DISTRIBUTED LEADERSHIP AND THE MALAYSIA EDUCATION BLUEPRINT: FROM PRESCRIPTION TO PARTIAL SCHOOL-BASED ENACTMENT

Purpose The purpose of the paper is to present and discuss the findings from research on the relationship between leadership theory and policy reform in Malaysia. The research was conducted in two dissimilar Malaysian states. Research Design The research was a multiple case-study design, with 14 schools (seven in each state). Within each school, interviews were conducted with principals (secondary schools), headteachers (primary schools), teachers, middle leaders, and senior leaders, to achieve respondent triangulation. Findings The findings confirm that the Malaysia Education Blueprint prescribes distributed leadership. Most schools embraced an allocative model, with principals sharing responsibilities with senior leaders in a manner that was indistinguishable from delegation. Research Implications A significant implication of the research is that policy prescriptions within major reform initiatives can lead to unintended consequences when applied in different cultural contexts. While distributed leadership is presented as ‘emergent’ in the international literature, it has been adapted for use in this highly centralised context, where structures assume a top-down model of leadership. Practical Implications The main practical implication is that principals and head teachers are more likely to enact leadership in ways which are congruent with their cultural backgrounds and assumptions than to embrace policy prescriptions, even in this centralised context. Originality The paper is significant in exploring a popular leadership model in an unfamiliar context. It also has wider resonance for other centralised systems which have also shown interest in distributed leadership but have been unwilling to embrace it in the ways assumed in the literature.


Introduction
Many countries are seeking to improve their education systems in order to enhance their competitiveness in an increasingly global economy.Referring to Asia's tiger economies, Hallinger (2004, p. 63) argues that 'global economic competition has raised the stakes for educational attainment, individually and collectively.Consumers now define the meaning of quality education globally, rather than locally or nationally'.The growing importance of international comparisons of student learning outcomes, notably the Programme for International Student Assessment (PISA), increases the visibility of different levels of performance and often informs national reform initiatives.
Many countries use international comparisons, such as PISA, as levers to evaluate their own education against other systems.Malaysia is one such country and its educational reform agenda is informed by the PISA scores.The Malaysia Education Blueprint (MEB) (Ministry of Education, 2013) is the major policy document driving reform.It is explicit about benchmarking Malaysian student performance against international norms.According to MEB (2013, E6), 'other systems are improving student performance more rapidly, and have found ways to sustain that momentum.The gap between Malaysia's system and these others is therefore growing.However, Hallinger (2010, 409) cautions against policy borrowing when seeking school improvement.Some 'education reforms have travelled around the globe far from their points of origin and often appear "foreign" upon arrival in South-East Asia'.
The MEB outlines an ambitious vision to raise Malaysia's learning outcomes from their current position in the bottom quartile of PISA scores in reading, mathematics and science: example in neighbouring Thailand (Hallinger and Lee, 2014).'Despite new system expectations . .., the predominant orientation of Thai principals remains largely unchanged' (Hallinger and Lee,p. 6).
The Blueprint makes several references to the significance of distributed leadership in achieving the Ministry's aims, stating that 'in line with international best practices, the Ministry will move towards a model of distributed leadership where effective, high quality school leadership permeates the entire organisation of all schools' (MEB, 2013, 18).However, this model, and most of contemporary leadership theory, was developed and honed in Western contexts, raising questions about its suitability for Asian contexts, including Malaysia, where education systems are highly centralised (Walker and Hallinger, 2015).This prompted the authors to conduct research on whether, to what extent, and in what ways, distributed leadership is practiced in Malaysian schools.The Blueprint's ambitious agenda is intended to bring about enhanced student outcomes through changing leadership practice from the dominant managerial role practiced in Malaysia and in many other centralised systems (Hallinger and Lee, 2014).However, such radical changes are difficult to achieve because of deeply-embedded cultural expectations within a society where 'power-distance' (Dimmock and Walker, 2002;Hofstede, 1991) is accentuated.
As in neighbouring Thailand, Malaysian principals are civil servants who function as line managers within the hierarchy of a highly centralised national system of education (Hallinger and Lee, 2014).This systemic culture suggests that shifting leadership practice from solo to distributed leadership will be difficult to achieve.

Distributed Leadership: An Outline Literature Review
The study began with a systematic review of the literature, in English and Bahasa Malaysia, using the search terms distributed, shared and teacher leadership (Bush, Ng, Abdul-Hamid and Kaparou, 2018).The inclusion of Bahasa Malaysia sources helps to address a fundamental weakness in much of the current literature; an almost total reliance on English language sources.Hallinger and Chen (2015, p. 21) note the problem of 'a hidden literature' with 'a substantial number of research papers . . .written in indigenous languages'.These sources are often ignored in systematic reviews, so their inclusion here provides a more robust starting point for the review of previous research on distributed leadership.

