bureaucratic logic made it difficult for them to comprehend interorganisational collaboration.They understood the economic and functional arguments for inter-organisational collaboration, however professional logics made it difficult to appreciate and accept its practical value.This resulted in gradual or incremental progress with piecemeal and fragmented collaboration.HR actors therefore sought avenues in some cases to blend their existing logics with a form of collaboration that was acceptableinformal collaboration (Smets et al., 2015).Powerful HR actors were instrumental in advocating a particular institutional logic and in having this influence both the collaboration and the type of TM practices that were developed and implemented.
We found that over time institutional logics worked in different ways.The initial efforts at collaboration could be described as unorganised and fragmented, with many dispersed activities undertaken largely arising from the different institutional logics that were at play.This evolved into something slightly less fragmented and focused on discussion of challenges around collaboration and the feasibility of implementing TM in the public sector.A third phase focused on the implementation of initiatives, such as the delivery of collaborative training.
However, what was most noticeable about the initial mobilisation stage was its informal and haphazard nature.We found significantly less evidence that these initial steps led to collaborations that facilitated the development of co-produced programmes and increased sophistication in the way that organisations approached TM.Knowledge sharing and programme initiatives were often undertaken by individual HR actors.In addition, some institutional logics continued to play a role over the duration of the implementation process.
For example, the professional logic which we equate with a HR professional logic had a particularly pervasive influence in shaping both the types of TM practices implemented and the nature of inter-organisational collaboration efforts.The strength of this professional logic resulted in many issues around implementation never coming to resolution and they were particularly influential in how HR actors developed and implemented TM, and their unwillingness to surrender autonomy to achieve a more formal type of inter-organisational collaboration.They were particularly salient in phase three when it came to taking action.In many ways the TM practices implemented were strongly aligned with these logics and the market logic component was diminished or taken out of the types of practices developed.
Theoretically, our study is one of the relative few empirical studies conducted in a public sector context which is highlighted as a unique context when it comes to the implementation of TM (Blom et al., 2018;Thunnissen, and Buttiens, 2017).While studies have generally adopted macro level approaches to the study of TM in this context, we focus on the micro or practice level and bring into focus the role that institutional logics play in the context of TM practice implementation.We specifically build on research by Alvehus (2018, p.39) who found that HR actors were 'active co-constructors of organisational reality, employing several forms of reconciliation' of conflicting logics.This is an important finding and it emphasises that much of the implementation literature in the context of HRM has to date taken too much of a macro level of analysis and assumed a relatively streamlined process of implementation (Trullen et al., 2020).Implementation of HR practices such as TM in public sector organisations is highly political and bounded within existing institutional logics.
Our study also contributes to the literature on institutional logics in the context of TM implementation in public sector organisations.In particular, we found evidence that institutional logics are blended and become blurred over time (Battilana, 2011).This finding supports institutional logics in other research contexts which reveals that multiple logics can be re-arranged and enacted in different ways (Pache and Santos, 2013) and the findings reveal the malleability of institutional logics (McPherson and Sauder, 2013).Our study findings reveal the agency of HR actors and their capacity to develop a much softer version of TM that better fitted their professional logics.The findings also shed light on the actual and intended gap in the context of TM practice implementation and the reality of significant gaps between intended and actual TM practices (Tyskbo, 2019).This is an important finding because it goes against a significant amount of the normative literature on TM emphasising the value of strategic approaches and their links with what happens on the ground (Collings et al. 2017).
Finally, in the context of understanding institutional logics the study reveals something of the latent nature of these logics and their potential to be enacted in situations where HR actors feel that their interests, professional role and position are threatened, suggesting that different logics will find it difficult to coexist with each other (Waldorff et al., 2013).

Limitations, future research and practice implications
First, we acknowledge the inherent limitations with convenience sampling, in this case our reliance on study participants opting in and accepting our invitation to participate.We focused on a distinct set of public sector organisations and acknowledge that these organisations are highly variegated.Moreover, while a longitudinal approach to our data collection would have been preferable in order to track participants' experiences over time, this was not possible within the scope of our study.We did, however, mitigate this limitation as far as practicable by surfacing the temporal issues or stages of implementation of the TM collaboration.In terms of future research questions there is scope to undertake research on the role of institutional logics in the context of the implementation of other HR practices and in contexts other than public sector organisations.It is possible that the influence of institutional logics many have less of an impact in SMEs, for example, where the owner manager is influential in decision making (Bertels and Lawrence, 2016).In the case of MNCs, additional factors will likely come into play related to cultural differences and institutional distance.There is also scope to conduct research on how differences in HR actor knowledge and competencies influence the implementation process and how they blend and mould logics to suit their context (Alvehus, 2018;Bévort and Suddaby, 2016).
Our findings have a number of practical implications.First, the decision to establish inter-organisational collaborations to deliver HR practices such as TM require careful analysis of the institutional logics that may facilitate or inhibit their establishment (see, Alvehus, 2018;Tyskbo, 2019).For example, we illustrate in Figure 2 the emergence of strong professional and bureaucratic logics often found within public sector organisations (cf.Ferry et al., 2019), which have the potential to conflict with the market-driven logic of HR practices associated with TM.This clearly will dictate the type soft practices that are implemented.It suggests that initially HR practitioners in public sector organisations should focus on talent development initiatives, self-assessment of performance and potential and practices that enhance career development.
These are more likely to gain traction.
Effort should be devoted to understanding the implications of these logics for the potential viability of the collaboration.Second, our findings highlight that TM should not be viewed as an independent, decontextualized set of practices.On the contrary, effective practices aimed at attracting, developing, deploying and retaining talent in public sector organisations are dependent upon organisational actors bringing different institutional logics to our understanding of TM (Boselie and Thunnissen, 2017).For TM to be successful in the public sector, it must accommodate the strong professional logic that operate in these organisations.This will require sensitivity to what can work and not work and, in terms of the latter, initiatives designed to stratify talent and categorise based on potential are likely to fail.
Third, inter-organisational collaborations between public sector organisations to deliver HR may work differently from that intended.In particular, goals that focus on the co-production of practices will be difficult to achieve in the short-term (Nabatchi et al., 2017).Therefore, it makes sense to identify small wins and focus on less controversial initiatives that have benefit for many of the collaborating organisations.

Conclusion
In this study we shed light on the role of conflicting institutional logics in a setting in which multiple organisations and HR actors came together to develop and implement TM practices in Scottish public sector organisations.We found that the interplay of institutional logics was in a constant state of flux over the duration of the collaboration and the proved resilient in the face of external mandates and the power of influential HR actors.Our study highlights important insights that can inform future research.
here] Salience of institutional logics throughout the inter-organisational collaboration process This aggregated theme incorporated two secondary codes: (a) challenges of blending the market logic of TM with existing logics and (b) the difficulties of managing different agendas, organisational boundaries and different TM challenges.
) actors had to navigate different challenges and logics in response to mandates for inter-organisational collaboration and become familiar with what was possible.Actors spent time orienting themselves with the changed landscape, seeking advice and seeking to translate what these mandates meant in terms of actual collaboration and the types of TM practices that could be implemented.HR actors were still unsure as to what collaboration would look like during phase 2 and what types of TM practices could be implemented.Sometimes powerful actors suggested that organisations should respond in aceremonial way with new mandates and had doubts about the value of collaboration.These doubts were often related to conflicts between a market logic notion of collaboration and legalistic-bureaucratic logics.For instance, a participant from the NHS explained that "The Scottish Government wanted this [collaboration] to happen for a long time.It seems that things don't happen unless you tell them and force it.Otherwise people find a way not to do it" (Iv9).