Leadership theory
There are numerous leadership theories, which seek to explain the behaviours and actions of school leaders.Yukl (2002, p. 4) argues that 'the definition of leadership is arbitrary and very subjective', but the following 'working definition' includes its main features: 'Leadership is a process of influence leading to the achievement of desired purposes.Successful leaders develop a vision for their schools based on their personal and professional values.They articulate this vision at every opportunity and influence their staff and other stakeholders to share the vision.The philosophy, structures and activities of the school are geared towards the achievement of this shared vision' (Bush and Glover, 2003, p. 5).
Theory is valuable and significant if it serves to explain practice and provide leaders with a guide to action.Theories are most useful for influencing practice when they suggest new ways in which events and situations can be perceived (Bush, 2011).This article tests the application of leadership theory in the specific context of Malaysia.Some of the most prominent leadership models are managerial, transformational, distributed and instructional leadership.These theories are mostly normative, with their advocates stressing their utility as the 'best' way to lead and manage schools (Bush and Glover, 2014;Leithwood et al, 1999).
Distributed leadership has become the most fashionable leadership model in the 21 st century, with numerous books and journal articles focusing on this theme (Bush, 2019;Harris, 2010).It provides the theoretical framework for the study because of its current popularity and because it is advocated in the Blueprint (see above).The rationale for this model is that leadership is too complex to be handled purely through solo leadership.By increasing leadership density, through empowering more leaders, there is potential for enhanced learning outcomes (Bush and Glover, 2014).
Distributed leadership is one of several models which stress shared approaches to leadership (Crawford, 2012).Collegial and participative leadership were popular shared approaches in the late 1900s but distributed leadership has become the normatively preferred leadership model in the 21 st century.Gronn (2010, p. 70) states that 'there has been an accelerating amount of scholarly and practitioner attention accorded [to] the phenomenon of distributed leadership'.Harris (2010, p. 55) adds that it 'represents one of the most influential ideas to emerge in the field of educational leadership in the past decade'.

Understanding distributed leadership
An important starting point for understanding distributed leadership is to uncouple it from positional authority.As Harris (2004, p. 13) indicates, 'distributed leadership concentrates on engaging expertise wherever it exists within the organization rather than seeking this only through formal position or role'.Harris (2010, pp. 55-56) defines it as: 'The expansion of leadership roles in schools, beyond those in formal leadership or administrative posts . . .[it] concentrates on the interactions rather than the actions of leaders'.Gronn (2010, p. 70) refers to a normative switch 'from heroics to distribution' but also cautions against a view that distributed leadership necessarily means any reduction in the scope of the principal's role.Indeed, Hartley (2010, p. 27) argues that 'its popularity may be pragmatic: to ease the burden of overworked headteachers'.Lumby (2009, p. 320) adds that distributed leadership 'does not imply that school staff are necessarily enacting leadership any differently' to the time 'when heroic, individual leadership was the focus of attention'.Bennett et al (2003, p. 3) claim that distributed leadership is an emergent property of a group or network of individuals in which group members pool their expertise.Harris (2004, p. 19), referring to an English study of ten English schools facing challenging circumstances (Harris and Chapman, 2002), says that there should be 'redistribution of power', not simply a process of 'delegated headship'.However, Hopkins and Jackson (2002) argue that formal leaders need to orchestrate and nurture the space for distributed leadership to occur, suggesting that it would be difficult to achieve without the active support of school principals.Heads and principals retain much of the formal authority in schools, leading Hartley (2010, p. 82) to conclude that 'distributed leadership resides uneasily within the formal bureaucracy of schools '. Bottery (2004, p. 21) asks how distribution is to be achieved 'if those in formal positions do not wish to have their power redistributed in this way?' Harris (2005, p. 167) argues that 'distributed and hierarchical forms of leadership are not incompatible' but it is evident that distribution can work successfully only if formal leaders allow it to take root.In their meta-analysis of distributed leadership, Tian et al (2016, p. 153) add that, 'in a distributed leadership setting, formal leaders should also be regarded as important "gate keepers", who either encourage or discourage others from leading and participating in organisational changes.' Gronn's (2010, p. 74) overview of four research projects leads him to conclude that principals retain considerable power.'Certain individuals, while they by no means monopolized the totality of the leadership, nonetheless exercised disproportionate influence compared to their individual peers'.Harris (2005, p. 167) argues that 'distributed and hierarchical forms of leadership are not incompatible' but it is evident that distribution can work successfully only if formal leaders allow it to take root.Gunter et al (2013, p. 563) argue that 'normative work tends to present the idea of distributed leadership as an imperative for practitioners as school improvers' and adds that the 'lack of substantial and robust data can make exhortations to adopt distributed leadership problematic' (Ibid, p. 565).

Distributed leadership and student outcomes
The interest in, and support for, distributed leadership is predicated on the assumption that it will bring about beneficial effects that would not occur with singular leadership.Leithwood et al's (2006, p. 12) important English study shows that multiple leadership is much more effective than solo approaches: 'Total leadership accounted for a quite significant 27 per cent variation in student achievement across schools.This is a much higher proportion of explained variation (two to three times higher) than is typically reported in studies of individual headteacher effects'.Leithwood et al (2006, p. 13) add that schools with the highest levels of student achievement attributed this to relatively high levels of influence from all sources of leadership.Distributed leadership features in two of their widely cited 'seven strong claims' about successful school leadership.Hallinger and Heck (2010) also found that distributed leadership was significantly related to change in academic capacity and, thus, to growth in student learning.

Limitations of distributed leadership
As suggested earlier, the existing authority structure in schools and colleges provides a potential barrier to the successful introduction and implementation of distributed leadership.'There are inherent threats to status and the status quo in all that distributed leadership implies' (Harris 2004, p. 20).Fitzgerald and Gunter (2008) refer to the residual significance of authority and hierarchy.
As noted earlier, the Blueprint (MEB 2013, E28) suggests a shift towards distributed leadership.'The aspiration is to create a peer-led culture of professional excellence wherein school leaders mentor and train one another, develop and disseminate best practice, and hold their peers accountable for meeting professional standards'.

Previous Research on Distributed Leadership in Malaysia
Perhaps because Malaysia has a highly centralised system, there is only limited research on distributed leadership in this context.Jones et al's (2015) study of principals' leadership practices in Malaysia provides evidence of principals' transformational and distributed practices attributed to their emerging accountability for school outcomes.They conclude that secondary school principals are 'increasingly seeing themselves as leaders who are responsible for change and empowering others' (Jones et al, 2015, p. 362).Abdul Halim's (2015) correlational study, involving 831 teachers in 17 residential and national secondary schools, found a significant positive relationship between distributed leadership and teachers' self-efficacy.The author reports that teachers' self-efficacy is relatively high in residential schools compared to national secondary schools.Boon and Tahir's (2013) survey of 600 senior and middle leaders in Johor involved three questionnaires on distributed readiness, work stress, and organisation commitment.By using structural equation modelling, they found positive relationships between the dimensions of leadership, work stress and work commitment among middle managers.Fook and Sidhu's (2009, p. 111) research showed evidence of 'distributing leadership . . .through the development of macro and micro management teams' to contribute to the management of change.Rabindarang et al's (2014) explanatory mixed methods study included a questionnaire, completed by 359 teachers, and interviews with four teachers.Their study established that distributed leadership reduces job stress among teachers in technical and vocational schools.Abdullah et al (2012) studied distributed leadership in a daily premier School in Selangor.They identified three elements of distributed leadership; sharing the school's goal, mission and vision, school culture (cooperative, collaboration and professional learning community), and sharing responsibilities.Zakaria and Abdul Kadir (2013) studied the practice of distributed leadership among teachers in a city in north Malaysia, based on demographic factors using the Distributed Leadership Inventory developed by Hulpia et al (2009).The findings showed that distributed leadership was only moderately practiced by the teachers in the city, for example in respect of participative decision making, cooperation within the leadership team, and leadership supervision.Norwawi's (2017) research on leadership in high performing schools showed evidence of distributed leadership but this appears to be 'allocative' (Bolden et al, 2009) rather than 'emergent' (Bennett et al, 2003), with principals delegating tasks to their senior and middle leaders rather than empowering them to act independently.
This limited body of literature shows some evidence of the existence of distributed leadership in some Malaysian schools, for example through team-work.It appears to have enhanced teacher self-efficacy and reduced teacher stress.Perhaps as a consequence, teachers feel empowered and may enhance their commitment.However, despite its normative emphasis in the Blueprint, the literature suggests two cautions.First, distributed leadership may be practiced only moderately.Second, the model appears to be allocative, consistent with the hierarchy, rather than emergent.More work is required to establish whether and how distributed approaches can be meaningful in this hierarchical context.The present research contributes to this knowledge 'gap'.
Teacher leadership is often aligned with distributed leadership as distribution often involves classroom teachers (Bush and Glover, 2014).Although teacher empowerment has been considered as an integral element of the attempt to move towards decentralisation from a highly centralised education system (Lee 1999), there is limited evidence within Malaysia (but see Jones et al, 2015).The Blueprint stresses the need to enhance 'attractive' pathways into leadership for teachers.This might include becoming subject specialists, focusing on developing curriculum and assessment.
Another dimension of teacher leadership highlighted in the Blueprint is that of master teachers.Lee (1999: 93) highlights the emergence of 'master teachers' in Malaysia, whose role was mainly targeted to 'pedagogical guidance to their own colleagues'.Bush et al's (2016) study of master teachers in Malaysia and Philippines, drawing on interviews with master teachers, principals and teachers, show that they occupy the hinterland between formal and informal teacher leadership.In both countries, their work is legitimised by their appointment to an established position with enhanced salary and status.They conclude that 'the advent of master teachers in both countries has succeeded in keeping talented and ambitious teachers in their classrooms but their leadership role is patchy and depends on personal variables rather than school or system endorsement' (ibid, p. 37).They also note that the development of teacher leadership has been limited because of the emphasis on the formal hierarchy.Ngang's (2012) research on teacher leadership in special education classrooms in China and Malaysia reveals that teacher leadership is evident in classroom management in both countries.The Malaysian evidence arises from a survey of 369 special education teachers in Peninsular Malaysia.The paper suggests the provision of training for teacher leadership, and capacity building.The role of teachers in building capacity within schools has attracted attention and Harris et al (2013, p. 217) argue that: 'In Malaysia, which aspires to be high performing, the Education Blueprint . . . is the clearest signal yet that collaborative professional learning is viewed as a potential strategy for securing educational improvement and change.It reinforces collective professional learning as a means of transforming education quality and performance'.Park and Ham's (2016) quantitative study of three countries, Australia, Malaysia and South Korea, found that an increased level of effective interaction between principals and teachers leads to consolidation of trust, and enhanced teacher collaboration.The Malaysia Education Blueprint (2013) also discusses the pathway to teacher leadership.However, the limited Malaysian research on teacher leadership tends to align it with formal roles, such as master teacher.This seems to limit the scope for 'emergent' teacher leadership, arising from personal initiative.
Distributed leadership appears to be allocative, consistent with the hierarchy, rather than emergent.Distributed leadership also appears to have enhanced teacher self-efficacy and reduced teacher stress.Perhaps as a consequence, teachers feel empowered and may enhance their commitment.
The review suggests a gap between distributed leadership theory, developed in Western contexts with high degrees of decentralisation, and leadership practice in centralised contexts such as Malaysia, where teachers feel constrained by the hierarchy.These insights provided the starting point for the authors' research on distributed leadership in 14 schools.

Research Methods
This paper reports the authors' study of the application of school leadership theory in selected schools in two Malaysian provinces, Selangor and Sarawak, funded by the Ministry of Education.The purpose of the study was to examine the nature and extent of distributed leadership in selected Malaysian schools, in order to assess whether, how and to what extent, the Ministry's advocacy of this model was borne out in practice.This led to the following research questions: 1. What leadership theories are manifested in Malaysian schools? 2. How, and to what extent, is distributed leadership practiced in Malaysian schools? 3. What is the relationship between distributed leadership and student outcomes in Malaysia?4. What combination of leadership practices is most effective in facilitating school improvement?The main focus of this paper is research question 2.
Following a sequential research design, the first phase of the research involved case studies of seven contrasting schools in Selangor, the political and economic heart of Malaysia.The second phase featured case studies of seven contrasting schools in Sarawak, an island state remote from the Malaysian peninsula.The research is a multiple case study design.Bassey (2012, p. 156) describes case study as 'an empirical enquiry which is conducted within a localised boundary of space and time'.
Malaysian schools are stratified into seven 'bands', according to Ministry of Education criteria about school effectiveness and achievement, with the most successful schools in band one and the least successful in band seven.The intention was to adopt a stratified sampling frame, with one school from each band in both states.This stratification was intended to establish whether distributed leadership was more prominent in those schools labelled as more effective through the banding process.In practice, it was not possible to include schools from all seven bands.In Selangor, all seven schools were in bands 2, 3 or 4. The sample included one urban and three rural primary schools, as well as one rural and two urban secondary schools.In Sarawak, the seven schools ranged from band one to band five.The sample included one rural and two urban primary schools, as well as four urban secondary schools.
The researchers scrutinised school documents and conducted several interviews in each school.The intention was to interview the principal (secondary schools), headteacher (primary schools), senior leaders, and middle leaders, to build a picture of the extent, nature and pattern of distribution in each school.Access was secured through the Ministry of Education's Planning and Research Division (EPRD).This facilitated access to the 14 schools and the intended participants were asked for their voluntary informed consent to take part in the research.Most readily agreed to participate but a few declined to do so.As a consequence, participant sampling differs, to some extent, across the 14 schools.In total, 95 interviews were conducted, 51 in Selangor and 44 in Sarawak.Interviews typically lasted between 45 and 60 minutes, and were then transcribed.Table 1 shows the interviewees at each school, identified by position.Principals were asked questions about their knowledge of the Malaysia Education Blueprint, especially Shift Five, Leadership, whether or not their leadership practices have changed since the Blueprint was published in 2013, and whether or not they regard themselves as distributed leaders, with probes about how, and to whom, leadership is distributed.The other participants were asked about their own leadership practices and also questioned about whether or not they regard their principals as distributed leaders, with probes about how, and to whom, leadership is distributed.

State
The findings offer a triangulated perspective on the extent and nature of leadership distribution in the case-study schools.The participants also responded to questions about instructional and transformational leadership but these dimensions are beyond the scope of this paper.
The data were analysed sequentially.First, a case study report was prepared for each school, integrating the data from all participants to build a picture of distributed leadership specific to each context.The findings were organised thematically, with most themes linked to the research questions.Second, the seven cases in each state were compared, leading to an overview of distributed leadership in these very different contexts.Third, the findings from each state were compared to build an overall picture of the nature of distribution in these 14 Malaysian schools.

Findings
Distributed leadership features prominently in the Malaysia Education Blueprint, although its focus is almost exclusively on distribution to official leaders, such as senior assistants, heads of department and subject leaders.This is a different interpretation from most of the established (mainly western) literature which stresses that distribution is emergent rather than being linked to the hierarchy, although this may be largely a normative view, even in the west.
Participants at all 14 case-study schools were able to identify aspects of distribution but most of their examples relate to collaborative activity, broadly defined, rather than distributed leadership.The thematic discussion below shows the different interpretations offered by the participants.The six themes are constructs, which emerged from the data, and may be regarded as distinctive features of distributed leadership in Malaysia:

Delegation
The prominence of allocative distribution is shown in the emphasis on delegation, explicit in five schools; E, F, one, two, and seven, and implicit in several others, for example in school six, where senior leaders are free to make co-curricular decisions.Delegation is a management concept, a device for allotting tasks and responsibilities.It aligns strongly with allocative leadership and may be regarded as consistent with leaders' and followers' expectations in a hierarchical system.Delegation is usually portrayed as a linear and vertical process, with a superordinate allotting tasks to a more junior colleague, implying an organisational hierarchy (Connolly et al, 2017).
The head teacher of school E believed that he should know everything that is going on in the school and he practised delegation in a manner that suggests controlled freedom.Initial planning had to be done in consultation with the head teacher.When plans were adapted to incorporate the head teacher's ideas and suggestions, the teachers were allowed to implement the plans or programmes.Staff acceptance of this stance is illustrated by the comments of the senior assistant (academic).
"We can make decisions but, of course, we refer to him . . .we still have a leader and, of course, the leader should have control in the institution.He will and should have the last say".
This view is confirmed by a senior teacher."He is not autocratic.But still in control".
Delegation to the senior assistants in school F was facilitated by the layout of the administrative block, with the principal's office at the end of the corridor.The general administrative office, and the four senior assistants' offices, was towards the front of the corridor, meaning that visitors had to go past all these various offices before reaching the principal's office.This facilitated the senior assistants handling most issues with students, teachers and parents.
"If . . .parents have anything [to discuss] they settle there and then [and] sometimes they just go back.So, we do a lot of distribution of power for the senior assistants".
School one also appears to operate an allocative model, with the principal 'passing down authority' to senior teachers, for example for discipline and to 'control the children', especially if he is out of the school.The senior assistant (academic) says that the level of distribution depends on the type of decision but we 'have to report to him' and such decisions are 'guided'.
Delegation in school two is primarily to senior positional leaders; "he will distribute based on our post" (senior assistant, academic).The assistant principal (student affairs) adds that "the principal will delegate all academic matters to the assistant principal, academic, to ensure that all programmes run smoothly . . .Anything about students, he will delegate to me".The HOD (languages) confirms that "he [the principal] will follow the hierarchy".
At school seven, the niche area coordinator thinks that the principal distributes power but in the form of 'delegating tasks'.'I think she gives us a lot of leeway . . .a lot of freedom'.However, she still needs to report to the principal.

Sharing the workload
A strong motivation for delegation is the need to share the workload, with recognition that the expectations of stakeholders, notably education officials at every level, cannot be met by the principal alone.This was evident at five of the case study schools.The head teacher of school A epitomized this view, stressing the need to work with teachers and the community; "I cannot work alone".The head teacher of school B was also aware that she could not do all the work herself and that she needed others to help: "I have only two hands, I cannot manage so many things and then I need people to do the things and then, if they have good ideas, why not let them do it?" The need to share the work is particularly evident in larger schools.The principal of school D, for example, was aware that he could not run the school alone and needed the help of all teachers, especially senior and middle leaders: "I need those senior teachers and senior assistants to observe them (teachers) and coach them.The same goes to students, because we have more than 2000 students, there is no way I could do it alone by myself . . .We have form teachers, my counsellors, they are my eyes and arms of the law . . .they are the ones who talk to the students and help the students".
The principal of school F also stated that his school is too big and that he needed several teachers to manage the school with him.Similarly, the principal of school seven claimed that 'you must have a team, you cannot want to work by yourself'.

Decision-making
The nature of the decision-making process provides important clues about the extent of distributed leadership in schools.Where this is tightly controlled by the principal, opportunities for teacher initiative are likely to be limited.This was illustrated at school one, where decisions were 'guided' by the principal, and at school four, where decision-making was subject to 'controls'.In contrast, where teachers are able to exercise agency, within only broad accountability parameters, distributed leadership should be able to thrive.While control was strongly evident in the case-study schools, there were a few examples of teacher initiative being encouraged.
The teachers of school B indicated that they were very much involved in decision making related to their subjects.Although important decisions were made by the head teacher, teachers were allowed to make decisions in their zones of expertise."Sometimes they need my decision.Sometimes they make their decisions too" (head teacher).This notion of decision-making zones provides a helpful way of understanding the scope for distributed leadership.
The principal of school D regarded distributed leadership as a way of developing his staff: "I told them (teachers) that I would like to build them up, if they can make decisions, that's good, but any major decision they want to make, it's good also if they discuss with . . . the senior management team because I don't want to be the only person here as the leader.I told them that I like to see parallel leadership because this is a big school".
However, the senior assistant (student affairs) was less confident about making independent decisions.He felt that, in matters pertaining to student welfare and discipline, it would be wise to follow rules and to consult the principal, who should make the final decision, although he could discuss matters with him: "I handle students' cases so I want everything to follow school rules and the circular.So, normally I will not make the final decision but I will try to give some suggestions or my opinions and then I will discuss with the principal and he will come to the final decision".
This example shows that distribution may be constrained by followers being ultra-cautious about exercising leadership, as well as by principal reluctance to share power.
The principal of school F distributed many leadership issues to his two senior assistants, academic, and student affairs.Usually, the teachers would liaise with the appropriate senior assistant, for academic or student-related issues.Matters would be resolved at that level and reported to the principal during the weekly meetings.Only serious issues were referred directly to the principal.The senior assistants were given autonomy to decide on matters at their respective levels.The principal believed in giving them 'space': "I give the teachers space to work.They have freedom to work and, when they have space, they work on their own.I believe they work honestly".This was attested by all the teachers who were interviewed.They felt it easy to work when the principal trusted them.They were not worried about whether what they were doing was right or wrong as they were given the space to do what they thought appropriate for their own students.
The principal of school F was a firm believer in internal promotion.He preferred middle leader posts be filled by teachers from the same school, because they knew the culture of the school and how things were done.He would personally train them for leadership posts.There were weekly meetings with the senior leadership team (SLT) members so that the principal would be briefed about what was happening in the various departments.
The school F example suggests that distribution is more likely to take root where principals have confidence in their colleagues.By encouraging internal promotion, principals are able to allow distribution, knowing that decisions are likely to be consistent with their values and priorities.

Trust
Trust is a key dimension if distributed leadership is to thrive.Senior leaders need to be confident that decisions are consistent with their values, and with school policies, as in the school F example discussed above.Teachers also need to feel confident to enact leadership without fear of negative consequences.In hierarchical systems, such as Malaysia, it is possible that risk-averse teachers may seek approval for their ideas even when this is not a requirement.
Participants at five schools discussed trust as a dimension of distributed leadership.School A has many experienced teachers and the head teacher trusted them to carry out their duties without having to monitor them.The work-focused culture of the school encouraged the development of mutual trust.Leadership is distributed, especially to senior teachers, who are often consulted by the head teacher.It is unsurprising that trust in senior leaders is greater than in the wider body of teachers.
School D is a two-session school, with an afternoon supervisor.The principal of school D stayed in the school until 6 p.m. on most days but he did not interfere with the afternoon supervisor.He trusted him but stayed until 6 p.m. to show his support: "I told parents that . . . the afternoon supervisor is the principal in the afternoon, you go and see him . . .I want him to be in-charge in the afternoon, I stay back until 6pm plus to show I am behind him, but I want him to feel that he can deliver also".(Principal) The principal of school F trusts his teachers and, for example, he allowed them to organize their own professional development programmes.However, he expected the teachers to provide a report of their activities every six months.
At school six, the senior assistant for student affairs implies a distributed approach in saying that the head gives her 'a lot of opportunities' to make decisions, and attributes this to the trust the head has in her.Three other participants also refer to the 'freedom' to develop and implement programmes, but not if they have financial implications.
Trust is also a feature of leadership at school four.The HoD (Languages) stresses that 'trust is important because . . .we have to work together as a team'.The principal provides opportunities for teachers to make decisions and she seeks advice from senior assistants and committees.However, this may be understood as consultation rather than distribution.'She [the principal] will consult with the teachers on everything.Once she has given her consent, then I will make the decision' [but] 'there are controls' (senior teacher).

Consultation
As noted above (school four), some participants referred to consultation when asked about distributed leadership.This is a weaker form of collaborative activity, because advice can be ignored or rejected and final decisions are taken by the principal.Several staff at school three refer to the head seeking their opinions but, as one teacher notes, 'there's discussion, but we always follow her [the principal's] decision . . . in the end, it's always her decision'.
Similarly, at school five, the senior assistant (co-curriculum) notes that the head allows the teachers to plan as a group, rather than taking all the decisions herself.A class teacher says that the leadership is willing to listen to teachers but this is consultation not distribution.

Autonomy
Autonomy is an important requirement for distributed leadership to thrive (Javadi et al 2017).Teachers need a degree of agency to initiate and enact new ideas.As professionals, they have specialist knowledge and skills but they also require the confidence and encouragement to act autonomously, and to understand how their autonomy may be constrained.
Autonomy for teachers, and especially for senior leaders, was evident at six of the case study schools, although it was constrained at four of them.
Senior leaders at school A have a degree of autonomy, as indicated by the senior assistant (academic): "If I can handle it, I will handle it.I don't push everything to the head teacher.She has given me the empowerment but I will inform her after I have solved the situation.If the problem is big, then I will meet the teacher with the head teacher".
The principal of school F did not dictate and teachers were given the autonomy to decide what they should do.The teachers were encouraged to meet him often to discuss their plans and ideas.Because of the large size of school F (more than 2000 students), the senior assistants were given a lot of authority to carry out their tasks.They, in turn, gave discretion to their teachers.There are many committees (including those for time-tables, text books, subjects, and exams).Work was distributed to the various committees, which were allowed to make decisions.However, they needed to report to the senior assistant who, in turn, had to report to the principal.
The school F principal appeared to offer opportunities for teachers to initiate ideas and to work on their own projects, a form of teacher leadership.The principal did not micro-manage their work, believing that, by giving them space and trust, the teachers would flourish professionally and personally.This led to a degree of autonomy, with teachers feeling free to do what they think best for their subjects and for their students.
The principal of school C wanted to know everything that was going on in the school, and this meant only modest autonomy for teachers and other leaders.This was to make sure that she was on top of everything and not caught by surprise if things went wrong.The senior assistant (academic) acknowledges the limited nature of autonomy, even for senior leaders: "I don't make a lot of suggestions but when I do she listens but I need to consult her first because she's the head of the school.Whatever happens later or when people call, she knows.It is easier for us that way" (SA Form 6) The teachers at school G mentioned that autonomy is granted to teachers, for example to conduct extra classes for the weaker students.The principal left it to individual teachers to plan the classes and to identify the students to attend such classes.This may be seen as an example of bounded autonomy within a hierarchy.
Another example of bounded autonomy was evident at school seven.One middle manager claimed that the principal 'gives us a lot of leeway' but she still needs to report to the principal.Similarly, at school one, the principal gives teachers the opportunity to make new policies but, in practice, 'we normally consult the senior assistant first'.

Discussion and Conclusion
The Malaysia Education Blueprint (2013) documents an ambitious attempt to transform the schools' system so that it is among the best in the world, for example in respect of PISA scores.One key dimension of this reform relates to school leadership.The Ministry of Education is exhorting principals and other leaders to move away from administrative leadership and to adopt distributed approaches.There is evidence of the beneficial effects of distributed leadership in international research and literature (e.g.Leithwood et al, 2006).However, there is much less data to support their efficacy in centralised contexts, including Malaysia.
The global popularity of distributed leadership arises from dissatisfaction with the limitations of solo leadership, linked to the hierarchy.In the international (mostly western) literature, it has been conceptualised as an emergent property, uncoupled from the formal roles of principals and head teachers (Bennett et al, 2003;Bush and Glover, 2014).The focus is on expertise, not positional authority, recognising that schools are professional organisations and that talent and knowhow are widespread.However, this perspective may be normative, offering a somewhat romanticised view, rather than reflecting significant empirical evidence about emergent distribution in western contexts.This model is arguably better suited to devolved education systems, such as those in England, Australia and the US, than the centralised systems evident in much of Asia, but distribution is often to formal leaders in both types of context, leading to Gronn (2010) to propose a hybrid model, combining bureaucratic and distributed approaches.
The advocacy of distributed leadership by the Malaysian Ministry of Education is recognition of its potential to enhance leadership density, and thus, potentially, to contribute to improved student outcomes.However, the Blueprint links distribution firmly to the hierarchy in two ways.First, the focus is firmly on middle and senior leaders holding formal roles in the structure.Second, the scope of distribution is circumscribed; leaders will be prepared to fully utilise the decision-making flexibilities 'accorded to them' (MEB: E28) (present authors' emphasis).The Blueprint's cautious approach to distribution is consistent with the notion of allocative distributed leadership (Bolden et al, 2009).This suggests an uneasy compromise between the freeflowing assumptions of distributed leadership theory and the rigid requirements of the hierarchy.There is evidence of allocative distribution, predominantly to senior leaders, in most of the case study schools.The school one leader, for example, only consults his senior assistants, usually through the senior assistant (academic).Whatever decisions are required, they have to go through the senior assistant before going to him.The senior assistant also acts as a 'gatekeeper' for teacher access to the head.Similarly, distributed leadership in school two is allocative, and based on hierarchy.
Allocative distribution has several similarities to the management concept of delegation (Connolly et al, 2017).The head teacher of school six, for example, delegates tasks to her senior assistants, a form of allocative distribution.They carry out duties as instructed by their head teacher.When asked about distributed leadership, participants in several schools referred to delegation.In school seven, a middle leader mentions 'delegating tasks', a 'top-down' process.The evidence from the Malaysian schools is that distributed leadership is almost indistinguishable from delegation, not least because principals remain in control and have firm reporting requirements.Hartley's (2010) view that distributed leadership is popular because it eases the burden of overworked head teachers appears valid in the Malaysian context, as shown in schools A, B, D and F, for example.This is very much a pragmatic view and has little to do with teacher empowerment.Where teacher decision-making is allowed, or encouraged, it is often limited to specific zones, notably classroom practice or subject leadership, rather than being a whole-school activity.
Trust is an important aspect of distributed leadership and this led to a measure of autonomy in some case-study schools, for example in school F, a very large school in Sarawak.Where principals trust their colleagues to take the 'right' decisions, they are more likely to cede autonomy, regarded by Javadi et al (2017) as a significant requirement for distributed leadership to be meaningful.
The 14 schools (seven in Selangor and seven in Sarawak) operate in different ways, and with various degrees of success, evidenced in part by their school bands (1-5).Despite the centralised education system, there is scope for individual agency, allowing principals to act as distributed leaders, as their personalities and contexts indicate.One common feature, however, is that leadership is still largely interpreted as 'headship', with little focus on the roles and actions of other senior and middle leaders.Despite its references to senior and middle leaders, the main focus of the Blueprint is on school principals and how they should be developed to lead their schools more effectively.Accountability in centralised systems is through the hierarchy and Malaysian principals are civil servants employed by the government and this inevitably limits their scope for individual agency.Their centrality means that they also act as 'gatekeepers' (Tian et al, 2016), who can choose to facilitate or inhibit distributed leadership.The evidence from the 14 case study schools is that principals have chosen to use their authority to limit distribution to senior colleagues and to retain overall control of all major decisions.As Hartley (2010) indicates, distributed leadership does not fit easily within school hierarchies.

Limitations
The research reported in this paper makes an important contribution to understanding the nature of distributed leadership in Malaysia and, more generally, in centralised contexts.However, it has three main limitations.First, it was not possible to achieve the planned stratified sample of two schools (one from each state) from each of the seven bands used to classify Malaysian schools.This limited the prospect of cross-band comparisons.Second, the data are limited to only two Malaysian states.Including all states may have modified the findings.Third, while the 14 schools provide helpful illustrative data, it is not possible to generalize the data to all 10,000 schools.

Implications of the research
The research has implications for policy, practice and theory.The implication for policy-makers is that 'big picture' announcements, such as advocating distributed leadership through the Malaysia Education Blueprint, may have limited impact at school level, particularly where it contradicts existing cultural assumptions, which privilege hierarchical leadership.For principals, the data indicate that partial enactment of distributed leadership helps in sharing leadership workloads but largely misses the opportunity to develop future leaders.The implication for theory is that the data indicate a need to modify conventional distributed leadership theory, which stresses 'emergence', to recognize that an allocative leadership model (Bolden et al, 2009) may be more appropriate for centralized contexts.Both sub-models serve to enhance greater leadership 'density', thus potentially securing the benefits of multiple leadership identified by Leithwood et al (2006).However, the Malaysian version of distribution is almost indistinguishable from the management concept of delegation, suggesting that the widely supported notion of distribution has been 'captured' to provide a more 'acceptable' label for traditional management activities.The belief that leadership may emerge from anywhere in the organization, a central feature of distributed leadership theory, is discredited by the Malaysian research which shows that it is a property of the hierarchy and subject to control by the principal.Further research is required to establish whether this constrained form of distribution is also evident in other centralized contexts.